摘要:

J. MATER. CHEM., 1991, 1(4), 591-593 59 1 Second-harmonic Generation from Mixed Crystals of p-Nitroaniline with Substituted Benzenes studied by the Powder Method Ryoka Matsushima,* Hideo Takeshita and Naomichi Okamoto Department of Applied Chemistry, Faculty of Engineering, Shizuoka University, Hamamatsu 432, Japan Second-harmonic generation (SHG) from mixed crystals of pnitroaniline ( pNA) as guest and various substituted benzenes as host has been measured by the powder method with a 1064nm YAG laser source. The SH intensities, which varied with crystallization conditions as well as host and mixing ratio, were as high as 100-189 x urea with pdihydroxybenzene or pdicyanobenzene as host. The SH intensity tends to increase with the size of the powdered crystals, implying a phase-matchable feature.The SHG activities were stable only at low (5 "C) temperature. It is speculated that epitaxial growth of pNA crystals on the surface of host crystals would lead to an SHG-activity morphology. Keywords: Second-harmonic generation; p-Nitroaniline; Substituted benzene Non-centrosymmetric alignment is essential for macroscopic polarizability, or second-harmonic generation (SHG), in solid- state polar molecules.' Various approaches have been tried to control molecular orientation, e.g. by the use of Langmuir- Blodgett layers, host-guest interactions,2 poling under high electric field^,^ as well as direct modification of guest mol- ecule~,~but SHG activities have tended to decrease as a result of thermal relaxation.Recently, highly stable and efficient SHG polymers have been developed by chemical cross-linking under a high electric field.' Epitaxial growth could be an effective and convenient way of obtaining highly aligned guest molecules with high SHG activitieq6 although the aligned crystals thus obtained may be thermally unstable. Occasionally, however, thermally stable and SHG-active crystals could be obtained via spontaneous alignment in crystals by suitable modifications and/or combi- nations of guest and host molecule^.^^*^ This paper is concerned with preliminary results on SHG from mixed crystals of p-nitroaniline (p-NA) as guest with substituted benzenes as host, implying a simple and effective method for spontaneous molecular alignment in crystals.Experimental Mixed crystals were obtained as follows, after preliminary examination of the solvent conditions, the temperature, and the evaporation rate. The host and p-NA, in a particular host :guest ratio (Table l), were dissolved in warm tetrahydro- furan: the solution was cooled and evaporated at 15 "C for 10 min under reduced pressure. The SHG signals were col- lected in a specially modified integrating sphere, and measured by the powder method' with a Q-switched Nd:YAG laser (1064 nm), using powdered crystals of urea (0.05-0.1 mm diameter) as standard. In order to collect the second-harmonic waves from samples, an integrating sphere and a small hemi- spherical mirror were set up in contact with the sample cell on the reflecting and transmitting sides, respectively, as described in detail in a recent pa~er.~~'~ Results and Discussion In Table 1 SHG activities are summarized for powdered crystals obtained from mixtures of p-nitroaniline (p-NA) as guest and various proportions of host molecules.The SHG activity was strongly dependent on the conditions of crystalliz- ation, viz. solvent, temperature, and/or the rate of solvent evaporation. Thus, even the freshness (purity) of solvent for crystallization significantly affected the SHG of p-NA-3 (Table 1). It is assumed that the solubilities of the solutes would be changed by such impurities as moisture, thus affecting the rate of crystal growth or morphology.' On the other hand, mixed crystals of rn-nitroaniline (m-NA), p-(N-isopropy1)nitroaniline(IPNA), or p-(N,N-dimethyl) nitroaniline (DMNA) as guest showed no enhancement of SHG activity.Thus, none of the mixed crystals of IPNA and N-alkylated p-nitroanilines featured enhancement of SHG Table 1 Second-harmonic generation from mixed crystals of p-NA as guest and substituted benzenes as host" host p-dih ydrox ybenzene p-dini trobenzene p-dicy anobenzene p-dicyano benzeneb p-diamino benzene p-cyanophenol p-dimethox ybenzene rn-nitroaniline mixing ratio of host :p-NA no. 1.o :0.0 1.O:O.l 1.0:0.5 1.0: 1.0 0.5 :1.0 0.1 :1.0 0.0 : 1.o 1 0.0 0.03 89.5 137.0 179.7 156.9 30.9 2 0.0 6.8 48.7 25.7 7.9 5.9 30.9 3 0.0 6.6 25.8 26.3 165.0 50.5 30.9 3 0.0 0.2 8.0 53.2 433.7 1.7 30.9 4 0.0 0.0 3.5 0.0 0.0 3.2 30.9 5 0.0 10.5 4.0 1.3 81.2 156.8 30.9 6 0.0 0.3 47.4 30.7 0.1 0.5 30.9 m-NA 12.1 14.8 65.9 103.7 97.3 132.1 30.9 a SH intensity relative to that of urea.Tetrahydrofuran solution containing p-NA and a host compound was evaporated over 10 min at 15 "C.* Less fresh tetrahydrofuran which might have contained some impurities such as moisture was used as solvent to compare with the fresh. J. MATER. CHEM., 1991, VOL. 1 activity (not exceeding the value 1.50 for pure IPNA) as shown in Table2. Similarly, SHG intensities from m-NA mixed with p-dihydroxybenzene, 1 (mixing ratio m-NA : 1 is shown in parentheses), were 12.1 (1 .O :O.O), 8.5 (1.O :0.l), 5.3 (1.0:0.5), 5.0 (l.O:l.O), 3.1 (0.5: l.O), 1.6 (O.l:l.O), 0.0 (0.0: 1.0) x urea.Intensities from DMNA mixed with various hosts in a 1 : 1 ratio were no larger than that of pure DMNA, 6.64 x urea: 0.27 (host l), 0.41 (2), 0.09 (3), 0.10 (4), 0.31 (5), 0.29 (6)x urea. Because of the sterically hindered amino group of IPNA, or the lack of hydrogen-bonding protons of DMNA, these guest molecules are unfavourable for intermolecular hydrogen bonding, while the bent structure of m-NA is incapable of linear head-to-tail alignment. Therefore, it may be assumed that linear alignment with head-to-tail hydrogen bonding is essential for the enhanced SHG of p-NA." On the other hand, hydrogen-bonding protons are not necessary with host molecules, since 2, 3, and 6 are still effective hosts for the enhanced SHG.It is also notable in Table 1 that a low mixing ratio, host :p-NA =0.1 : 1.O, features relatively high SHG activites. From these results we speculate tentatively that the host substrates play a role in initial nucleation to offer a small crystalline surface on which the guest crystal can grow (as for epitaxial growth) with extensive intermolecular hydrogen bonding to give an SHG-active morphology. Fig. 1 illustrates the SHG intensity of p-NA-3 (1.0 :0.5) as a function of the size of the powdered crystals. The SHG intensity tends to increase with the crystal size, implying a phase-matching ability. The SHG activities were, however, thermally unstable; thus SHG activity of p-NA-1 (1.O :0.5) decreased with temperature from 179.7 (room temperature) to 129.5 (50 "C) and 0.17 (80 "C)on heating for 5 min under nitrogen atmosphere.Only at low temperatures is the SHG activity retained, as shown in Fig. 2. Table 2 Second-harmonic intensities (x urea) of mixed crystals of IPNA as guest with N-alkylated p-nitroanilines as host' mixing ratio of IPNA: host N-alkyl group of host 1.O:O.l 1.0:0.5 l.O:l.O 0.5:1.0 0.1 : 1.0 methyl 0.67 0.13 0.08 b 0.00 bethyl 0.15 0.50 0.25 0.08 dimethyl 0.16 0.51 0.47 0.27 0.17 diethyl 1.50 1.10 0.87 0.79 0.12 'Diethyl ether was used as solvent, and evaporated over 20 min at 25 "C. Very weak. 0 150-B 10 0-> CI.-rn Q, '0.-Cc. I v) 0 50-0 0 O 00 I 1 I 0 1000 2000 time/h Fig.2 Thermal stability of the SHG activity of p-NA-1 (1.0: 1.0) at 25+5 "C (0)and 5_+3"C (0) Conclusions The mixed crystals of p-nitroaniline as guest and substituted benzenes as host, under suitable conditions of mixing ratio and rate of crystallization, show phase-matchable SHG activi-ties as high as 100-200 x urea owing to spontaneous molecular alignment in the crystals. The SHG activities are, however, stable only at lower temperatures. Since no such enhancement of the SHG activity is observed with m-nitroaniline or p-(N,N-dimethy1)nitroanilineas guest, it is assumed that linear alignment of guest molecules with head-to-tail hydrogen bond- ing is essential for the enhanced SHG activities of the mixed crystals.References (a) D. J. Williams, Nonlinear Optical Properties of Organic and Polymeric Materials, ACS Symp. Ser. 233, American Chemical Society Washington D.C., 1983;(b) D. J. Williams, Angew. Chem., Int. Ed. Engl., 1984,23, 690. (a) G. Berkovic, T. Rasing and Y. R. Shen, J. Opt. SOC.Am. B, 1987,4, 945; (b) I. R. Girling, N. A. Cade, P. V. Kolinsky, R. J. Jones, I. R. Peterson and M.M. Ahmad, J. Opt. SOC.Am. B, 1987, 200-4, 950; (c) G. C. Cross, I. R. Girling, I. R. Peterson, N. A. Cade 1-> and J. D. Earls, J. Opt. SOC.Am. B, 1987,4,962;(d)T. Miyazaki, u) T. Watanabe and S. Miyata, Jpn. J. Appl. Phys., 1988,27, L1724;Q)c .-I v) e/100--me< m-50 100 150 200 particle size/,um Fig.1 Variation in SHG activity with crystal size of p-NA-3 (1.O :0.5) (e) G. J. Ashwell, E. J. C. Dawnay, A. P. Kuczyniski, M. Szablewski, I. M. Sandy, M. R. Bryce, A. M. Grainger and M. Hasan, J. Chem. Soc., Faraday Trans., 1990,86, 1117;(f)S. D. Cox, T. E. Gier and G. D. Stucky, Chem. Mater., 1990,2, 609. (a) K. D. Singer, J. G. Kuzyk, W. R. Holland, J. E. Lalama, R. B. Comizzoli, H. E. Katz and M. L. Schiling, Appl. Phys. Lett., 1988,53, 1800; (b) M. Eich, A. Sen, H. Looser, G. Bjorklund,J. D. Swalen and D. Y. Yoon, J. Appl. Phys., 1989,66, 2559;(c) M. A. Mortazavi, A. Knoesen, S. T. Kowel, B. G. Higgins and A. Dienes, J. Opt. SOC.Am. B, 1989,6, 733. (a)J. F.Nicoud and R. J. Twieg, Nonlinear Optical Properties of Organic Molecules and Crystals, ed., D.S. Chemla and J. Zyss, J. MATER. CHEM., 1991, VOL. 1 593 Academic Press, Orlando, 1987, vol. 1, pp. 227-296; (b)W. Tam, B. Guerin, J. C. Calabrese and S. H. Stevenson, Chem. Phys. Lett., 1989, 154, 93. J. Calabrese, Chem. Phys. Lett., 148, 136;(c)D. Chen, N. Okamoto and R. Matsushima, Opt. Commun., 1989, 69, 425; (d) D. Chen, N. Okamoto, T. Sasaki, S. Tasaka and R. Matsushima, IEZCE 5 6 7 M. Eich, B. Reck, D. Y. Yoon, C. G. Willson and G. C. Bjorklund, J. Appl. Phys., 1989, 66, 3241. S. Tasaka, T. Abe, R. Matsushima and N. Okamoto, Jpn. J. Appi. Phys., 1991, 30,296. (a) W. Tam and J. C. Calabrese, Chem. Phys. Lett., 1988, 144, 79; (b) Y. Wang, W. Tam, S. H. Stevenson, R. A. Clement and 8 9 10 Trans., 1991, 74, 946. S. K. Kurtz and T. T. Perry, J. Appl. Phys., 1968, 39, 3798. S. R. Hall, P. V. Kolinsky and R. Jones, J. Crystal Growth, 1986, 79, 745. T. W. Panunto, Z. U. Lipkowska, R. Johnson and M. C. Etter, J. Am. Chem. Soc., 1987, 109, 7786. 'Paper 1/01080H; Received 4th February, 1991

ISSN:0959-9428    

DOI:10.1039/JM9910100591

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

J. MATER. CHEM., 1991, 1(4), 595-596 Synthesis and Properties of a New Family of Phases, Li,XYO,: X=Al, Ga; Y =Si, Ge Chnoong Kheng Leet and Anthony R. West University of Aberdeen, Department of Chemistry, Meston Walk, Aberdeen AB9 2UE, UK A family of new phases, Li,XYO,, has been synthesised by solid-state reaction. They are all thermodynamically stable, melting incongruently in the range 1165-1250 "C. Electrically, they are insulators and show very low electronic conductivity at high temperatures. Unit cells are orthorhombic, typically 5.4 A x 15.8 A x 4.9 A, with space-group Pnam. The phase diagram of the binary join Li,GeO,-LiAIO, is given, showing the presence of incongruently melting Li,AIGeO, and a limited range of LiAIO, solid solutions. Keywords: Lithium aluminosilicate; Lithium aluminogermanate During a study of new lithium-ion conducting ceramics in the system Li4Si04-LiGaSi04,1*2 the presence of an additional phase of unknown structure was indicated from the X-ray powder diffraction data.A subsequent phase dia- gram study of the system Li20-Ga203-Si02 showed its formula to be approximately Li14Ga6Sis026, although it was not obtained as a single phase.3 Follow-up work on the Al- and Ge-containing analogous systems has shown that similar phases are also formed in these systems. We have now confirmed their stoichiometry to be Li3XYOs: X=Al, Ga; Y=Si, Ge. These results are presented here, together with crystallographic, phase diagram and electrical property data for the new phases. Experimental Starting materials were reagent-grade Li2C03, A1203, GeO,, and Ga203.All were used as received since tests showed that drying was not necessary. Reagents were weighed out, mixed using an agate mortar and pestle, and using acetone to form a paste, dried and heated in Au- or Pt-foil boats at 700 "C for 2 h to eliminate CO,. The temperature was then raised to 950-1 100 "C,depending on composition. Trial and error tests on a considerable number of compositions were required to find the optimum conditions (i.e. temperature and time of firing) for reaction to go to completion. At the outset, the major objective and difficulty, was to determine the composition(s) of the new phases. Initial experi- ments were carried out in the system Li,0-A1,03-Ge0,; various compositions in the region of 'Li14A16Ge5026' were prepared, reacted and analysed by X-ray powder diffraction.Gradually, it became clear that the true composition was displaced somewhat from the supposed composition and may be close to Li3A1Ge05. It was, however, very difficult to prepare a phase-pure sample for any composition, by using a single-step reaction of the three oxide/carbonate reagents. Since the composition Li3A1Ge05 may be regarded as an equimolar mixture of Li,Ge03 and LiAlO,, stock quantities of these two phases were prepared and reacted. Using this procedure, a phase-pure sample of the new phase Li,AlGeO,, was obtained. For all subsequent work, on this and the other systems, a similar two-step reaction procedure was used.Phase analysis was carried out by X-ray powder diffraction with a Stoe Stadi psd diffractometer, Cu-Kcc, radiation. For conductivity measurements, gold-paste electrodes were t Permanent address: Chemistry Department, Universiti Pertanian Malaysia, 43400 Serdang, Selangor, Malaysia attached to opposite pellet faces, dried, fired at 200-600 "C, and the assembly loaded into a conductivity jig; a.c. impedance measurements were made using two instruments: Solartron 1250/1286 for the frequency range 0.65 Hz-65 kHz; Hewlett- Packard 4192A RF bridge for the range 100 Hz-13 MHz. Melting points were determined approximately by placing pellets in a furnace at various preset temperatures and observing whether or not melting occurred. Results and Discussion Using the procedures described above, pure samples of the three new phases, Li3A1GeOs, Li3GaGe05 and Li3GaSiOS, were prepared. Reaction conditions, for the mixtures of Li2Y03 and LiXO,, were 2 days at 1050 "C, 2 days at 1050 "C and 2 days at 1100 "C (pelleted samples) for the above new phases, respectively.The X-ray powder diffraction data were indexed success- fully, by trial-and-error Visser methods. An abbreviated listing is given in Table 1; a more complete set is available from the authors and will be submitted to the JCPDS file. Unit-cell data are summarised in Table 2; systematic absences indicate the space group to be Pnarn. The fourth phase in this group, Li3A1Si05, has been synthe- sized, but never as a pure sample.This phase was detected Table 1 Indexed X-ray powder diffraction data Li,AlGeO , Li,GaGeO, Li,GaSiO, dobs/A I dobs/A I dobs/A I ~~ 020 7.9297 13.5 7.8305 58.3 110 5.1084 17.1 5.I728 9.2 011 4.6570 4.3 4.71 11 6.9 4.6282 43.6 120 4.4592 100.0 4.5031 100.0 4.4435 100.0 040 3.9583 13.1 3.96 11 4.5 130 3.7504 5.7 031 3.5796 85.5 3.6068 90.4 3.5506 76.4 111 3.5263 39.0 3.5666 26.7 3.51 10 4.9 121 3.2896 35.6 3.3236 36.5 3.2742 29.6 140 3.1905 1.5 3.1662 12.8 131 3.0106 8.6 2.9654 21.4 150 2.7302 49.0 2.7445 49.1 2.7038 90.7 200 2.6985 25.7 210 2.6608 2.9 2.6586 12.4 060 2.6377 5.0 2.5488 4.6 002 2.4362 43.4 2.4642 49.7 2.4227 52.5 230 2.4023 11.5 2.4277 15.0 2.3960 8.O 151 2.3819 26.1 2.3974 44.8 2.3620 43.7 201 2.3607 13.8 Table 2 Unit-cell data Li,AlGeO, 5.397(1) 15.825(1) 4.873(1) 416.17(6) Li,GaGeO, 5.467(1) 15.861(2) 4.929(1) 427.4q4) Li,GaSiO, 5.398(1) 15.633(4) 4.845(1) 408.8(2) after heating at 1050-1 100 "C for up to 4days.It is possible that longer heating times would yield a single-phase product but significant lithia loss by volatilisation may also become a problem, thereby making it difficult to obtain a single-phase sample. Further work is required, probably using alternative reaction pathways, in order to obtain single-phase samples of Li3A1Si05. In order to obtain more information on the thermal stability and melting behaviour of the new phases, a brief study of the phase diagram Li,GeO3-LiA1O2 was made; the results are shown in Fig.1. Melting information was obtained from the appearance of pelleted samples after they had been placed inside a furnace for a few minutes at various preset tempera- tures. From this, it was clear that Li3A1Ge05 melted incongru- ently to LiAlO, and liquid at 1165f10 "C and that a eutectic existed between Li3A1Ge05 and Li,Ge03 at 11 10 & 10 "C. To confirm the incongruent melting of Li,A1Ge05, a sample was heated briefly at 1180 "C and quenched. X-Ray diffraction of the product showed a rather poor-quality pattern with evidence of significant amounts of LiA102, together with smaller amounts of Li3A1Ge05 and Li2Ge03. This showed LiA10, to be present at the temperature of heating, together with liquid which subsequently crystallised to give a mixture of Li,Ge03 and Li3A1Ge05 on cooling.There was no evidence for any solid-solution formation with either Li3A1Ge05 or Li,Ge03, but LiAlO, does appear to form a limited range of solid solutions. The phase diagram, Fig. 1, may be regarded as a true binary join within the ternary system Li,0-A1203-Ge02. This is I I I1 I I I1 1 II Li,GeO, 20 40 60 80 Li A102 LiAIO, composition (mol%) Fig. 1 Phase diagram for the join Li,Ge03-LiA10,; compositions studied are marked 0 J. MATER. CHEM., 1991 VOL. 1 I I I I I I 0.8 1.0 1.2 1.4 1.6 1.8 103 KIT Fig.2 Arrhenius conductivity plot for Li,AlGeO, (a) and Li,Ga- GeO, (b) Activation energies are 1.02 and 1.08 eV, respectively, (a) 0,heating; @, cooling; (b) @, heating; 0,cooling. because all the phases that are present on the diagram have compositions which also lie on this join. The other two new phases, Li3GaGe05 and Li3GaSi05, also appear to melt incongruently, in similar fashion, at 1250&10 and 1210&10 "C, respectively. The phase diagrams of their corresponding Li,Y03-LiGa02 joins have not been determined, however, but almost certainly will be more com- plex owing to the appearance of phases such as Li5GaSi208.1*2 Conductivity data for two of the new phases are presented as Arrhenius plots in Fig. 2. The data are very similar, giving linear plots over the temperature range 300-900 "C with activation energies of 1.05 eV. The conductivity values are very low, 2x10-' Q-'cm-' at 300°C. The conducting species appear to be electronic since no evidence of electrode- polarisation effects was seen in the impedance data., This is therefore in complete contrast to the Al, Ga-doped Li4Si04/ Li,GeO, materials, which exhibit high lithium-ion conduc- tivity, lop6 0-' cm-' at 25 0C.1*596. References 1 P. Quintana, F. Velasco and A. R. West, Solid State lonics, 1988, 34, 149. 2 P. Quintana and A. R. West, J. Solid State Chem., 1989, 81, 257. 3 P. Quintana and A. R. West, Br. Ceram. Trans. J., 1989, 88, 17. 4 J. T. S. Irvine, D. C. Sinclair and A. R. West, Adv. Muter., 1990, 2, 132. 5 A. Garcia, G. Torres-Treviiio and A. R. West, Solid State lonics, 1990,40/41, 13. 6 C. K. Lee and A. R. West, in preparation. Paper 1/00564B; Received 6th February, 1991

ISSN:0959-9428    

DOI:10.1039/JM9910100595

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

J. MATER. CHEM., 1991, 1(4), 597-610 La,- ,Sr,CuO,-,: Structural, Magnetic and Transport Measurements on Antiferrornagnets, Insulators and Superconductors Matthew J. Rosseinsky,*."#t Kosmas Prassides*Tb and Peter Day" a Inorganic Chemistry Laboratory, Oxford University, Oxford OX1 3QR, UK School of Chemistry and Molecular Sciences, University of Sussex, Brighton BN 1 9QJ, UK lnstitut Laue Langevin, 38042 Grenoble, France The structural, magnetic and conducting properties of the La2-$rxCu0,-b system (01 ~50.13,0.01 16 10.04) are examined as functions of temperature, and combined with assignment of the formal Cu oxidation state. High-resolution powder neutron structural results confirm that the correct space group is Abma, ruling out charge-density wave behaviour as well as other Fermi-surface instability mechanisms, as being responsible for the tetragonal-to-orthorhombic phase transition.The reduced magnitude and eventual disappearance of the orthorhombic distortion as the formal Cu oxidation state increases is well described by the better matching of the (La,Sr)-0 and Cu-0 bond distances in the two layers, (Ln202) and (CuO,), resulting in the tolerance factor increasing into the stability range of the K,NiF, structure. The structural data reveal that there is a smooth variation of all structural parameters as the formal Cu oxidation state increases, except in the bond-length ratio (dax-deq)/(dax+deq),where d,, and deqare the axial and equatorial Cu-0 bond distances of the severely elongated CuO, octahedra. This ratio peaks at the formal Cu oxidation state nearest to +2 in our series (i.e.+2.01) to a value of 0.1207(3), decreasing at both higher and lower oxidation states. An electronic Jahn-Teller mechanism is found superior to a superexchange mechanism in explaining the orbital ordering, indicating that the holes formed on Sr2+ doping have c*character. Analysis of the magnetic Bragg scattering resulting from antiferromagnetic ordering of the copper spins leads to the evaluation of the staggered Cu moment which is reduced on doping and disappears at a formal Cu oxidation state of between +2.01 and +2.04. The data support localised models for the Cu spin system and provide no evidence for spin-density wave behaviour. Analysis of resistivity data at low temperatures shows the occurrence of variable-range hopping even after the long-range magnetic order has disappeared, demonstrating that the electronic states at the Fermi level are localised.Doping by oxygen vacancies and/or Sr2+ cations leads to the introduction of impurity states into the Mott-Hubbard gap. Each carrier extends over two or three copper sites, with the localisation length < increasing to beyond the mean interdopant separation on approaching the metal-insulator transition. As the temperature increases, there is a crossover to a transport mechanism based on excitation to the mobility edge in the overlapping upper-Hubbard impurity and valence bands with the metal-insulator transition being of the Anderson type. Keywords: Lanthanum-strontium-copper-oxygen la tor transition 1.Introduction Although numerous binary and ternary transition and B-subgroup metal oxides are metallic conductors, very few are' suaerconductors. Mixed valency' seems to be a prerequi- site for &perconductivity in these systems, e.g. in the spinel2 Lil +xTi2-x04 and the pero~skites~.~ BaBi, -,Pb,03 and Ba, -,K,Bi03; this may be because strong electron-phonon coupling is thereby asvured, despite t& relatively low density of states at the Fermi energy. However, traditional ideas about superconducting ceramics have changed dramatically with the discovery of superconducting oxides based on ternary cup rate^;^ these phases have proved to be immensely versatile, leading to a whole series of structurally distinct families of superconducting oxides with T, as high as 126 K6 However, despite the large progress made in raising the T, in this class of materials, no clear theoretical explanation of the phenom- enon of superconductivity in these and related systems has yet emerged.A common structural feature in the copper oxide ceramics has been the presence of layers of corner-sharing square Cu04 units. These CuO, layers are present in both hole- and electron-doped superconductors. For example, in the 1-Present address: AT&T Bell Laboratories, Murray Hill, NJ 07967, USA. system; Superconductivity; Antiferromagnetism; Metal-insu- La, -$~,CUO,-~ series (O/T phases), they form part of the strongly elongated octahedral CuO, units; in the Nd2-,Ce,CuO, -series (T' phase^),^ Cu has a square-planar co-ordination, while in the Nd, -,-,Ce,Sr,CuO, -series (T* phases),8 the CuO, layers are formed by corner-sharing square-pyramidal Cu05 units.The La, -,Sr,CuO, series of compounds have proved important prototypes to use as a test bed for evaluation of the mechanism of superconductivity in all the 'high-T,' materials. They are structurally simple; their stucture is based on that of K2NiF4 and there is only one Cu site per unit cell. One can also systematically vary the formal Cu oxidation state by simple chemical doping since the La: M" ratio controls the filling of the conduction band. Furthermore, simple variation of the dopant cation M does not have a monotonic influence on the magnitude of T, with changing ionic radius (for x =0.15, T, is as high as 42 K for Sr, 36 K for Ba, and 26 K for Ca).Finally, the existence of the parent compound in this series, L~,CUO~-~, with its own unusual magnetic behaviourg introduces the competition between magnetism and superconductivity in any model for the 'high- T,' compounds. The importance of the question of how lattice distortions and their temperature and composition evolution are related to the transition from insulating to metallic to supercon- ducting behaviour was immediately recognised. lo Initial 598 suggestions that charge-density wave onset in orthorhombic LazCu04-d is suppressed by IIA cation doping to give a tetragonal superconducting phase'', l2 were disproved by early powder neutron diffraction studies," and showed the import- ance of using high-resolution powder neutron diffraction to accurately determine the crystal structure.Similar arguments apply to the accurate knowledge of the magnetic structure since various model^'^-'^ emphasize the role of strong elec- tronic correlations, prompted by the high TN (ca. 250 K) of La,Cu04. For example, it was originally suggested" that the antiferromagnetic transition is a spin-density wave instability of the square two-dimensional Fermi surface at x =0, being rapidly suppressed on doping and giving a superconducting ground state. Furthermore, there has been some ambiguity in the literature16 about the true orthorhombic space group, adopted by the members of this series of oxides as a function of temperature and composition; besides Crn~a,'~*'~-'~ the space groups Pccn2' and Crnrnrn" have also been proposed on the basis of single-crystal X-ray diffraction and low-resolution powder neutron work.A possible monoclinic dis- tortion from orthorhombic symmetry has also been pro- Finally, for Ba2+ doping levels in the vicinity of x =0.12, a phase transition to a low temperature tetragonal phase (Cmca+P4,/ncrn) occurs.z4 The existence of the low- temperature tetragonal phase has a detrimental effect on the superconducting properties of these materials, strongly affect- ing the transport proper tie^.,^ In this paper, we report a systematic examination of the structural, magnetic and electron transport behaviour in the La, -,Sr,CuO, -model system for samples of well defined formal copper oxidation state in the vicinity of the metal- insulator transition. We have used four-probe conductivity measurements and high-resolution powder neutron diffraction together with careful analytical measurements to investigate such behaviour across the insulator-to-metal-to-superconduc-tor transition which occurs as a function of x and 6 in this system, with T, reaching a maximum of 42 K at x=O.15 and 6 =o.26 2.Experimental In order to prepare samples with highly homogeneous dopant distributions, the solution-based citrate sol-gel technique was empl~yed.'~,~~Samples (20 g) of Laz -,Sr,Cu04 -(x =0,O.O 1, 0.03 and 0.06) were prepared in this way; firing temperatures of 650 (2 h) and 950 "C (48 h, one intermediate regrind) were used. The La1.87Sro.13Cu04-d sample used in this work was prepared by the carbonate co-precipitation technique." Powder X-ray diffraction confirmed that the samples were monophasic.Elemental analysis of Sr and Cu was performed by atomic absorption spectrometry. Sample homogeneity at the microscopic level was confirmed by energy-dispersive X-ray analysis on 15 crystallites of each sample in a JEOL 2000 FX electron microscope. Data to determine the La :Cu ratio were collected with the beam incident on the edges of the thinnest ~rystallites.,~ Oxygen concentration was evalu- ated by thermogravimetric reduction of 100 mg of the sample at 950°C in flowing 95%Ar-5%H2 on a Stanton Redcroft STA 785 thermobalance.The sample temperature was raised to 950 "Cat 10 "C min-' in flowing dry Ar (with no observable weight loss, except for a small amount of moisture at 100 "C), maintained at 950 "C for 30 min, followed by introduction of the Ar-H, mixture. Reduction proceeded to completion within 20 min (the flow of reductant was maintained until no further weight loss was observed for 30 min). Oxygen concen- J. MATER. CHEM., 1991, VOL. 1 tration was evaluated according to the reaction scheme: La, -xSr,Cu04-s(s)+ (1 -x/2)La,03(s) +xSrO(s)+Cu(s)+(1 -6)H20(g) The identity of the reaction products was confirmed by powder X-ray diffraction on samples cooled in the Ar flow.Measurements were repeated twice with consistent results. The copper oxidation state in was also deter- mined independently by iodometric titration using the double- titration method of Nazzal et aL3' Electrical conductivity measurements were performed between 300 and 4.2 K using the four-probe technique in an Oxford Instruments CF200 cryostat. Silver paint was used to make contact onto samples of dimensions 2mm x2mmx4mm cut from the sintered pellets with a diamond saw. Contact resistances were between 7 and 15 Q. The ratio of nested :unnested resistances was greater than 100:1. A.c. mutual inductance measurements show that the x =0.13 sample is superconducting with an onset temperature of 36 K and a transition width (10-90Y0) of 4 K.Powder neutron diffraction profiles of samples with x =0, 0.01, 0.03 and 0.06 were measured at 4.2 K on the D2b high- resolution powder diffractometer at the Institut Laue Lange- vin, Grenoble. The instrument was operated in its high- resolution mode at a mean neutron wavelength of 1.5942A. A profile of L~,CUO,-~ was also recorded at 300 K. The samples were mounted in vanadium cans placed in an ILL 'orange' cryostat. Diffraction data were collected in steps (28) of 0.025'. Diffraction profiles of the La1.87Sro.13Cu04-6 sample were recorded on the high-resolution powder diffractometer (HRPD) at the ISIS pulsed neutron source, Rutherford Appleton Laboratory, with the sample at the 1 m position over a range of temperatures.3. Results 3.1 Chemical Stoichiometry The analysis for Sr and Cu by atomic absorption spectrometry indicates that all the samples prepared by the citrate route have Sr concentrations close to those of the precursor prep- arations. In contrast, the sample prepared by the pH-adjusted carbonate route" had a slight Sr deficiency compared to the starting composition, i.e. La1.87Sr0,13C~04-dinstead of Lal~8sSro.15Cu04-s.The results of the oxygen concentration determination by thermogravimetric analysis (TG) reduction in flowing 95%Ar-5%Hz are shown in Table 1. We also note that the iodometric titration of L~,CUO,_~ gives a value of +1.90 for the formal oxidation state of Cu, in good agreement with the TG results. Furthermore, the electron-microscopy results show that homogeneous samples are prepared with no detectable vari- ation of La :Cu ratios, and hence, no consequent segregation.In all but the x=O.O6 sample, the Srz+ concentration was below the detectability limit. The samples were shown explicitly to correspond to the desired composition to within one standard deviation. Using calibration constants Table 1 Thermogravimetric analysis results and deduced stoichio- metries in the La, -,Sr,Cu04-b samples studied 6 deduced composition Cu oxidation state 0.04(1) La2CuO3.9, +1.92 0.04(1) .99sr0.01cu03.96 +1.93 0.01(1) La1 .97sr0.03cu0j.99 +2.01 0.01(1) La1 .9&0.06CU03 .99 +2.04 O.Ol(1) La, .87sr0.13cu03.99 +2.11 J. MATER. CHEM., 1991, VOL. 1 599 K(La-La/Cu-Ka)=0.94 and K(Sr-Kcr/Cu-Ka)= 1.61,,' rl;metal-insulator)which decreases with increasing x. The activa- the energy-dispersive X-ray analysis results led to compo- tion energy E, also decreases smoothly.The validity of more sitions .99(3)cu04 -6, La1.~9(4)~r0.01~u~4-complex alternatives to the simple Arrhenius model was also S, Lal.98(3)sr0.03cu04 -8 and .95(6)Sr0.05(2)CU04-S for the investigated. The anti-adiabatic small-polaron model3, samples with intended Sr2+ dopant levels of x=O, 0.01, 0.03 proved the most satisfactory, except for the x=O.O3 material and 0.06, respectively. The small error limits reflect the where a totally satisfactory fit could not be obtained. Inclusion precision of the measurements rather than the accuracy of of the temperature-dependent polaron mobility did remove the energy-dispersive X-ray analysis which is ca.10%. We the change of sign in the temperature dependence.7 The note that the largest spread in composition is shown by the parameters extracted from both the Arrhenius and anti- x=O.O6 sample. It will be shown later that this sample is in adiabatic small polaron models of the resistivity are collected the region of the phase diagram close to the metal-insulator in Table 2. transition. It is a feature of materials close to such transitions Below ca. 70 K, the Arrhenius plots developed marked that they are susceptible to phase separation as the material curvature. In view of the three-dimensional variable-range becomes unstable close to the transition.The reduced micro- hopping observed in La, -.LiXCuO4-pellets and single scopic homogeneity of this sample then may reflect this crystals,34 a temperature dependence of the resistivity of the tendency. Finally, we find no evidence for La non-stoichi- form ometry, although it should be noted that La-deficient phases exist owing to the possibility of forming intergrowths of P =Po(T/To)"2 exp C(TO/T)"I (2) LaCuO, perovskite layers3, was assumed35 at low temperatures. Owing to the limited temperature range over which variable range hopping is 3.2 Conductivity found, it is difficult to evaluate the exponent v unambiguously. Two independent fitting techniques were employed: (i) the The temperature dependence of the resistivity of gradient of a plot of log [In (pT-'12)]versus T was evaluated La,-xSrxCuO,-S samples with x=O, 0.01, 0.03 and 0.06 is by least-squares fitting; (ii) the exponent v was increased from shown in Fig.1. There is an upturn in the resistivity that 0.16 to 0.60 in increments of 0.01 and a least-squares fit to a moves to lower temperatures with increasing dopant level x. plot of In p versus T-' was carried out for every value of v; The magnitude of the resistivity also decreases. Assuming an the quality of the fit was evaluated by the magnitude of the Arrhenius behaviour appropriate for broad-band semi-conductors, a plot of the logarithm of resistivity against t Supplementary material available (SUP 56834, 2 pages); details T-' shows that the slope changes sign at a temperature from Editorial Office.'\,0 , I '?., 3 La2Cu03.96 -1 Q CJ) ............-..--2 ......................"1 160 200 TIK Fig. 1 Temperature dependence of the resistivity for samples L~,-,S~,CUO,-~ with x=O, 0.01, 0.03 and 0.06 Table 2 Parameters extracted from the Arrhenius and small polaron fits to the conductivity data of the La2-,Sr,Cu04-d samples studied composition E,"/meV a0"/Q-' cm-' Eab/meV tpdbleV La2cu03.96 10 40 25 0.01 La 1.9gSr0 .01 3.96 5 35 18 0.01 La1.97Sr0.03Cu03.99 1 28 13 0.03 La1.94Sr0.06Cu03.99 0.6 300 I1 0.03 ~~~~~ ~~~~~~~~~~~~~~~~ ~ ~~~~~ ~ a Arrhenius model: Q =o0exp (-E,/kBT); Anti-adiabatic small polaron model:33 Q =(n~1'2e2u2t~,)/(2hk~~2E~~2T3/2)exp (-E,/kBT). discrepancy index p, given by The dimensionality of the hopping process was then evaluated from the dependence of variable-range hopping (VRH) in d dimension^^^^^^ on the dimension d, using the expression: pccexp (-1/P+') (4) Thus it was found that LazCuO3.,, shows the presence of a Coulomb gap in the density of states,37 whereas the doped materials showed two- or three-dimensional VRH behaviour.Table 3 summarizes the preferred exponents v on the basis of the above tests and the corresponding variable-range hopping mechanism. 3.3 Powder Neutron Diffraction 3.3.1 La2-xSr,Cu04-a; x=O-0.06 The aims of this part of the study were to evaluate, in detail, the structural evolution in the La2-,Sr,Cu04-, series as the formal Cu oxidation increases and to address the question of itinerant uersus localised electron behaviour by investigating the effect of cation doping on the periodicity of the antiferro- magnetic order.Long scan times of up to 18 h were used in an effort to measure the extremely weak magnetic scattering from the S =4Cu2+ ions above the background. As a conse- quence, the much more intense nuclear scattering was recorded with excellent counting statistics. The raw data were merged and, after background subtrac- ti~n,~*profile refinements were performed by using the Rietveld profile method,39 and incorporating a pseudo-Table3 Variable-range hopping fits to the conductivity data of the La, -.Sr,CuO, -samples studied composition .96 La1.99Sr0.01 3.96 .Wsr0.03cu03.99 La 1.94Sr0 .06cu03 .99 2D =two-dimensional; 3D V mechanism" 1/2 Shklov~kii-Efros~~ 113, 1/4 1/4 1/4 2D, 3D 3D 3D =three-dimensional. J.MATER. CHEM., 1991, VOL. 1 Voigt function peak shape des~ription.~' The starting model for all the refinements was the 22K structure" of La1.87Sr0.13C~03.99in space group Abma. The refinements were completed with anisotropic temperature factors on the (La/Sr) and the oxygen sites. Oxygen deficiency was handled by constraining the total oxygen content to be that given by the chemical analysis and allowing refinement to determine the best oxygen distribution over axial and equatorial sites. The observed, calculated and difference profiles for La1~99Sro~olCu03,96at 4.2 K are shown in Fig. 2.The pos- itional parameters from the refinements are given in Table4 with the bond lengths and angles for the copper and La/Sr sites in Tables 5 and 6, respectively. Initial inspection of the bond lengths resulting from the x =0.03 refinement revealed a Cu-Oo,, bond length of 2.422(1)A which was unexpectedly long in comparison with the other compounds. To ensure that the refinement had not converged to a false minimum, the z coordinate of the axial oxygen was initially started from several different positions. The refinements always converged to the same point. Refinement of the profiles was also carried out in the Cmmm and Pccn orthorhombic subgroups of the tetragonal Z4/mmrn space group which describes the K2NiF4 structure.These were also proposed to account for the structures of this series of compounds20,21 and describe different possible structural distortions of the square-planar layer of corner-sharing CuO, octahedra in the parent K2NiF4structure. The Cmmm space group describes a charge-density wave with a periodic expan- sion-contraction distortion of the CuO, units with a modu- lation wavevector 4CDw=(OOc*). In Pccn there is a rigid tilt of the Cu06 octahedra about the [loo] direction of the 14/mmm cell, i.e. the Oe,-Cu-Oe, bond direction. The Cu atoms are also located at a centre of symmetry and are surrounded by three crystallographically independent 0 atoms. This would result in the existence of two unequal Cu-Oe, bonds on the basal plane with corresponding unequal transfer integrals.Refinement of the profiles using the Cmmm space group and initial parameters from ref. 21 gave clearly inferior results to the other space groups which have only one Cu site; for example, full refinement of the low-temperature profile of La2Cu03.,, led to a weighted-profile R factor R,, =0.226. As a result, the Cmmm space group was discounted. It was 6000 5000 4000 (/) 3000 c C s 2000 1000 0 I -1000 -2000 11 IIIIIIIIIIIIIIIIII~IIIIII IIIIII~III~IIH~II I I 1111 ni IIIIIII IIIIIII ~IIIII~III~IIIIIIIIII~I~IIIIIII IIIIIIIII I 11 h, 4 I I I I 77 1 I I I 1 T-Fig. 2 Observed (points), calculated (full curve), and difference profiles for La,~9,Sro~06Cu03~9, at 4.2 K J. MATER.CHEM., 1991, VOL. 1 601 Table 4 Final parameters derived from the Rietveld refinements of La2 -,ST,CUO~-~ (space group Abma)" X 0.0 0.0 0.01 0.03 0.06 T/K 300 4.2 4.2 4.2 4.2LaiSr X 0.0069(2) 0.0088(2) 0.0086(2) 0.0077(2) 0.0076( 2) z 0.36 145(6) 0.36 166(6) 0.36 I 68(5) 0.361 27(6) 0.36103(6)0.79( 3) 0.39(2) 0.44(2) 0.42(2) 0.45(2)0.72(3) 0.25(2) 0.24(2) 0.26(2) 0.31(2)0.58(2) 0.23(2) 0.1 8( 2) 0.22(2) 0.25(2)0.14(4) 0.03(3) 0.03(3) -0.03(3) -0.06(3)0.63(2) 0.30(2) 0.33(2) 0.31(2) 0.33(2)-0.330(3 -0.040l(2) -0.0393(2) -0.0379(2) -0.0346(2)0.1837( 1 0.18367(9) 0.18341(9) 0.18365(9) 0.1 8 300( 9 1.42(6) 0.72(5) 0.68(4) 0.63(4) 0.89(4)1.89(5) 0.99(4) 0.90(4) 0.92(4) 0.83(3)0.51(4) 0.32(4) 0.23(3) 0.22(4) 0.22(3) -0.22(6) -0.33(4) -0.28(4) -0.23(4) -0.18(5)7.94(2) 7.98(2) 7.97(2) 8.O 1 (2) 7.98(2)0.0076( 0.0083( 1) 0.0082( 1) 0.0076(1) 0.0068(1)0.62(4) 0.50(3) 0.44(2) OSO(4) 0.63( 3) 0.52(3) 0.37(3) 0.36(3) 0.46(3) 0.51(3)1.17(6) 0.60(5) 0.50(2) 0.43(5) 0.4 1 (4) 0.23(3) O.OO(2) 0.05(3) O.Ol(2) -0.02( 3) 7.90(2) 7.86(2) 7.87(2) 7.95(2) 7.98(2)5.40359( 6) 5.4 1622(5) 5.41232(6) 5.40381(6) 5.38992(5) 5.35783(5) 5.33438(4) 5.333 12(5) 5.331Iq5) 5.32823(5) 13.1599(2) 13.1 148(1) 13.1209(1) 13.1325(1) 13.15 12(2) 9.4 9.2 9.5 10.0 9.0 2.9 1.9 1.9 1.9 2.1 3.6 3.1 3.0 3.1 3.0 1.3 0.9 0.9 0.9 1.o " The atoms were refined in the following positions in the unit cell.La/Sr, O(1) in 8f (m):(x, 1/2, z); Cu in 2b (2/m): (0,0,O); O(2) in 8e (2): (1/4, 1/4, 4.The R factors are defined as follows:3994oR,, =[cwiI x(obs)-~(~alc)l~/~w~Y~(obs)]'/~;R:, =[(N -P + C)/cwiY,2(obs)]lj2; Rmod(l)= cII,(obs)-Zi(calc)l/cZi(obs);Re=(N -P + C)/CZ,(obs) with N =number of observables, P =number of parameters and C =number of cons train ts. Table 5 Selected copper-oxygen bond distances and angles in La, -,Sr,CuO, at 4.2 K -Oeq/OX cu -oe,/A Cu-Oax/A cu-o,,-CU/O oeq-cu (dax -deq)/(dax+deq) 0.0 1.9037( 1) 2.4 17( 1) 173.40(9) 88.9( 1) 0.1 188(3) 0.0 1 1.9026( 1) 2.416(1) 173.5(1) 89.0(1) 0.1 I89(3) 0.03 1.9004( 1) 2.422(1) 173.9( 1) 89.1(1) 0.1207(3) 0.06 1.8968(1) 2.413(1) 174.q 1) 89.2(I) 0.1 198(3) 0.13" 1.8859(3) 2.395( 1) 176.6(6) 89.7(I) 0.1 189(3) " Ref. 10 at 22 K. Table 6 Selected La/Sr-0 bond distances in La,-,Sr,CuO,-, at 4.2 K 0.0 2.349(1) 3.032(1) 2.5 15( 1) 2.7379( 3) 2.579(1) 2.680(2) 2.654(1) 0.01 2.353(1) 3.024( 1) 2.5 18( 1) 2.7 364(3) 2.579(1) 2.679(1) 2.653( 1) 0.03 2.346(2) 3.007(I) 2.525(1) 2.7349( 1) 2.585( 1) 2.680(2) 2.653( 1) 0.06 2.353(1) 2.979( 2) 2.534(2) 2.7302( 3) 2.593(I) 2.675( 1) 2.65 I( I) 0.13" 2.3 7 6( 7) 2.87( 1) 2.59(1) 2.72 l(2) 2.60 l(9) 2.65 l(9) 2.642(9) considerably more difficult to distinguish between the Pccn perature factors) of the two independent in-plane oxygen and Abma space groups on the basis of our data.Initial atoms. The quality of the Pccn refinement was also inferior parameters were taken from ref. 20. Pccn gave marginally in terms of the larger estimated standard deviations (up to lower R factors than Abma and the values of the calculated five times larger) found for the positional parameters. Inspec- structure factors were very similar.However, any improve- tion of the observed and calculated Fourier maps (cf. section ment on moving from the Abma to the Pccn space group was 3.3.2) also leads to the conclusion that Abma is the correct rejected as statistically insignificant at the 99.5% confidence space group for the structure of the compounds level on the basis of the Hamilton significance applied La2 -,Sr,CuO, -x =0-0.06 at 4.2 K. This conclusion was to the integrated intensity R factors. Pccn was also rejected also reached in a single-crystal neutron diffraction study on as the correct space group on the basis of very large corre- La2C~O~95Li0.0504,which showed that Abma was correct and lations (> 95%) observed in the variance-covariance matrix explained Pccn on the basis of a twinning law, although between the refinement parameters (z-coordinates and tem- extinction prevented refinement of temperature factors.I6 J.MATER. CHEM., 1991, VOL. 1 The search for magnetic scattering led us to a careful investigation of the low-angle part of the 4.2 K diffraction profiles for extra intensity near the background level. Weak peaks were clearly visible above the background level, while they were absent from the 300 K diffraction profile of La2Cu03.96. Data evaluation was complicated by the occur- rence of weak 1/2 scattering from the (1 10) nuclear peak which overlapped with the magnetic (010) peak (indexed on the Abma unit cell).All other magnetic peaks predicted by the collinear model of Vaknin et aL9 were masked by A/2 scattering or background, and hence no refinement of magnetic structure was possible. Two overlapping Gaussians were fitted to the region of the (OlO), peak in order to determine the variation of peak position and integrated intensity with dopant level x. The (010) peak remained centred at the commensurate (010) position within a 0.1" precision of our data, and its intensity decreased markedly until it was not observed for x=O.O6. In order to evaluate the Cu moment, using I, =CF~molo/sin8 sin 28 (5) we normalised with respect to the structure factor of the weak (040) nuclear reflection and used the antiferromagnetic form factor Aq)=0.835 at q= 1.164A-', measured by Freltoft et ~1.~'mOIOis the multiplicity of the (010) reflection.The moment direction is taken parallel to u with modulation of its direction along b. The magnetic structure factor is given by where p is given by y0SJq) with yo the gyromagnetic ratio of the neutron, Aq) the form factor, S the spin, and (4) the magnetic interaction term given by q =sin a= 1 (ais the angle between the moment direction and the scattering vector). The variation of the Cu moment calculated in this way as a function of Cu formal oxidation state is presented in Fig. 3. 3*3*2La1.87Sr0.13CU03.99 The structural characterisation of this sample has been described before." There is a second-order tetragonal-to- orthorhombic phase transition at 180 K, characterised by co- operative rotation of the Cu06 octahedra about the [llO] direction of the tetragonal Z4/mmm space group.Here we report a re-examination of these data, considering the alterna- tive space group Pccn." As in the previous section, the x2 values were very similar for both the Abma and Pccn structure refinements. The site symmetry of the La/Sr cations is reduced from m to 1 in going from Abma to Pccn, i.e. the cations are moved off the mirror plane. These cations occupy general rn0.0 I I I U 1.90 1.95 2.00 2.05 Cu oxidation state Fig. 3 Average calculated moment per Cu atom as a function of formal Cu oxidation state in La, -xSrxCu04-6 positions in Pccn and their position in the nine-co-ordinate site is not fixed along any crystallographic axis.During Rietveld refinement of the Pccn structural model, the y coordinate of the La/Sr cation oscillated about zero. Reduction of the applied shift by a factor of 0.01 combined with the introduction of a slack constraint on the La,Sr-Cu distance, allowed the refinements to converge but, as in section 3.3.1, the Pccn refinement was judged inferior to that obtained in Abma. In particular, the z coordinates of the two indepen- dent in-plane oxygen atoms differed by an order of magnitude less than their estimated standard deviations, and the e.s.d.s obtained in Pccn were again considerably larger. As the two structures differ in the axis about which the octahedral tilt occurs, inspection of difference Fourier sections at x=O.25 in the (100) plane was undertaken in order to investigate whether there was any evidence for two different oxygen positions. This showed unambiguously the superiority of the Abma refinement, with large differences between Fobs and Fcalc apparent for Pccn (Fig.4).It was therefore concluded that there was insufficient evidence for two distinct oxygen pos- itions to prefer Pccn to its supergroup Abma. 4. Discussion 4.1 Crystal and Electronic Structures The orthorhombic structure of the x<O.O6 and x=O.13 (T<180 K) phases results from alternate rigid tilting of the Cu06 octahedra about the [I lo] direction of the tetragonal unit cell, which becomes the b axis of the enlarged orthorhom- bic unit cell.The two cells are related by a 45" rotation: a0 =(21'2aT + 2E)cos CI (7) bo =(21'2aT-2E) where E is the elastic deformation of the basal plane and CI is the tilt angle about the [l lo] axis. The octahedral tilt can be taken as the primary order parameter of the tetragonal+or- thorhombic transiti01-1,~~ while the elastic shear does not lead to a change in the unit-cell dimensions and behaves like a secondary order parameter. In Fig. 5, we plot the orthorhom- bic distortion (a-b)/(a+b) as a function of formal Cu oxi- dation state at 4.2 K. The occurrence of the tilting transition in La2Cu04 and its suppression at room temperature in superconducting La1.85Sr0.15C~04led to suggestions that the transition" was of a Peierls nature leading to a gap at E~.This was discarded when high-resolution powder neutron diffraction" showed that doping only served to reduce the transition temperature for the tilting transition.In fact, although the vector Q= (1/2)(u*+b*), where u* and b* are reciprocal primitive unit vectors, nests the square two-dimensional Fermi surface, the tilting nature of the distortion cannot open a gap as it does not create two crystallographically inequivalent copper sites. Weak coupling theories were then built on the assumptionu that the van Hove singularity in the two-dimensional density of states of the tight binding square lattice splits upon the T+O transition. If the transition was driven by an instability of the two-dimensional Fermi surface, it was suggested that the orthorhombic distortion would change the Brillouin zone from a square to a rectangle, the two saddle points becoming inequivalent and the van Hove singularity in the density of states would split into two parts, symmetric with respect to half filling.However, this would require two different Cu-0 distances in the CuOz sheets with two different Cu-0 hopping integrals. The distinction between Pccn and Abma is of vital importance in this respect as there are two different Cu-0 distances in the plane in Pccn. We have shown that Pccn is not the correct space group, discounting the possibility J. MATER. CHEM., 1991, VOL. 1 0 ........... ......-* .. ... ...... .........: ................................. .... .... . . .. I. '....,..: ... ..' ...... .. ... .......... 0 :..__. . . ._..... ..': .-.. *.. -0.32 ClOO] -3 0.32 .... 0.32....*.,.V; f...........................: .I ... ... ...........; ;' ........ ..,......... . . ........ . .. -. .....a: ._.. .. ..... .. .... ;... ...................... *.. . .. ..,:. ; .. ........... .__....... rn c .. 0 0 Y ....... ... . .. .....L.. ........:. . ..... .*. ._. . ,..*; : ! ............ ........... ... ...... ..... : ; ........ ...... .. .. ....,. ..1. r -. ..... .. .. ,. ,.,. .. ..... ..;. ....t : ..:. ._ .......... ..-0.32 .. ...:.. ..-18 lii Fig. 4 Observed (a),(c),and calculated (b),(d),Fourier maps at x =0.25 in the (1 10) plane of (a) and (b)the Abma and (c) and (d) Pccn structural models for La,,,,Sr,,,,CuO,,,, of such a Fermi surface instability mechanism. In Aha, the four Cu-0 near-neighbour distances remain equal and very similar in magnitude to the tetragonal phase, and (a,,co), (b,,co) are maintained as mirror planes. The Brillouin zone becomes a truncated rhombus and the Fermi surface for half 0 filling is a rectangle;43although this surface is distorted from square, its four corners are equivalent under the symmetry0 operations of the lattice and the orthorhombic phase, like the tetragonal, has a single van Hove singularity.Electronic energy stabilisation due to the band Jahn-Teller effect observed in the A-15 compounds, moving the Fermi energy away from the van Hove singularity and reducing the density of states at E~,is also not responsible for the T-tO phase transition. Alternatively, the tilting distortion and its variation with 1.90 1.95 2.00 2.05 2.10 2.15 copper oxidation state may be explained by a non-electronicCu oxidation state mechanism, noting that the stability of the layer perovskites Fig. 5 Dependence of the orthorhombic distortion (a-b)/(a+b) on is determined by the matching of the intralayer distances. formal Cu oxidation state Then a tolerance factor4' for an A,B04 system may be 604 defined as t =dAo/21/2d~o (8) where dAOand dBO are the bond lengths in nine- and six-fold co-ordination, respectively.Use of the Shannon-Prewitt ionic radii for La2Cu04 gives t =0.86, indicating that the K2NiF4 structure will be subject to distortion and that pressure will be exerted on the Cu-0 bonds in the basal plane to shorten them below their ionic values.46 Using the method of the tolerance factor is given by t =$A/21/2fiB, where PB and are invariant values associated with nine- and six-co-ordinated A-0 and B-0 bond distances, respectively. The distances are related via fiB+21'2+A= O.996V1l3,where V is the volume of the unit cell. This method gives t =0.83. Thus, purely ionic considerations show that the La-0 distance in the La202 bilayers is too short relative to the in-plane Cu-0 distance for the undistorted structure to be stable. The effect of oxidation by Sr2+ doping is to increase the tolerance factor as the Cu-0,, bond lengths decrease and the (La,Sr) effective ionic radius increases.Thus, we may qualitatively understand the reduced magnitude of the tilting on doping and its eventual disappearance at x =0.248in terms of the tolerance factor increasing into the stability range of the Z4/mmm structure. The orthorhombic distortion due to the tilting produces better matching of the bond lengths in the two layers as buckling of the Cu02 layers leads to Cu-Cu distances below the sum of the two Cu-0 distances (Table 5). This leads to a reduction of the Cu-Cu antibonding interaction by dimin- ishing the overlap integral with the bridging oxygen atom.The (La,Sr) co-ordination polyhedron is in the shape of a square antiprism with one capped face and the (La,Sr) ions displaced off the centre of the antiprism. In the T phase, each ion occupies a site of symmetry C4", with four equidistant, oxygen neighbours in the rock-salt layer, four in the adjacent Cu02 layer and one axial oxygen of another layer. The effect of tilting is to increase the average (La,Sr)-0 distance (Table 6). Very importantly, the (La,Sr)-Oax bond length increases monotonically as the tilt angle increases. In the tetragonal phase, distances in the (La,Sr),O, planes are longer than the (La,Sr) distances to the oxygen atoms of the Cu02 planes; on tilting, however, one of the (La,Sr)-0 bonds in the lanthanide bilayer contracts significantly and its value is taken below the La,Sr distances to 0 atoms belonging to the CuO, planes.This is accompanied by the weakening of one of the (La,Sr)-0 bonds in the (La,Sr),O, layers. The net effect is that the pressure exerted on the basal plane decreases as the tilting angle increases and the bonding within the Cu02 layers is strongly affected by the relative disposition of the lanthanide ions within their co-ordination polyhedron. If we consider the combined effect of Sr2+ doping and tilting on the lanthanide co-ordination, we note that the La-Oax distance should increase smoothly (note the discrepancy at x =0.03)as the formal Cu oxidation state increases, but should decrease as the tilt angle is reduced upon doping.The first effect combined with the larger size of Sr2+ predominates. In general, upon doping the difference between long and short (La,Sr)-0 bonds decreases, tending towards the average value of the tetragonal phase, whereas the average (La,Sr)- 0 bond length decreases smoothly on doping. We now discuss the variation of the shape of the distorted CuO, octahedron with change in valence electron count given in terms of formal oxidation state, deduced from the observed chemical stoichiometry. Table 5 shows that the Cu-0 in-plane bond length contracts smoothly as a function of formal Cu oxidation state, indicating that, upon oxidation, electrons are removed from Cu-0 antibonding orbitals.A similar monotonic increase towards 180" is observed in the Cu- J. MATER. CHEM., 1991, VOL. 1 O,,-Cu angle, decreasing the tilting and producing improved overlap between the x2-y2 orbitals of the Cu atoms via the bridging oxygens. In contrast, variation of the Cu-0 axial bond length with x seems anomalous, with La, .97Sro.03Cu03,99 displaying an unexpectedly long bond. The site symmetry of the metal site for the I4/mmm space group is D4h.The resulting tetragonal crystal-field component is expected to produce some bond-length anisotropy. This is indeed observed, for example, in La2Ni0, (Ni" has a d* configuration and is not a Jahn-Teller ion), where (dax-deq)/ (dax+deq)=0.054 and the (c/a)=3.282.,'~,~As an empirical rule, it is observed that c/a ratios of the order of 3.3 are always found in the absence of Jahn-Teller distortion^.^' The formal Cu oxidation state in the series of compounds La2-xSr,Cu04-d is directly related to the number of anti- bonding electrons per Cu atom. Upon oxidation the contrac- tion of the Cu-0 in-plane bonds is larger than the contraction in the Cu-0 axial bonds, leading to a reduced (dax-deq)/(dax+deq) ratio.Then the anomaly in the Cu-0 axial bond lengths as a function of x can be rationalised by consideration of the Cu oxidation state and the Jahn-Teller effect. A co-operative Jahn-Teller effect in L~,CUO,.~, (and related compounds like LaSrMnO,) results in a large c/a ratio (3.451) compared to the Nil* analogues owing to the ferrodistortive coupling of the Jahn-Teller distortions at the Cu" centres.If we consider the dimensionless quantity (dax-deq)/(dax +deq) to represent the magnitude of the Jahn- Teller distortion (Fig. 6), this is clearly a maximum for a Cu oxidation state of +2.01 and decreases when the oxidation state changes from +2. We may apply the linear Jahn-Teller E-e problem" to the CuO, units in La2-,SrXCuO4-~; this results in the familiar 'Mexican hat'-type adiabatic potential for a two-fold degenerate electronic E term (dZ2, dX2-,,2) inter-acting with the two-fold degenerate vibrations (Qt, QJ. From the observed axial and equatorial bond lengths, we can estimate the radius po of the circle at the bottom of the trough.The depth from the degeneracy point of the two surfaces at p=O [p is the polar coordinate defined by (Q:+Q:)1/2] to the bottom of the trough at po is the measure of the Jahn-Teller stabilisation energy, EIT.As we see from Table 7, the maximum stabilisation energy [E,, = 19.12(5)hvY where v z500 cm-is the frequency of the Jahn-Teller active mode] occurs at x=O.O3, which corresponds to a formal Cu oxidation state of +2.01, with the number of Jahn-Teller active Cu" centres being at a maximum and co-operative ordering of the distortions of the CuO, octahedra occurring through the shared 0 atomss1 in an analogous fashion to three-dimensional perovskites. A mean-field treatment shows na"0.120+mo*121* 0.118! .. ' l . 1.90 1.95 2.00 2.05 2.10 2.15 Cu oxidation state Fig. 6 Dependence of the Jahn-Teller distortion on formal Cu oxi-dation state J. MATER. CHEM., 1991, VOL. I 605 Table 7 Jahn-Teller parameters in La, -xSrxCu04-6 0.0 0.395 l(5) 18.5 1 (5)0.01 0.3952(5) 18.52(4)0.03 0.4015(5) 19.12(5)0.06 0.3974(5) 18.73(5)0.13 0.39 19(5) 18.22(5) “~~=(4/3’”)[(d,,-d,,)/3];E,, =(1/2)pirnv2; rn is the mass of an oxygen atom and hv z 500 cm-* for the Jahn-Teller active mode. the order parameter in co-operative Jahn-Teller transitions is maximised when the number of contributing centres is largest,52i.e. at the oxidation state closest to 11. This implies that oxidation via Sr2+ doping is removing a* rather than n* antibonding electrons, as the a* electrons are responsible for the Jahn-Teller distortion.The possible complication with the above interpretation of the data is that superexchange as well as the conventional Jahn-Teller electron-phonon coupling mechanism can lift the degeneracy of the x2-y2 and z2 orbitals. Khomskii and K~gel~~have shown that antiferromagnetic superexchange favours ferrodistortive orbital ordering; this possibility must therefore be considered in view of the very large anti- ferromagnetic copper-copper superexchange interaction (Jc--cu z1000 K54,55)found in these systems. This mechanism would then allow n*,rather than a*, holes to frustrate the antiferromagnetic superexchange via the strong ferromagnetic potential exchange between the Cu x2-y2 orbitals and 0pn: holes in orthogonal orbitals [Fig.7(a)]. However, the distor- X 2-y PX x 2-y * cu 0 cu (b) tX 2-y x *-y 2PCI tion persists to higher doping levels where no magnetic long- range order is observed, and the maximum in the staggered moment at x=O.O does not correspond to that in the Jahn- Teller distortion. 4.2 Magnetic Ordering Despite the difficulties experienced in the analysis of the magnetic reflections because of their weakness, certain con- clusions may be tentatively arrived at. The ordered staggered moment is reduced on doping and disappears between a Cu oxidation state of +2.01 and +2.04. The absolute value of the moment estimated from our data is in reasonable agree- ment with the 0.48(15)pB found earlier in La2C~03,98,9 Fur-thermore, neutron measurements on single crystals have shown that local moments with correlation lengths inversely proportional to the hole-hole separation persist in the metallic and superconducting La2 -,Sr,Cu04 -The rapid reduction of the moment on doping may be rationalised as follows. Photoelectron spectroscopy and NMR measurements on oxidised samples in both the YB~,CU~O~-~ and the L~,-,S~,CUO~-~ systems imply that the states at cF may have up to 70% oxygen ~haracter.’~ The structural data of the previous section would then suggest that 2pa holes on the oxygen sublattice couple antiferromagnetically to a* holes on both neighbouring Cu2 ions, giving a net ferromagnetic + Cu-Cu interaction that will frustrate the Nee1 state [Fig.7(b)].The Cu-0 superexchange dominates the antifer- romagnetic Cu-Cu superexchange, as it is to second-order in the copper-oxygen transfer integral, and its energy has +-t If ++-Vdd It Fig. 7 Models for the reduction of Cu moment on doping. (a) n* hole producing ferromagnetic S= 3/2 unit and frustrating Cu-Cu antiferromagnetic superexchange; (b)frustration of the Nkel state by 0 r~* holes; (c)destruction of the Niel state by hole hopping within its localisation length; (d)antiferromagnetic next-nearest-neighbour coupling due to available Cu’”level been estimated as ca. 4000 K.'* This contrasts with the much more benign effects of non-magnetic diluents, such as Mg2+, below the percolation threshold (x, =0.59 on a two-dimen- sional square lattice59) on antiferromagnetic order in the prototype two-dimensional antiferromagnetic insulators Rb2MnF4 and Rb2CoF4.60 This model, incorporating the spectroscopic evidence of oxygen 2p character at E~ in the oxidised samples, can qualitatively account for the experimental observations, but in view of the large hole radii (more than twice the Cu-Cu distance, as deduced from the conductivity measurements), there are two other possibilities arising from dominant Cu character at E~,depending on whether hole motion is real or virtual.(i) Hopping of the hole within its localisation length will destroy the time-averaged moments within this radius as the spins are flipped at each passage of the hole through the site with a frequency thopmuch greater than Jt:-cu [Fig.7(c)]. Since ca.4% of holes lead to a 100% reduction in moment, the holes must spread to 25 Cu sites. The localisation length at the metal-insulator transition is ca. 15 A, corresponding to a radius of ca. 4 Cu-Cu bond lengths and containing ca. 40Cu sites. So the large hole radius makes such a model possible irrespective of the relative copper and oxygen contri- butions to the wavefunctions at E~.(ii) The availability of an empty x2-y2 level at low-spin d8 Cu"' states produces an antiferromagnetic next-nearest-neighbour interaction due to virtual hopping of two Cu" spins to the intervening Cu"' site which would frustrate the next-nearest-neighbour ferromag- netic coupling of the Nee1 state in La2Cu03.96 [Fig. 7(d)].This interaction is weaker than both Jcu--cuand Jcu--o but it disrupts a larger number of bonds. It may be estimated as 4t4/~~dUdd=4J~u-cu(t/Vdd)2,with Udd and Gd the intra- and inter-site repulsion parameters, respectively, and t the transfer integral; such long-range interactions could well be important in a system with very strong Cu-Cu coupling. We commented earlier that the (OIO)M magnetic peak does not shift with band filling within the resolution of our data. This observation may be used to test the predictions of itinerant models for the Cu" spin system. In such cases, the spontaneous magnetism in La2Cu04 would result from a spin-density wave produced by perfect nesting; doping would change the band filling and lead to an incommensurate spin- density wave.The incommensurate peak positions may then be generated by using a simple model, based on the nesting properties of a square two-dimensional Fermi surface.I2 Assuming that nesting is retained for small dopant concen- trations, the change in the nesting vector may be calculated from the requirement that the volume enclosed by the Fermi surface is equal to the number of electrons. This leads to the following expression for the dependence of the nesting vector Qon doping level x: Q={[I- x/21(4a), C1-x/21(4a)> (9) Hence if there is incommensurate spin-density wave behaviour, we expect a 0.3"shift from the commensurate position when x=O.O3. The fits indicate that this does not occur, favouring non-itinerant magnetic models.4.3 Conductivity The conductivity data show variable-range hopping (VRH) at low temperatures, indicating that for Cu oxidation states < +2.04, the states at E~ are localised rather than extended. The intrinsic gap for the creation of charged excitations in La2Cu04 is 2.0 f0.1 eV, as determined by photoconductivity experiments;61 carriers introduced by the dopants are lifted into the correlation gap between the CulI/I1' and the Cu*/" couples, to form impurity states. In the reduced samples, these J. MATER. CHEM., 1991, VOL. 1 are associated with oxygen vacancies and are pulled below the conduction band by the reduced ligand-field destabilis- ation of the e, electrons.The activation energy observed in La2Cu03,96 (Table 2) corresponds to the energy required for thermal ionisation of the bound electron into the upper Hubbard band. The x=O.Ol sample is compensated as the Sr2+ dopants empty the upper Hubbard impurity band, resulting in a reduced activation energy. The oxidised samples with formal Cu oxidation state greater than I1 have hole carriers associated with Sr2 dopant sites which are raised + above the lower Hubbard sub-band; compensation in this case occurs through the simultaneous presence of Sr2+ dopants and 0 vacancies. Owing to the limited temperature range over which VRH behaviour is observed, the determination of the exponent v is not totally unambiguous.Nonetheless, La2Cu03.96 can tenta- tively be said to have v = 1/2 due to the existence of a Coulomb gap in the density of states arising from the absence of compensation in the impurity band. The value of the To parameter [eq. (2)] derived from the VRH data allows evalu- ation of the localisation lengths through the expre~sion:~~ To=e2/4n~,~05,where E, is the average relative permittivity of the material, and taken to be z20, yielding 5=4 A. Thus 5 is of the order of the distance (R) between impurity states, given approximately by (Ni,;l3) z17 &where Nimpis the num- ber density of impurity ions), and impurity conduction (trans- port not involving wavefunctions associated with the +unperturbed Cu2 background) must be considered.The localisation length 5 may be further used to estimate the effective mass of the carriers through the expression62 (m*/ rne)=~/Erao,with a. the Bohr radius, yielding (rn*/rne)%2,in good agreement with that determined by measurements of the frequency-dependent relative permitti~ity.~~ This leads to an estimate of the dopant ionisation energy through the relationship Ei=(13.6/e2)(rn*/m,) of 0.06 eV. We have assumed that the carriers are bound to the impurity atoms in hydrogen- like orbits. La, .99Sro.01Cu03.96 shows either two- or three-dimensional VRH behaviour, while La1~97Sro~03Cu03~99and La, .94Sro.06C~03 display three-dimensional VRH behav- .99 iour. To now allows evaluation of the product t3N(&F),where N(E~)is the density of states at the Fermi energy through the relation35 kB TO =4qc/5 N(EF ) (10) with qc a dimensionless constant characteristic of the perco- lation network36 in three dimensions. Similarly, in two dimen- sions we have kBTO= 3/52N(EF) (1 1) However, to make further progress we need to evaluate This is done in two ways: (i) by assuming that the density of states is so large that is pinned and N(+) is constant across the series and equal to the value obtained by band-structure calculations (3 x lo2' states eV-' ~m-~);'~ and (ii) by calculat- ing N(E)within the space of singly occupied impurity states, assuming a two-dimensional tight-binding model for the Hub- bard sub-band in question: N(E)=[4N/z2(8t-E)]K[E/(~~-E)] (12) with K an elliptic integral of the first kind.64 The density of states in mid-band (~~4t) is then given approximately by N(E)=(Nim,/2n2t)In [16t/(~-4t)l (13) The Fermi energy is then evaluated from the condition N(E)dE =N(1-y)/2 J.MATER. CHEM., 1991, VOL. 1 where y is the fractional deviation from half filling the Hubbard sub-band, determined by the level of compensation. Evaluation of eqn. (14) at cF=4t in the limit (1 -u)% ulnu, where u= 1 -(+/4t) gives EF M4t(1-y) allowing estimation of cF and hence N(+) from eqn. (13), as a function of Cu oxidation state. The transfer integral is approximated to that arising from the overlap of two hydro- genic wavefunctions of the form exp (-(R)/l). The variation of localisation lengths, derived in the two approximations, with oxidation state is shown in Fig.8. We may have confidence in the values of the localisation lengths thus derived as the two models yield similar values. It is notable that as becomes greater than the interlayer distance 42, three-dimensional variable-range hopping occurs. The percolation pathway can now be completed within three dimensions rather than two, as the localisation length allows interlayer hops. The oxidation state of +2.04 corresponds to the disappear- ance of ordered magnetic moments in the o* orbitals and is in the vicinity of the transition to the metallic and supercon- ducting phase. It also corresponds to the radius of the impurity wavefunction becoming greater than the average interparticle separation.In this case, 5=2(R) (Table 8) and the uncorre- lated single-particle hopping model will break down; if the carriers are polaronic, they will exclude a carrier from hopping into a region within the localisation length of another. oo derived from the Arrhenius equation (Table2) is also much higher at this oxidation state. The localisation lengths may be used to calculate other parameters of interest associated with the carriers in the insulating region of the phase diagram such as the transfer 30 1J 25 -0 5. 20-0 C0,-05 15-0.-c u).-2 10-0. -e5-0 I'I'I'I0 ' Table 8 Parameters derived from the localisation lengths for the impurity bands in La, -.Sr,CuO, - ~~~~ ~ ~ ~ ~ Sr2+ concentration x 0.00 0.01 0.03 0.06 localisation length r/8, interdopant separation (R)/A effective mass (m*/m,) ionisation energy E,,cU/eV transfer integral t/eV" Hubbard U/eVb t/ u disorder broadening V/eV 4 17 2 0.06 0.07 0.12 0.6 0.07 6 17 1.8 0.05 0.11 0.06 1.8 0.07 14 15 0.9 0.03 0.07 0.03 2.3 0.06 20 12 0.6 0.02 0.05 0.02 2.5 0.08 " t =(3/2Me2/4n~o~,t)Cl +(1/6X(R)/5)fl exp (--(R)/5); !=+((W/O(5/8)(e2/4nq,&,<);V =e2/4n~O~,( where E, is the average relative R)permittivity of the material. integral t, the Hubbard repulsion parameter U and the disorder energy V.65 These are collected in Table 8.We now discuss these in view of the possible impurity transport mechanisms. Polaron formation is generally of importance in narrow band materials and in this case where the carriers are localised, a local lattice distortion is expected.However, in all cases except La2Cu03.96, the localisation lengths are larger than the Cu-Cu distances and the small polaron model is not strictly valid; the polarisation clouds will overlap to reduce the hopping energy significantly (intermediate polaron) or produce large effective mass band transport (large polaron). Two important parameters are the polaronic transfer integral tpoland the Holstein coupling constant g. For small polarons, the transfer integral is given by tpo,=t exp (-E,/hcuo), where E, is the polaron binding energy and oois an optical phonon frequency; using E, =0.2 eV5' and oo=0.075 eV,66 we obtain t,,,=0.01 eV, of the same order as oo with the Holstein coupling constant estimated as 3, much lower than the lower limit for adiabatic behaviour to be valid.67 Hence the anti- adiabatic model is expected to be more appropriate, as found experimentally.Thus although the inclusion of a temperature- dependent mobility term makes the Arrhenius plots mono- tonic and allows an approximate one-function description down to the temperature range at which VRH is important, its interpretation as small polaron hopping is inconsistent with the carrier radii and the activation energies. Thus while we expect localised narrow band carriers to have some polaronic character, there is no quantitative evidence for a hopping energy. In narrow bands with kBTxW, the mobility is expected to have a complex temperature dependen~e.~~ In La2Cu04+ Preyer et aL6*find that the in-plane Hall mobility pHxl cm2 V-' s-' is temperature independent, whereas Che~ng~~et al.prefer a diffusive mobility term pz(eD/kBT) to explain the high-temperature behaviour of Pb-doped La,Cu04. The Arrhenius energies (Table 2), though their exact values depend on the existence or not of a temperature-dependent mobility, may be discussed within the framework of doped, compensated semiconductors. The carrier-dopant ionisation energies Ei in Table 8 are calculated using ~,=20 and hence are upper bounds as the relative permittivity is expected to increase with increasing carrier concentration. The Eivalues are too high to explain the observed activation energies and impurity-band transport competes with valence/conduction band transport.The Hubbard splitting of the impurity band is opposed both by the bandwidth W=2zt and by the spread in site energies, governed by the disorder which becomes more important for heavy carriers. Even though the vaues for t and U (Table 8) are only approximate, it can clearly be seen that t/U increases as the formal oxidation state increases. The disorder potential may then be sufficient to cause the reduced Hubbard gap to disappear. If transport via doubly occupied impurity sites in the upper Hubbard sub-band was operative c2 = U -W would decrease rapidly with concentration, with nobeing only weakly dependent on carrier density.Both these criteria are fulfilled for the low oxidation states (< +2.01) and such a mechanism appears a good candidate for transport in this regime. Hopping between localised states on opposite sides of cF may occur if the disorder potential causes the two sub-bands to overlap. The condition (R) % 5 allows use of the Miller-Abrahams formula for one-phonon nearest-neigh- bour hopping for x=O and 0.01,70 giving EMIA=0.07and 0.02 eV, respectively. Given that overlap occurs between the two Hubbard impurity bands, or the upper Hubbard band with the edge of the conduction or valence bands, Anderson localisation of the states in the band tails will occur, and excitation to the mobility edge at cC (Fig. 9) then competes with near-neighbour hopping: o=ooexp [-(E~-cF)/kBq, Udd cun/cu= cu'/cu* E Impurity band cuu/cu= -2 N (E) Fig.9 Impurity band-structure (a)at low carrier concentrations and (b)close to the metal-insulator transition in La, -xSrxCu04-d where a,(= 0.03e2/ha)is the minimum metallic conductivity, assuming the inelastic scattering length is equal to a, the interdopant spacing, at higher temperatures. Taking a = 15 A gives a. =50 R-' cm-' which is of the correct order of magnitude and would explain the similar values of a, for oxidation states -= +2.04. The mobility edge mechanism can crossover directly to VRH at low temperature without an intermediate lower activation energy regime due to hopping in localised states in the tail of the valence band if N(+) is high, as is the case here.35 Considering the magnitude and variation of the 0, prefactor, excitation to the mobility edge in the upper Hubbard band/valence band at high temperatures with a crossover to VRH between localised states at E~ at low temperatures seems preferable.The Sr2 lower Hubbard band + will be narrow as the Sr2+ dopants bind one charge but the separation of the upper Hubbard band from the valence band will be much reduced as the binding of an extra electron to Sr2+ will be greatly reduced [Fig. lO(a)]. It is therefore J. MATER. CHEM., 1991, VOL. 1 possible that the upper Hubbard band will overlap the edge of the valence band and will merge into it at higher dopant concentrations [Fig.1 O(b)].The metal-insulator transition in the Cu oxidation range +2.04 to +2.05 implies that the Hubbard splitting between the impurity bands disappears in this concentration range, with the transition being of the Anderson type, i.e. non-zero N(E~)with the wavefunctions of the carriers changing from localised to extended as crosses E,. The sharp increase of a. at +2.04 indicates the system is close to delocalisation. 5. Conclusions We have presented structural, magnetic and transport measurements on La, -xSrxCu04-samples of well defined formal Cu oxidation state. The structural evolution in the L~,-,S~,CUO~-~series as the formal Cu oxidation state increases has been investigated. The orthorhombic space group Aha, resulting from alternate rigid tilting of the Cu06 octahedra along the [llO] direction of the tetragonal unit cell, describes the orthorhombic phase throughout the compo- sition and temperature range investigated.No evidence of charge-density wave formation and Fermi surface instabilities is found and there is no splitting of the van Hove singularity in the density of states in the metallic samples. The mechanism of the tetragonal-to-orthorhombic phase transition was rationalised in terms of tolerance-factor arguments, based on the mismatch in size between the lanthanide bilayers and the CuO, sheets; doping by the larger and more basic Sr2 + results in the tolerance factor increasing towards the stability range of the tetragonal phase, through the relief of pressure exerted on the Cu02 basal plane.Closer examination of the distortion of individual CuO, units reveals that the Cu-0 bond lengths do not change monotonically with increasing oxidation state. The (dax-deq)/(dax+deq)ratio peaks at a formal Cu oxidation state of +2.01; this was interpreted in terms of the co-operative Jahn-Teller effect as implying that the holes pro- duced by oxidation occupy a* orbitals rather than n*. The magnetic neutron diffraction measurements show no incommensurate behaviour and weigh against the interpret- ation of the long-range magnetic order observed in the system as a spin-density wave instability of the square two-dimen- sional Fermi surface, suggesting a Mott-Hubbard model. The magnetic long-range order disappears at a Cu oxidation state of +2.04, i.e.before the metal-insulator transition and the onset of superconductivity. The rapid destruction of the N6el state on doping is attributed to frustration. Location of the holes in the pa oxygen orbitals will lead to ferromagnetic Cu-Cu nearest-neighbour interactions (one bond disrupted per hole) that will frustrate the antiferromagnetic order. Alternatively, the availability of empty x2-y2 orbitals in low- spin Cu"' could similarly lead to frustration through the introduction of antiferromagnetic Cu-Cu next-nearest-neighbour interactions (four bonds disrupted per hole). The magnetic behaviour on hole doping is in contrast to the benign effect of Cu' 'electron doping' on magnetic long-range order in Nd, -,CeXCuO4 Finally, the nature of the carriers associated both with oxygen vacancies and with Sr2+ dopants, as the metal- insulator transition is approached was clarified.Doping intro- duces impurity states into the large intrinsic Mott-Hubbard gap, in analogy with classical doped semiconductors. The conductivity measurements reveal variable-range hopping behaviour at low temperatures, showing that the states near E~ are localised in the samples even after the ordered moments have disappeared. The data allowed the evaluation of the localisation lengths 5 of the carriers as a function of formal Cu oxidation state. They increase monotonically upon oxi- J. MATER. CHEM., 1991, VOL. 1 609 dation, and at a composition close to the metal-insulator 24 J.D.Axe, A.H. Moudden, D. Hohlwein, D. E. Cox, K. M. transition they become larger than the mean inter-donor separation. The transport mechanism at higher temperatures is more difficult to elucidate; however, close to the metal- insulator transition, excitation to a mobility edge in the upper impurity Hubbard sub-band seems reasonable. Variation of go close to the transition indicates that it is of Anderson type, with E~ crossing the mobility edge E, which lies in the overlapping upper Hubbard impurity band and valence band. The importance of narrow impurity-band phenomena at higher oxidation states where the onset of superconductivity 25 26 27 28 Mohanty, A. R. Moodenbaugh and Y. Xu, Phys. Rev. Lett., 1989, 62, 2751.M. K. Crawford, W.E. Farneth, E. M. McCarron 111, R. L. Harlow and A. H. Moudden, Science, 1990, 250, 1390. R. J. Cava, R. B. van Dover, B. Batlogg and E. A. Rietmann, Phys. Rev. Lett., 1987, 58, 408. H. H. Wang, K. D. Carlson, U. Geiser, R. J. Thorn, H. C. I. Kao, M. A. Beno, M. R. Monaghan, T. J. Allen, R. B. Proksch, D. L. Stupka, J. M. Williams, B. K. Flandermeyer and R. B. Poeppel, Inorg. Chem., 1987, 26, 1474. M. J. Rosseinsky, K. Prassides and P. Day, Physica C, 1989, 161, 21. occurs merits serious attention since impurity-band effects 29 A. K. Cheetham and A. J. Skarnulis, J. Anal. Chem., 1981, 53, could well still be operative with the high density of states possibly allowing unconventional physics. 30 1060. A. I. Nazzal, V. Y. Lee, E. M. Engler, R.D. Jacowitz, Y. Tokura and J. B. Torrance, Physica C, 1988, 153-155, 1367. We thank SERC for financial support, the Institut Laue Langevin and the Rutherford Appleton Laboratory for pro- vision of neutron time, A. w. Hewat, J. K. Cockcroft, w. I. F. 31 32 33 34 A. M. Chippindale, personal communication. R. J. D. Tilley and A. H. Davies, Nature (London), 1987,326, 859. I. G. Austin and N. F. Mott, Adv. Phys., 1969, 18, 41. M. A. Kastner, R. J. Birgeneau, C. Y. Chen, Y. M. Chiang, D. R. David and R. M. Ibberson for help with the experiments and the data analysis and S. T. Bramwell, M. Kurmoo and J. M. F. Gunn for helpful discussions. M.J.R. thanks Merton College, Oxford for a Junior Research Fellowship. 35 Gabbe, H. P. Jennsen, T. Junk, C. J. Peters, P.J. Picone, T. Thio, T. R. Thurston and H. L. Tuller, Phys. Rev. B, 1988, 37, 111; R. J. Birgeneau, C. Y. Chen, R. D. Gabbe, H. P. Jenssen, M. A. Kastner, C. J. Peters, P. J. Picone, T. Thio, T. R. Thurston and H. L. Tuller, Phys. Rev. Lett., 1987, 59, 1329. N. F. Mott and E. A. Davis, Electronic Processes in Non-Crystal- References 36 line Materials, Clarendon Press, Oxford, 1978. V. Ambegaokar, B. I. Halperin and J. S. Langer, Phys. Rev. B, 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Mixed Valence Compounds in Chemistry, Physics, Biology and Geology, ed. D. B. Brown, Reidel, New York, 1980; Mixed Valency Compounds: Applications in Chemistry, Physics and Biology, ed. K. Prassides, Kluwer, Dordrecht, 199 I. D. C. Johnston, H. Prakash, W.H. Zachariasen and R. Viswantham, Muter. Res. Bull., 1973, 8, 777. C. W. Chu, S. Huang and A. W. Sleight, Solid State Commun., 1976, 18, 977. R. J. Cava, B. Batlogg, J. J. Krajewski, R. Farrow, L. W. Rupp, A. E. White, K. Short, W. F. Peck and T. Kometani, Nature (London),1988,332, 814. J. G. Bednorz and K. A. Miiller, 2.Phys. B, 1986, 64, 189. R. M. Hazen, L. W. Finger, R. J. Angel, C. T. Prewitt, N. L. Ross, C. G. Hadidiacos, P. J. Heaney, D. R. Veblen, Z. Z. Sheng, A. El Ali and A.M. Herman, Phys. Rev. Lett., 1988, 60, 1657. Y. Tokura, H. Takagi and S. Uchida, Nature (London), 1989,337, 345. H. Sawa, S. Suzuki, M. Watanabe, J. Akimitsu, K. Kolcirlo, H. Asaio, F. Izuki and E. Takayama-Muromachi, Nature (London), 1989, 337, 347. D. Vaknin, S. K. Sinha, D.E. Moncton, D. C. Johnston, J. Newsam, C. R. Safinya and H. King, Phys. Rev. Lett., 1987, 58, 2802. P. Day, M. J. Rosseinsky, K. Prassides, W. I. F. David, 0.Moze and A. S. Soper, J. Phys. C, 1987,20, L429. J. D. Jorgensen, H. B. Schuttler, D. G. Hinks, D. W. Capone 111, K. Zhang, M. B. Brodsky and D. J. Scalapino, Phys. Rev. Lett., 1987, 58, 1024. L. F. Mattheiss, Phys. Rev. Lett., 1987, 58, 1028. P. W. Anderson, Science, 1987, 235, 1196. P. G. de Gennes, C.R. Acad. Sci. Paris, 1987, 30511, 345. M. Cyrot, Solid State Commun., 1987, 62, 821. J. M. Delgado, R. K. McMullan, G. D. de Delgado, B. J. Wuensch, P. J. Picone, H. P. Jennsen and D. R. Gabbe, Phys. Rev. B, 1988, 37, 9343. D. McK. Paul, G. Balakrishan, N. R. Bernhoeft, W. I. F. David and W. T. A. Harrison, Phys.Rev. Lett., 1987, 58, 1976. H. J. Kim and R. Moret, Physica C, 1988, 156, 363. R. J. Cava, A. Santoro, D. W. Johnson Jr. and W. W. Rhodes, Phys. Rev. B, 1987, 35, 6716. M. Onoda, S. Shamoto, M. Sat0 and S. Hosoya, Jpn. J. Appl. Phys., 1987, 26, L363. T. Kajitani, T. Onozuka, Y. Yamaguchi, M. Hirabayashi and Y. Syono, Jpn. J. Appl. Phys., 1987, 26, L1877. R. V. Kasowski, W. Y. Hsu and F. Herman, Solid State Commun., 1987, 63, 1077. H. H. Wang, U. Geiser, R. J. Thorn, K. D. Carlson, M. A. Beno, M. R. Monaghan, T. J. Allen, R. B. Proksch, D. L. Stupka, W. K. Kwok, G. W. Crabtree and J. M. Williams, Inorg. Chem., 1987, 26, 1190. 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 1971, 4, 2612. B. I. Shklovskii and A. L. Efros, Electronic Properties of Doped Semiconductors, Springer Verlag, Berlin, 1984.A. W. Hewat, personal communication. H. M. Rietveld, J. Appl. Crystallogr., 1969, 2, 65. J. K. Cockcroft, personal communication. W. C. Hamilton, Acta Crystallogr., 1965, 18, 502. T. Freltoft, G. Shirane, S. Mitsuda, J. P. Remeika and A. S. Cooper, Phys. Rev. B, 1988, 37, 137. J. P. Pouget, G. Noguera and R. Moret, J. Phys. (Paris), 1988, 49, 375. S. Barisic, I. Batistic and J. Friedel, Europhys. Lett., 1987, 3, 1231. P. Ganguly and C. N. R. Rao, J. Solid State Chem., 1984, 53, 193. K. K. Singh, P. Ganguly and J. B. Goodenough, J. Solid State Chem., 1984, 52, 254. P. Poix, J. Solid State Chem., 1980, 31, 95. R. M. Fleming, B. Batlogg, R. J. Cava and E. A. Rietman, Phys. Rev. B, 1987, 35, 5337.A. Rabeneau and P. Eckerlin, Acta Crystallogr., 1958, 11, 304. I. B. Bersuker, The Jahn-Teller Effect and Vibronic Interactions in Modern Chemistry, Plenum Press, New York, 1984. J. Kanamori, J. Appl. Phys., 1960, 31, 14s. R. Englman and B. I. Halperin, Phys. Rev. B, 1970, 2, 75. D. L. Khomskii and K. I. Kugel, Solid State Commun., 1973, 13, 763. K. P. Lyons, P. A. Fleury, J. P. Remeika, A. S. Cooper and T. J. Negran, Phys. Rev. B, 1988, 37, 2393. L. J. de Jongh, Eur. J. Solid State Inorg. Chem., 1990, 27, 221. Y. Endoh, K. Yamada, R. J. Birgeneau, D. R. Gabbe, H. P. Jenssen, M. A. Kastner, C. J. Peters, P. J. Picone, T. R. Thurston, J. M. Tranquada, G. Shirane, Y. Hidaka, M. Oda, Y. Enomoto, M. Suzuki and T. Murakami, Phys. Rev. B, 1988, 37, 7443. M. Takigawa, P. C. Hammel, R. H. Heffner, Z. Fisk, K. C. Ott and J. D. Thompson, Physica C, 1989, 162,285. A. Aharony, R. J. Birgeneau, A. Coniglio, M. A. Kastner and H. E. Stanley, Phys. Rev. Lett., 1988, 60,1330. W. Kang, G. Collin, M. Riboult, J. Friedel, D. Jerome, J. M. Bas-sat, J. P. Coutures and P. Odier, J. Phys. (Paris), 1987, 48, 1181. R. J. Birgeneau, R. A. Cowley, G. Shirane, J. A. Tarvin and H. J. Guggenheim, Phys. Rev. B, 1980, 21, 317. T. Thio, R. J. Birgeneau, A. Cassanho and M. A. Kastner, Phys. Rev. B, 1990, 42, 10800. N. W. Ashcroft and N. D. Mermin, Solid State Physics, Holt-Saunders, New York, 1975. C. Y. Chen, R. J. Birgeneau, D. R. Gabbe, H. P. Jenssen, M. A. Kastner, P. J. Picone, N. W. Preyer and T. Thio, Physica C, 1989, 162, 1031. R. Combescot and J. Labbe, Phys. Rev. B, 1988,38, 262. 610 J. MATER. CHEM., 1991, VOL. 1 65 66 67 68 N. F. Mott, Adu. Phys., 1972, 21, 785. M. J. Rosseinsky, K. Prassides, P. Day and A. J. Dianoux, Phys. Rev. B, 1988, 37, 2231. A. J. Bosman and H. J. van Daal, Adu. Phys., 1970, 19, 1. N. W. Preyer, R. J. Birgeneau, C. Y.Chen, D. R. Gabbe, H. P. Jenssen, M. A. Kastner, P. J. Picone and T. Thio, Phys. Rev. B, 69 70 71 S. W. Cheong, Z. Fisk, R. S. Kwok, J. P. Remeika, J. D. Thomp- son, G. Gruner, Phys. Reu. B., 1988,37, 5916. A. Miller and E. Abraham, Phys. Rev., 1960,120, 745. M. J. Rosseinsky, K. Prassides and P. Day, Znorg. Chem., 1991, 30. 1989,39, 11563. Paper 1/0061OJ; Received 11th February, 199 1

ISSN:0959-9428    

DOI:10.1039/JM9910100597

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

J. MATER. CHEM., 1991, 1(4), 611-619 61 1 Molecular Engineering of Liquid-crystalline Polymers by Living Polymerization. Part 13t.-Synthesis and Living Cationic Polymerization of (S)-( -)-2-Methylbutyl 4’4a-Vinyloxy)-alkoxybiphenyl-4-carboxylate with Undecanyl and Hexyl Alkyl Groups Virgil Percec,* Qiang Zheng and Myongsoo Lee Department of Macromolecular Science, Case Western Reserve University, Cleveland, OH 44 106, USA The synthesis and living cationic polymerization of (S)-(-)-2-methylbutyl 4-(I I-vinyloxy)undecyloxybiphenyl-4-carboxylate (13-11) and (S)-(-)-2-methylbutyl 4’-(6-vinyloxy)hexyloxybiphenyl-4-carboxylate (13-6) are described. Polymers with degrees of polymerization from 4 to 26 and polydispersities G1.10 were synthesized and characterized by differential scanning calorimetry (DSC) and thermal optical polarized microscopy.When determined from first heating and cooling DSC scans, poly(l3-11) exhibits monotropic s,, si and s, (unidentified) mesophases over the entire range of molecular weights. When determined from second and subsequent heating and first and subsequent cooling scans, poly(l3-11) with degrees of polymerization <10 exhibit enantiotropic s,, st and s, mesophases, and crystallization on the heating scan, while those with higher degrees of polymerization exhibit enantiotropic s,, si and s, mesophases only. Regardless of the thermal history of the sample, poly(l3-6) exhibits an enantiotropic s, phase, and the polymers with degrees of polymerization > 12 exhibit an enantiotropic s, mesophase as well.Keywords: Living cationic polymerization; Chirality; Smectic C phase; Vinyl ether Liquid-crystalline polymers exhibiting chiral mesophases, i.e. Experimental cholesteric’ and chiral smectic C (s,*),~are of both theoretical Materialsand technological interest. Liquid crystals exhibiting chiral smectic A (s:) mesophases were only recently discovered3 and 4-Hydroxybiphenyl (97%), dimethylsulphate (99Oh +), HBr to our knowledge, polymers exhibiting s: mesophases have (48% in H20), 9-borabicyclo[3.3.1]nonane (9-BBN dimer, not yet been reported. Side-chain liquid-crystalline polymers crystalline, 98%), 1 1-bromoundecan- 1-01 (98%), butyl vinyl exhibiting s: mesophases were reported from several different ether (98%), tetrabutylammonium hydrogen sulphate (TBAH) However, there is very little understanding (all from Aldrich), 1,lo-phenanthroline (anhydrous, 99”/0), laboratorie~.~,~-’~ of the molecular design of side-chain liquid-crystalline poly- palladium(I1) diacetate (both from Lancaster Synthesis), acetyl mers displaying s,*mesophases, and of the influence of various chloride (990/) and S-(-)-2-methylbutan-l-o1(95%) (both architectural parameters of these polymers on their dynam- from Fluka) were used as received.Methylene chloride (Fisher) The ability to synthesize side-chain liquid-crystalline was purified by being washed with concentrated sulphuric ic~.~*~-” polymers by a living polymerization technique would provide acid several times until the acid layer remained colourless, a very useful preparative tool which gives access to polymers then with water, dried (MgS04), refluxed over calcium hydride with narrow molecular weight distribution and controllable and freshly distilled under nitrogen before each use.Dimethyl molecular weight. sulphide (anhydrous, 99%, Aldich) was refluxed over 9-BBN Group-transfer polymerization of mesogenic methacrylates and then distilled under nitrogen. Trifluoromethanesulphonic can be used to prepare side-chain liquid-crystalline polymers acid (triflic acid, 98%, Aldrich) was distilled under vacuum. with narrow molecular weight distribution and controllable molecular weights. “3’’ However, this polymerization does Techniquesnot tolerate functional groups sensitive to the nucleophilic growing species and can be used mostly for the polymerization ‘H NMR (200 MHz) spectra were recorded on a Varian XL- of methacrylates.Cationic polymerization can be used to 200 spectrometer. Infrared (IR) spectra were recorded on a polymerize, under living conditions, mesogenic vinyl ethers Perkin-Elmer 1320 infrared spectrophotometer. Relative mol- containing a large variety of functional especially ecular weights of polymers were measured by gel permeation with the initiating system CF3S03H-S(CH3)2.20 chromatography (GPC) with a Perkin-Elmer Series 10 LC The goal of this paper is to describe the synthesis and living instrument equipped with LC-100 column oven and a Nelson cationic polymerization of (S)-(-)-2-methylbutyl 4-(1 1 -vinyl- Analytical 900 series integrator data station.A set of two oxy)undecyloxylbiphenyl-4-carboxylate(13-11)and (S)-(-)-2- Perkin-Elmer PL gel columns of 5x102 and 104A with methylbutyl 4’-(6-vinyloxy)hexyloxybiphenyl-4-carboxylate CHC1, as solvent (1 cm3 min- I) were used. The measurements (13-6). The mesomorphic behaviour of poly(l3-11) and were made at 40 “C using the UV detector. Polystyrene poly(13-6) will be discussed as a function of molecular weight. standards were used for the calibration plot. High-perform- ance liquid chromatography (HPLC) experiments were per- formed with the same instrument. Absolute number average molecular weights were determined by ‘H NMR spectroscopy t Part 12. V. Percec, C. S. Wang and M. Lee, Polym.Bull., in the by analysing the chain ends of the resulting polymers. A press. Perkin-Elmer DSC-4 differential scanning calorimeter 612 equipped with a TADS data station was used to determine the thermal transitions, which were reported as the maxima and minima of their endothermic and exothermic peaks. In all cases, heating and cooling rates were 20 "C min-' unless specified. Glass-transition temperatures (Tg) were read at the middle of the change in the heat capacity. First heating scans differed from second and subsequent heating scans. However, second and subsequent heating scans were identical. A Carl- Zeiss optical polarized microscope (magnification 100x) equipped with a Mettler FP 82 hot stage and a Mettler FP 800 central processor was used to observe the thermal tran- sition and to analyse the anisotropic textures.Synthesis of Monomers Monomers were synthesized as outlined in Scheme 1. 4-Methoxybiphenyl (2) 4-Hydroxybiphenyl (127.8 g, 0.75 mol) was dissolved in aque- ous sodium hydroxide (2.250 dm3, 1.5 mol dm-3) at 55 "C. Dimethyl sulphate (189 g, 1.5 mol) was added slowly so that the temperature did not exceed 60 "C. The temperature was then raised to 70 "C for 30 min and the reaction mixture was allowed to cool to room temperature. The resulting white 1 2 2 - Br(CH2),-OH + n-C,HgCH=CHz Cl?ICI CHp=CH -O(CH,)@r 7 -n 8 9 -n (n=6.11) 4 CHJCOZH 13-11 Scheme 1 Synthesis of (S)-(-)-2-methylbutyl 4'-(1 1 -vinyloxy)undecyl-oxybiphenyl-4-carboxylate (13-11) and (S)-(-)-2-methylbutyl 4-(6-vinyloxy)hexyloxybiphenyl-4-carboxylate(13-6) J.MATER. CHEM., 1991, VOL. 1 precipitate was filtered and recrystallized from 95% ethanol to yield white crystals (68.5 g, 50%); purity, 99.8% (HPLC); m.p., 87.0-87.5 "C (lit.,21" 89.0 "C; lit.216 80.5 "C); dH (CDCl,, TMS) 3.81 (s, 3 H, OCH3), 6.92 (d, 2 H, 3-H, 5-H), 7.22 (d, 2 H, 2-H, 6-H), 7.35 (m, 2 H, 2'-H, 6'-H), 7.47 (m, 3 H, 3-H, 4-H, 5-H). 4-Acetyl-4-Methoxybiphenyl (3) 4-Methoxybiphenyl (2) (56 g, 0.303 mol) was dissolved in 500 cm3 dry methylene chloride in a 2000 cm3 three-necked round-bottom flask, equipped with dropping funnel and con- denser. The solution was cooled to 0-2 "C. Anhydrous alu- minium chloride (48.4 g, 0.36 mol) was added quickly to give a green solution.Acetyl chloride (28.5g, 0.363 mol) was then added over 20-30 min, and the reaction mixture was refluxed for 2 h. Ice-cooled, concentrated HC1 (300 cm3) was added to the cooled mixture to decompose the yellow-green complex. Then water (200 cm3) was added and the mixture was heated to remove most of the methylene chloride. A light yellow- brown solid separated from the reaction mixture. The resulting solid was washed twice with diethyl ether (200 cm3 each time) to remove the isomeric 3-acetyl-4'-methoxybiphenyl, which is soluble in diethyl ether. The solid was then recrystallized from isopropyl alcohol (24 cm3 g-') to yield colourless flakes (64 g, 93%); purity, 99.5% (HPLC); m.p., 153-154 "C (lit.,21b 153-154 "C); 8H (CDCl,, TMS) 2.61 (s, 3 H, COCH,), 3.88 (s, 3 H, OCH3), 7.00 (m, 2 H, 3'-H, 5'-H), 7.58 (d, 2 H, 2'-H, 6'-H), 7.64 (d, 2 H, 2-H, 6-H), 7.98 (d, 2 H, 3-H, 5-H).4'-Methoxybiphenyl-4-carboxyIicAcid (4) A sodium hypobromite solution, prepared at 0-5 "Cby adding bromine (34cm3, 0.659mol) very slowly into a solution of sodium hydroxide (95.2 g, 2.38 mol) in water (450 cm3), was added slowly to a solution of 4-acetyl-4'-methoxybiphenyl (3) (34 g, 0.15 mol) in 1100 cm3 of 1,4-dioxane over 1 h. The temperature of the reaction mixture was allowed to rise to 35-40 "C. After being stirred for an additional 15 min, the sparingly soluble sodium salt solution was treated with enough sodium hydrogen sulphite (52.5 g, 0.505 mol) to destroy the excess of hypobromite.The hot solution was then acidified to yield 4'-methoxybiphenyl-4-carboxylicacid (29 g, 84.5%) of sufficient purity to be used in the reduction step; m.p., 253- 254 "C (lit.,22 258 "C); aH (C2H6]acetone, TMS) 3.86 (s, 3 H, OCH,), 7.05 (m, 2 H, 3'-H, 5'-H), 7.68 (d, 2 H, 2'-H, 6'-H), 7.75 (d, 2 H, 2-H, 6-H), 8.08 (d, 2 H, 3-H, 5-H). 4'-Hydroxybiphenyl-4-carboxylicAcid (10) Compound 4 (29 g, 0.127 mol) was dissolved in 1160 cm3 of boiling acetic acid. A solution of 48% hydrobromic acid (230 cm3, 2.03 mol) was added and the reaction mixture was heated at reflux temperature for 12-13 h. The reaction mixture was then poured into water (3 dm3), and allowed to cool to room temperature. The resulting precipitate was separated, washed with water and dried to yield the crude solid product (27 g, 99%0).The product was purified by dissolving in meth- anol, and treating the solution with activated carbon to give a colourless solution, which was poured into water to yield a white powder (25 g, 91%); m.p., 293-294 "C (lit.,23 293- 294 "C); dH (['H,]acetone, TMS) 3.03 (s, 1 H, OH), 6.98 (d, 2 H, 3'-H, 5'-H), 7.58 (d, 2 H, 2'-H, 6'-H), 7.71 (d, 2 H, 2-H, 6-H), 8.07 (d, 2 H, 3-H, 5-H); v,,, (Nujol, KBr plate)/cm-' 3300-2400 (OH of COOH), 1670 (C=O). S-(-)-2-Methylbutyl Toluene-p-sulphonate (6) S-(-)-2-Methylbutan-l-o1 (44g, 0.5 mol) was added slowly to a solution of toluene-p-sulphonyl chloride (190.6 g, 1 mol) in dry pyridine (300 cm3) at 0 0C.24The reaction mixture was stirred overnight at room temperature.The resulting solution J. MATER. CHEM., 1991, VOL. 1 was poured into water (300 cm3) and extracted with diethyl ether. The ether layer was dried (MgS04) and the diethyl ether was removed on a rotary evaporator to yield a colourless oil (120 g, 99"/), purity, 99% (HPLC); dH (CDC13, TMS) 0.85 [m, 6 H, CH(CH3)CH2CH3], 1.02- 1.39 (m, 2 H, CHCH2CH3), 1.61 [m, 1 H, CH2CH(CH3)CH2], 2.40 (s, 3 H, ArCH3), 3.83 (m, 2 H OCH,), 7.22 (d, 2 H, 3-H, 5-H), 7.71 (d, 2 H, 2-H, 6-H). 1,lO-Phenanthrolinepalladium(II) Diacetate To a stirred solution of palladium(I1) diacetate (1.783 g, 7.94 mmol) in dry benzene (60 cm3) was added a solution of 1,lO-phenanthroline (1.5 g, 8.32 mmol) in benzene (70 cm3) over 30min.The mixture was stirred for 4h and the yellow precipitate was filtered and washed with benzene and light petroleum to yield a yellow solid (3.2g, 99"/), m.p., 233- 234 "C (lit.,25, 234 "C); 8H (CDC13, TMS) 2.16 [s, 6 H, OCO(CH,)J, 7.77 (m, 2 H, 4-H, 7-H), 7.99 (s, 2 H, 5-H, 6-H), 8.56-8.67 (m, 4 H, 2-H, 9-H). 11-Brornoundecyl Vinyl Ether (9-11)19' A solution of 11-bromoundecan-1-01 (8 g, 0.23 mol), 1,lO- phenanthrolinepalladium(1I) diacetate (0.41 8 g, 1.03 mmol), dry chloroform (20cm3) and butyl vinyl ether (85cm3) was heated at reflux overnight (12-14 h). The resulting light-yellow solution, obtained after gravity filtration, was distilled on a rotary evaporator to remove the excess butyl vinyl ether and chloroform.The remaining yellow oil was purified by column chromatography (silica gel, CH2C12 as eluent) to yield a light- yellow oil (8.1 g, 92%); 6, (CDC13, TMS) 1.27-1.81 (m, 18 H, OCH2[CH2I9), 3.38 (t, 2 H, CH2Br), 3.64 (t, 2 H, =CHOCH2-), 3.96 (d, 1 H, =CH2, trans), 4.14 (d, 1 H, =CH2, cis), 6.39-9.51 (m, 1 H,=CH-0-). 6-Brornohexyl Vinyl Ether (9-6) Compound 9-6 was synthesized by the same procedure as the one used for in the synthesis of 9-11.6-Bromohexan-l-o1(12g, 66.3 mmol) and 1,lo-phenanthrolinepalladium(rr) diacetate (0.872 g, 2.16 mmol), dry chloroform (20 cm3) and butyl vinyl ether (177 cm3) were heated at reflux overnight. The resulting product was purified to produce a light-yellow oil (12.2g, 89Yo). 6H (CDCl3, TMS) 1.42-1.84 (m, 8 H, OCH2-[CH2]4), 3.38 (t, 2 H, CH,Br), 3.65 (t, 2 H, =CHOCH2), 3.93 (d, 1 H, =CH2, trans), 4.14 (d, 1 H, =CH2, cis), 6.38-9.49 (m, 1 H,=CHO-).Potassium 4'-Hydroxybiphenyl-4-carboxylate (1 1) Compound 10 (20 g, 0.293 mol) was dissolved in methanol (500 cm3). The solution was titrated with a solution of KOH (1 mol dm-3) in CH30H using phenolphthalein as indicator. The solution was then poured into diethyl ether (1.5 dm3) to give a white precipitate. The precipitate was filtered and dried to yield the product (20.2g, 96%). The formation of the potassium carboxylate was confirmed by IR. After complete reaction, the carbonyl peak of the carboxylic acid at 1670 cm- was shifted down to 1585 cm-' owing to the more single-bond character of the carbonyl group of the potassium carboxylate.(S)-(-)-2-Methylbutyl4'-Hydroxybiphenyl-4-carboxylate(12) To a solution of potassium salt 11 (20.2 g, 0.08 mol) and TBAH (4 g) in dry DMSO (300 cm3) was added (S)-2-methyl- butyl toluene-p-sulphonate (20.1 g, 0.0826 mol) (6).After being stirred at 60 "C for 20 h, the clear light-yellow solution was poured into water (1.2 dm3). The resulting precipitate was filtered off. The crude product was dissolved in methanol (400cm3) to give a light-brown solution, which was treated with activated carbon to produce a colourless solution. The white solid, obtained after the solvent was distilled, was 613 recrystallized from a mixture of methanol and water (1.25 :1.0, v/v) to yield crystals (22.1 g, 98.2%), purity, 99% (HPLC); m.p., 115.8 "C (DSC); 6, (CDCl,, TMS): 1.02 [m, 6 H, CH(CH3)-CH2CH3], 1.33-1.49 (m, 2 H, CHCH,CH,), 1.90 (m, 1 H, -CH-), 4.24 (m, 2 H,-OCH2-), 5.18 (s, 1 H, -OH), 6.98 (d, 2 H, 3'-H, 5'-H), 7.56 (d, 2 H, 2'-H, 6-H), 7.63 (d, 2 H, 2-H, 6-H), 8.1 1 (d, 2 H, 3-H, 5-H).(S)-(-)-2-MethyEbutyl 4'-( 1 1 -Vinyloxy)undecyloxybiphenyl-4-carboxylate (13-11) To a mixture of potassium carbonate (5.9 g, 0.0378 mol) and acetone (90 cm3) was added ester 12 (4.3 g, 0.015 mol). After being stirred for 2 h at 60 "C, the mixture turned yellow. Then, 11-bromoundecyl vinyl ether (4.0 g, 0.014 mol) and dry DMSO (5cm3) were added and the reaction mixture was stirred for 20 h at 60 "C. The reaction mixture was poured into water (250cm3) to give a white precipitate, which was extracted with chloroform.The chloroform solution was dried (MgS04) and the solvent was removed in a rotary evaporator. The resulting solid was recrystallized from methanol to yield the monomer 13-11 (5.2g, 75%), purity, 97% (HPLC). The monomer was further purified by column chromatography (silica gel, CH2C12 as eluent) to give 4.3 g (62%); purity, 99.9% (HPLC); m.p., 48.0 "C (DSC); 6H (CDCl,, TMS) 0.99 (m, 6 H, CH(CH3)CH2CH3), 1.29 (m, 16 H, OCH2CH2[CH,],, and CHCH2CH3), 1.63 (m, 2H, CH2CH,0Ar), 1.78 (m, 3 H,=CHOCH2CH2- and -CH2CH(CH3)CH2-), 3.64 (t, 2 H,=CHOCH2-), 3.98 (m, 3 H, -CH20Ar and =CH2, trans), 4.13-4.17 (m, 3 H,-C02CH2-and =CH2 cis), 6.40-6.51 (m, 1 H,=CHO-), 6.96 (d, 2 H, 3'-H, 5'-H), 7.54 (d, 2 H, 2'-H, 6'-H), 7.60 (d, 2 H, 2-H, 6-H), 8.07 (d, 2 H, 3-H, 5-H).(S)-(-)-2-Methylbutyl 4-(6-Vinyloxy)hexyloxybipheny~-4-carboxylate (13-6) Compound 13-6was synthesized by the same procedure as the one used for the preparation of 13-11. Starting with alcohol 12 (5 g, 0.0176 mol), bromoether 9-6 (3.646 g, 0.0176 mol) and potassium carbonate (6.9 g), ester 13-6 (5.5 g, 76%) was obtained, purity, 99% (HPLC); m.p., 38.5 "C (DSC); 6H (CDC13, TMS) 0.97 [m, 6 H,-CH(CH3)CH2CH3], 1.29 (m, 2H, CHCH,CH,), 1.46 (m, 4H,= CHOCH2CH2[CH2I2CH2), 1.65 (m, 2 H,-CH2CH20Ar-), 1.78 [m, 3 H,-OCH2CH2- and -Ch2CH(CH3)CH2-1, 3.65 (t, 2 H,=CHOCH,-), 3.96 (m, 3 H, =CH2, trans, and -CH20Ar), 4.14-4.18 (m, 3 H, =CH2, cis, and-C02CH2-), 6.40-6.51 (m, 1 H, =CHO-), 6.96 (d, 2 H, 3'-H, 5'-H), 7.50 (d, 2 H, 2'-H, 6'-H), 7.58 (d, 2 H, 2-H, 6-H), 8.06 (d, 2 H, 3-H, 5-H). Cationic Polymerizations Polymerizations were carried out in glass flasks equipped with Teflon stopcocks and rubber septa under argon atmos- phere at 0 "C for 1 h.All glassware was dried overnight at 180 "C. The monomer was further dried under vacuum over- night in the polymerization flask. Then the flask was filled with argon, cooled to 0°C and the methylene chloride, dimethyl sulphide and triflic acid were added viu a syringe. The monomer concentration was ca. 10 wt.% of the solvent volume and the dimethyl sulphide concentration was 20 times larger than that of the initiator. The polymer molecular weight was controlled by the monomer :initiator ([MIo :[II0) ratio. After the polymerization was quenched with ammoniacal methanol, the reaction mixture was precipitated into meth- anol.The filtered polymers were dried, and precipitated from methylene chloride solution into methanol several times until GPC traces showed no trace amounts of unreacted monomer. The polymerization results are summarized in Tables 1 and 2. Table 1 Cationic polymerization of (S)-(-)-2-methylbutyl 4-(11-vinyloxy)undecyloxybiphenyl-4-carboxylate(13-11)"and characterization of the resulting polymersb ~ ~~~~~~~ ~~~~~~~~ ~ DP phase-transition temperatures/ "C and corresponding enthalpy changes/kJ mol -polymer [MI,: [I], yield (%) M, x lop3 M,/M, GPC NMR heating cooling ~~~ ~ 4 88 1.8 1.07 4 5 g -1.4 k 52.8 (19.8) SA 92.1 (6.23) i i 84.4 (6.10) SA 42.9 (0.121) sC* g -3.1 SX 5.8 (1.05) SX 13.4 (-0.71) sC* 26.4 (1.30) k -0.11 (3.76) SX -8.1 g 32.7 (-1.17y k 47.5 (10.2) sA91.5 (6.14) i 7 93 3.0 1.10 6 6 g 4.2 k 56.8 (18.1) SA 104.8 (6.14) i i 97.5 (5.85) sA44.9 (0.142) sc* g 1.2 SX 15.5 (2.72) sC* 31.7 (-) k 37.4 (-0.627)" k 48.4 (2.59) SX -1.7 g (4.01) SA 104.5 (5.89) i 10 90 4.7 1.08 10 10 g 6.9 k 56.2 (16.8) SA 108.7 (6.02) i i 102.5 (5.81) SA 46.3 (0.159) sC* g 5.2 SX 19.4 (2.63) sC* 35.2 (-) k 38.2 (-0.418)" k 48.3 9.5 (2.72) sx 1.8 g (1.42) SA 108.5 (5.77) i 14 90 6.7 1.09 14 12 g 10.1 k 56.8 (15.4) sA 114.4 (5.56) i i 108.4 (5.43) SA 47.1 (0.180) sC* g 6.8 SX 20.8 (2.13) sC* 51.1 (0.142) SA 114.3 (5.56) i 12.7 (2.01) sx 5.1 g 18 93 8.2 1.08 17 18 g 11.1 k 57.2 (12.9) sA 118.6 (5.68) i i 112.3 (5.35) SA 48.4 (0.159) sC* g 8.2 SX 24.9 (2.09) sC* 52.2 (0.159) SA 118.1 (5.56) i 15.6 (2.05) sx 7.9 g 23 91 11.5 1.07 24 24 g 10.5 k 60.4 (15.0) sA 119.8 (5.56) i i 113.8 (5.31) SA 48.2 (0.159) sC* g 9.5 SX 24.0 (2.09) sC* 52.3 (0.121) sA 119.7 (5.27) i 15.7 (2.09) sx 7.5 g 30 92 12.1 1.10 26 26 g 11.5 k 62.2 (15.5) sA 122.8 (5.60) i i 115.4 (5.14) sA 48.8 (0.142) sC* g 10.1 SX 26.1 (2.34) sC* 53.3 (0.142) SA 121.8 (5.23) 16.7 (2.59) sx 7.9 g ~~ ~~~~~ ~~~~~ ~ Polymerization temperature, 0 "C; polymerization solvent, methylene chloride; [MI, =0.208; [(CH,),S], :[I], =20; polymerization time, 1 h.Data on first line are from first heating and on second line are from second heating scan.Crystallization during heating. Table 2 Cationic polymerization of (S)-(-)-2-methylbutyl 4-(6-vinyloxy)hexyloxylbiphenyl-4-carboxylate"(13-6) and characterization of the resulting polymersb ~~ ~ ~ DP phase-transition temperatures/ "C and corresponding enthalpy changes/kJ mol polymer [MI,: [I], yield (YO) M,XIO-3 M,/M, GPC NMR heating cooling 5 70 2.9 1.04 6 6 g 2.4 SA 90.8 (4.39) i i 83.5 (4.51) SA -4.1 g g 1.3 SA 90.4 (4.47) i 8 78 3.9 1.09 10 9 g 8.2 SA 95.9 (4.60) i i 87.7 (4.43) SA -0.5 g g 7.5 SA 95.5 (4.56) i 12 87 5.0 1.06 12 15 g 16.9 sA 103.2 (4.51) i i 95.1 (4.60) SA 8.8 g g 14.2 sA 102.6 (4.56) i 18 86 6.9 1.07 17 18 g 20.1 sX 40.1 (0.920) sA 106.9 (4.47) i i 99.8 (4.51) SA 29.9 (0.794) SX 13.5 g 19.3 SX 38.6 (0.418) sA 106.6 (4.68) i 23 91 7.5 1.10 19 23 g 25.9 SX 45.6 (1.630) sA 108.5 (4.47) i i 101.2 (4.60) sA 38.3 (0.501) sx 18.2 g 24.1 SX 46.7 (0.209) sA 108.3 (4.56) i 30 85 10.0 1.09 25 31 g 29.3 sX 49.7 (1.05) sA 111.7 (4.72) i i 102.4 (4.72) SA 48.1 (0.418) SX 21.4 g 27.9 sX 58.0 (0.418) sA 110.3 (4.47) i Table 1, except [M],=0.244.As for Table 1. J. MATER. CHEM., 1991, VOL. 1 Although polymer yields are lower than expected owing to losses during the purification process, conversions are almost quantitative in all cases. Results and Discussion In the area of low molar mass liquid crystals, there are some empirical rules that can be used to design compounds dis- playing chiral smectic C (s;) mesophases.26 Such rules are not available for the design of side-chain liquid-crystalline poly- mers exhibiting s; phase~.~-'~A classic example comes from our laboratory where repeated attempts to synthesize side- chain liquid-crystalline polymers exhibiting s; mesophases led to polymers exhibiting an sA me~ophase.~~.~~Therefore, we decided to perform a series of systematic investigations aimed to derive some empirical rules useful for the molecular engin- eering of side-chain liquid-crystalline polymers exhibiting s; mesophases.In some previous publications, we have reported the synthesis and characterization of some polymers contain- ing various polymer backbones and spacer length, and side groups derived from 4-[S-(-)-2-methyl- l-butyoxyl-4'- (hydroxy)-a-methylstilbene.'5~27' The influence of polymer backbones, spacer length and mesogenic group length on the ability to generate a sz mesophase was discussed.Here we will first discuss the synthesis and living cationic polymerization of (S)-(-)-2-methylbutyl 4'-( 1 1 -vinyloxy)un- decyloxybiphenyl-4-carboxylate13-11 and (S)-(-)-2-methyl- 9hmh me 615 butyl 4'-(6-vinyloxy)hexyloxybiphenyl-4-carboxylate13-6. In the second part, we will discuss their mesomorphic behaviour. Scheme 1 outlines the synthesis of vinyl ethers 13-11 and 13-6. The cationic polymerization of both monomers was initiated with the system CF3S03H-SMe2 and was performed at 0 "C in CH2C12.'9*20 The polymerization mechanism is described in Scheme 2.It is essential that the monomers used in these polymerization experiments are completely free of pro- tonic impurities. In order to achieve this degree of purity, after the purification by conventional techniques, the monomer is purified by passing through a chromatographic column con- taining silica gel and using methylene chloride as eluent. Poly- merization results are summarized in Tables 1 and 2. In both tables, conversions are less than quantitative owing to polymer losses during the purification process. However, at the end of the polymerization, HPLC and GPC traces showed that the monomer conversion was ca. 100%. Although the molecular weights determined by GPC and reported in Tables 1 and 2 are relative to polystyrene standards, they demonstrate that the ratio of [MIo: [II0 provides a very good control of the polymer molecular weight.In addition, all polydispersities of poly(l3-11) and poly(13-6) are <1.10. Absolute number aver- age molecular weights and degrees of polymerization were determined by 200 MHz 'H NMR spectroscopy. A representa-tive 'H NMR spectrum together with its protonic assignments is presented in Fig. 1. Degrees of polymerization were deter- mined by measuring the ratio of the doublet at 6 6.97 us. the j CH3CH (CH2CH), -2 CH2CHOCH3 I I I 0 0 0 CH2 CH2 CH2 i CH2 CH2 CH2 n (CH2)n -4 (CH2)n -4 (CH2)n -4 o CH2 CH2 CH2 I CH2 CH2 CH2 gI I I 0 0 08\ \ \ c=o c=o c=o I I I ? ? ? CH2 CH2 CH2 f CH -CH~*CH-CH~ *CH-CH~ CH2 CH2 CH2\ k CH3 CH3 CH3kp r H-8 7 6 5 4 6 (PPm) O*P (1.29) m,n (156) I flk9179\N V (1.18) 3 2 1 Fig.1 200 MHz 'H NMR spectrum of poly(l3-11) with theoretical DP=4 J. MATER. CHEM., 1991, VOL. 1 12 - I0 s=10- Q COpR' (R'=-CHp HCHpCH3) F CH3 8\ LOpW COpR' QQQ QQQ COpR' CO2W COpR' COpR' cop COpR' Scheme 2 Cationic polymerization of 13-11 and 13-6 broad triplet at 6 4.64.The degrees of polymerization deter- mined by NMR are summarized in Tables 1 and 2 and, unex- pectedly, especially for the case of poly(l3-11), they agree quite well with the results obtained by GPC. Fig. 2 presents the plots of M, determined by GPC and NMR and MJM, us. t U0 Q, 26 --.s3 u 8-Q m' 6-X 2= 4-O! I I I I I I 0 5 10 15 20 25 30 : 5 [Mlo/[Ilo I -A-A-A-A-A-A-0 5 10 15 20 25 30 35 [MI0 /[I10 Fig. 2 The dependence of the number average molecular weight (M,) determined by GPC (0)and by NMR (m)and of the polydispersity (M,/M,) of (a)poly(l3-11) and (b) poly(13-6) on the [MIo: [IlO ratio (A) II'i'2vSC*sA -20 10 40 70 100 130 -20 10 40 70 100 130 -20 10 40 70 100 130 T/"C T/"C TIT Fig. 3 DSC traces displayed during (a) the first heating scan (b), the first cooling scan and (c) the second heating scan by poly(l3-11) with different degrees of polymerization (DP) determined by GPC.Values of DP are printed on the left-hand side of each DSC scan J. MATER. CHEM., 1991, VOL. 1 [M]o/[I]o obtained for the polymerizations of 13-11 [Fig.2(a)] and 13-6 [Fig. 2(b)]. These plots demonstrate that within this range of molecular weights both monomers poly- merize through a living polymerization mechanism. As expected, the plots of absolute and relative M,us. [M]o/[I]o provide different slopes [Fig. 2(b)]. Fig. 3 presents the DSC traces of poly(l3-11) with various degrees of polymerization. As we can observe from this figure, the DSC curves of the first heating scan [Fig. 3(a)] differ from those of the second heating scan [Fig. 3(c)]. However, second and subsequent heating scans exhibit identical DSC traces. First and subsequent cooling scans also exhibit identical DSC traces [Fig. 3(b)]. On the first heating scan, all polymers exhibit a glass-transition temperature followed by a crystalline phase which melts into an S, mesophase.The s,-isotropic transition temperature has a stronger dependence on molecu- lar weight than that of the melting transition temperature [Fig. 3(a),Table 13. On the cooling DSC scans, all poly (13-11)exhibit an isotropic-s, followed by sA-sz and sE-sx phase transitions [Fig. 3(b)]. The nature of the sx phase was not identified. On the second heating scan, the phase behaviour of poly(l3-11) is strongly dependent on the molecular weight of the polymer [Fig. 3(c)]. Poly(l3-11) with degrees of polymeriz- ation <10 undergo the transition from S, to sz phase followed by crystallization through endothermic and exothermic peaks. The crystalline phase melts into an S, phase [Fig.3(c), Table 11. Poly(l3-11) with DP=4 has an additional exother- mic peak on the second heating scan [Fig. 3(c)]. This peak is due to the completion of the sx phase formation. The endo- therm due to sx-sz and the exotherm of the sz-k phase tran- sitions overlap [Fig. 3(c)]. Therefore, poly(l3-11) with degrees of polymerization <10 exhibit very narrow enantiotropic sx, sz and S, mesophases and a crystallization on the heating scan when their data are collected from the second DSC scans. Poly(l3-11) with degrees of polymerization >10 do not crys- tallize on the heating scan [Fig. 3(c)]. Subsequently, their DSC traces show very distinct transitions from sx to s: and from sz to S, mesophases. Therefore, when the thermal transitions of these polymers are collected from second and subsequent heating and first and subsequent cooling DSC traces, they exhibit a quite broad enantiotropic s,?j phase [Fig.3(b), (c), Table I]. However, if these polymers are annealed within their sz phase below the melting transition temperatures determined from the first heating scan [Fig. 3(a), (c)], they crystallize. This means that under equilibrium condition, poly(l3-11) exhibit an enantiotropic S, and a monotropic sc* mesophase. The difference between the behaviour of low and high molecular weight poly(l3-11) is determined by the difference between the crystallization ability of these two series of polymers. The low molecular weight polymers exhibit a high rate of crystalliz- ation and therefore, can crystallize on the second heating scan, while the higher molecular weight polymers have a much lower rate of crystallization and, subsequently, they do not crystallize on the second heating scan.This represents a classic example of the 'polymer effect' which shows how the kinetically con- trolled crystallization process affects the stability of a thermo- dynamically controlled mesophase. Similar examples of this behaviour are available both from 0ur'~7" and from other laboratories. l8 The thermal transition temperatures collected from Fig. 3 are summarized in Table 1 and plotted in Fig. 4(a) (data from the first heating scan), Fig. 4(b) (data from the first cooling scan) and Fig. 4(c) (data from the second heating scan). Plate 1 presents representative optical polarized micrographs of the sAand sc* phases exhibited by poly(l3-11).The sA meso-phase displays a focal conic-texture (Plate la). The S,-S; phase transition is accompanied by the formation of equidistant lines on the focal conic te~ture.~~~'~~'~*'~ The DSC traces of the first and second heating and of the 617 130, 704 k30i 10--04--0-0-/0-0 glassy/O-1 0 I I I 1 I 1301 (b) / SA 70-9i2 50-A-A-A-A-A-/ 30-SC' degree of polymerization 130 I / o~~-~-o-o-5-0-glassy 0 5 10 15 20 25 30 degree of polymerization Fig. 4 The dependence of phase-transition temperatures on the degree of polymerization determined by GPC of poly(l3-11): (a)data from the first heating scan, 0,q;.H, T(k-s,); 0,T(sA-i);(b) data from the first cooling scan, H, T(1-sA); A, T(s~-s$);+, T(s$-s,); 0, q;(c) data from the second heating scan, 0,q; 0, T(sx-sE); H, T(k-SA); A, T(S,*-SA);0,T(SA-i) J.MATER. CHEM., 1991, VOL. 1 t t 0 0 U a Q, 25.0DSA Jb -10 20 50 80 110 -10 20 50 80 110 -10 20 50 80 110 T/"C T/"C T/"C Fig. 5 DSC traces displayed during (a) the first heating scan, (b) the first cooling scan and (c) the second heating scan by poly(13-6) with different degrees of polymerization determined by GPC (DP). DP is printed on the top of each DSC scan first cooling scans of poly(13-6) with various molecular weights 120 I I are presented in Fig. 5(a)-(c). Over the entire range of molecu- lar weights, these polymers exhibit an enantiotropic sAphase.Polymers with degrees of polymerization 17 exhibit also an enantiotropic sx phase. The only difference between the first and second or subsequent heating scans consists of the fact that the enthalpy changes of the s,-sA phase-transition tem- peratures from the second heating and first cooling scans are 0 lower than those from the first heating scans (Table 2). This is i2 60-&---*--A due to the close proximity of the sx phase to the glass tran- 40-fA' sx sition of poly(13-6). Owing to this proximity, the sx phase is kinetically controlled. A similar behaviour was observed with 20-other polymer^.'^*'^ The phase-transition temperatures col- lected from the first and second heating scans are plotted in 0-Fig.6(a),while those from the first cooling scan in Fig. 6(b). As we can observe from both Fig. 5(a) and 5(b),the slopes of -20 ! I the dependences of sX-sA and sA-sX transition temperatures I I I I I 0 5 10 15 20 25 30 us. the degree of polymerization are steeper than the slope of degree of polymerization the dependence of < us. degrees of polymerization. Conse- quently, poly(13-6) with degrees of polymerization <17 exhibit120 a virtual sx mesophase. These results can be explained by using the thermodynamic schemes published previously.28 A general discussion on the polymer backbone effects on the phase behaviour of side-chain liquid-crystalline polymers will be published elsewhere.29 Some brief discussion on the same topic has been already p~blished.~'*~' Additional experiments on the synthesis and characterization of side-chain liquid-crystal- line polymers and copolymers with narrow molecular weight 40g60] distribution, well defined molecular weights, and exhibiting A /A-sz mesophases are in progress.The availability of these well defined polymers will allow the elucidation of the influence of various parameters such as molecular weight and polydispers- ity on the dynamics of the sz .mesophases exhibited by side- chain liquid-crystalline polymers. Financial support from the Office of Naval Research, DARPA -204 II I I 1 I 0 5 10 15 20 25 30 and an unrestricted Hercules Incorporated Aid-to-Education degree of polymerization grant is gratefully acknowledged.Fig. 6 The dependence of phase-transition temperatures on the degree Referencesof polymerization of poly(13-6) (determined by GPC): (a) data from the first heating (fh)and the second heating scan (sh), 0, (fh); A, 1 V. P. Shibaev and Ya. S. Freidzon, in Side Chain Liquid Crystal T(sX-SA) (fh);0,T(SA-i) (fh); 0, Tg(sh); A, T(s~-s,)(sh); .,T(SA-i) Polymers, ed. C. B. McArdle, Chapman and Hall, New York, (sh);(b)data from the first cooling scan, .,T(i-sA); A, T(s,-s,); 0,T, 1989, p. 260. J. MATER. CHEM., 1991, VOL. 1 Plate 1 Representative optical polarized micrographs (100 x) of (a) the sA mesophase displayed by poly(l3-11) (DP= 10) at 90 "C on the cooling scan; (b)the sc* mesophase displayed by poly(l3-11) (DP= 10) at 31 "C on the cooling scan V.Percec et al. (Facing p. 618) J. MATER. CHEM., 1991, VOL. 1 619 2 3 4 5 6 7 8 9 10 11 12 P. LeBarny and J. C. Dubois, in Side Chain Liquid Crystal Poly- mers, ed. C. B. McArdle, Chapman and Hall, New York, 1989, p. 130. J. W. Goodby, M. A. Waugh, S. M. Stein, E. Chin, R. Pindak and J. S. Patel, J. Am. Chem. SOC., 1989, 111, 8119. (a)V. P. Shibaev, M. V. Kozlovsky, L. A. Beresnev, L. M. Blinov and N. A. Plate, Polym. Bull., 1984, 12, 299; (b) V. P. Shibaev, M. V. Kozlovsky, N. A. Plate, L. A. Beresnev and L. M. Blinov, Polym. Sci. USSR, 1987,29,1616; (c) L. M. Blinov, V. A. Baikalov, M. J. Barnik, L. A. Beresnev, E. P. Pozhidayev and S. V. Yablon- sky, Liq. Cryst., 1987, 2, 121. (a) G. Decobert, F. Soyer and J.C. Dubois, Polym. Bull., 1985, 14, 179; (b)J. M. Guglielminetti, G. Decobert and J. C. Dubois, Polym. Bull., 1986, 16, 411; (c) J. C. Dubois, G. Decobert, P. LeBarny, S. Esselin, C. Friedrich and C. Noel, Mol. Cryst. Liq. Cryst., 1986, 137, 349; (d) G. Decobert, J. C. Dubois, S. Esselin and C. Noel, Liq. Cryst., 1986, 1, 307; (e) S. Esselin, L. Bosios, C. Noel, G. Decobert and J. C. Dubois, Liq. Cryst., 1987, 2, 505; (f)S. Esselin, C. Noel, G. Decobert and J. C. Dubois, Mol. Cryst. Liq. Cryst., 1988, 155, 371. (a) R. Zentel, G. Reckert and B. Reck, Liq. Cryst., 1987, 2, 83; (b) S. Bualek and R. Zentel, Makromol. Chem., 1988, 189, 797; (c) H. Kapitza and R. Zentel, Makromol. Chem., 1988, 189, 1793; (d) S. Bualek, H. Kapitza, J. Meyer, G. F. Schmidt and R. Zentel, Mol.Cryst. Liq. Cryst., 1988,155,47;(e)S. U. Vallerien, R. Zentel, F. Kremer, H. Kapitza and E. W. Fischer, Makromol. Chem. Rapid Commun., 1989, 10, 333; (f)R. Zentel, H. Kapitza, F. Kremer and S. U. Vallerien, in Liquid Crystalline Polymer, ed. R. A. Weiss and C. K. Ober, ACS Symposium Series 435, Washington DC, 1990, p. 207. S. Uchida,, K. Morita, K. Miyoshi, K. Hashimoto and K. Kawasaki, Mol. Cryst. Liq. Cryst., 1988, 155, 93. N. Koide, K. Uehara and K. Iimura, Mol. Cryst. Liq. Cryst., 1988, 157, 151. T. Suzuki, T. Okawa, K. Ohnuma and Y.Sakon, Makromol. Chem. Rapid Commun., 1988, 9, 755. M. Dumon, H. T. Nguyen, M. Mauzac, C. Destrade, M. F. Ach-ard and H. Gasparoux, Macromolecules, 1990, 23, 355. D. M. Walba, P. Keller, D. S. Parmar, N. A. Clark and M.D. Wand, J. Am. Chem. SOC., 1989, 111, 8273. H. Kapitza, R. Zentel, R. J. Twieg, C. Nguyen, S. U. Vallerien, F. Kremer and C. G. Willson, Adu. Mater., 1990, 2, 539. 17 18 19 20 21 22 23 24 25 26 27 28 W. Kreuder, 0.W. Webster and H. Ringsdorf, Makromol. Chem. Rapid Commun., 1986, 7, 5. (a) T. Sagane and R. W. Lenz, Polym. J., 1988, 20, 923; (b) T. Sagane and R. W. Lenz, Polymer, 1989,30,2269;(c) T. Sagane and R. W. Lenz, Macromolecules, 1989, 22, 3763. (a)J. M. Rodriguez-Parada and V. Percec, J. Polym. Sci., Polym. Chem. Ed., 1986,24, 1363; (b)R. Rodenhouse, V. Percec and A. E. Feiring, J. Polym. Sci., Polym. Lett., 1990, 28, 345; (c) V. Percec and M. Lee, J. Polym. Sci., Polym. Chem. Ed., 1991, 29, 327; (d) V. Percec and M. Lee, Macromolecules, 199 1, 24, 1017 and 2790; (e)V.Percec and M. Lee, Polymer, in the press; (f) V. Percec and M. Lee, Polym. Bull., 1991,25, 123; (8)H. Jonsson, V. Percec and A. Hult, Polym. Bull., 1991, 25, 131; (h)V. Percec, A. Gomes and M. Lee, J. Polym. Sci., Polym. Chem., submitted; (i) V. Percec, C. S. Wang and M. Lee, Polym. Bull., in the press. C. G. Cho, B. A. Feit and 0.W. Webster, Macromolecules, 1990, 23, 1918. (a) S. H. Korzeniowski, L. Blum and G. W. Golcel, Tetrahedron Lett., 1977, 22, 1871; (b)G. W. Gray, J. B. Hartlay and B. Jones, J. Chem. SOC., 1955, 1412. W. S. Johnson, C. D. Gutsche and R. D. Offenhauer, J. Am. Chem. SOC., 1946, 68, 1649. L. F. Fisher and C. K. Bradsher, J. Am. Chem. SOC., 1936, 58, 1738. B. Hahn and V. Percec, Macromolecules, 1987, 20, 2961. J. E. McKeon and P.Fitton, Tetrahedron, 1972,28, 233. (a)G. W. Gray and J. W. Goodby, Mol. Cryst. Liq. Cryst., 1976, 37, 157; (b)G. W. Gray and J. W. Goodby, Mol. Cryst. Liq. Cryst., 1978, 48, 127; (c) D. Demus, H. Demus and H. Zaschke, FluJige Kristalle in Tabellen, VEB Deutscher Verlag fur Grundstoffin- dustrie, Leipzig, 1974, I; 1984, 11; (d) D. M. Walba, C. S. Slater, W. N. Thurmes, N. A. Clark, M. A. Handsky and F. Supon, J. Am. Chem. SOC., 1986, 108, 5210. (a) V. Percec and B. Hahn, Macromolecules, 1989, 22, 1588; (b) B. Hahn and V. Percec, Mol. Cryst. Liq. Cryst., 1988, 157, 125; (c) V. Percec and B. Hahn, J. Polym. Sci., Polym. Chem. Ed., 1989, 27, 2367; (d) V. Percec, B. Hahn, M. Ebert and J. H. Wendorff, Macromolecules, 1990, 23, 2092; (e) V. Percec and C. S. Wang, Macromol. Reps, in the press. (a) V. Percec and A. Keller, Macromolecules, 1990, 23, 4347; (b) A. Keller, G. Ungar and V. Percec, in Advances in Liquid Crystal- line Polymers, ed. R. A. Weiss and C. K. Ober, ACS Symposium 13 14 15 16 T. Kutazume, T. Ohnogi and K. Ito, J. Am. Chem. Soc., 1990, 112, 6608. H. M. Colquhoun, C. C. Dudman, C. A. O’Makoney, G. C. Rob- inson and D. J. Williams, Adu. Muter., 1990, 2, 139. (a) V. Percec and C. S. Wang, J. Macromol. Sci.-Chem., in the press; (b)V. Percec and C. S. Wang, J. Macromol. Sci.-Chem., in the press; (c) V. Percec, C. S. Wang and M. Lee, Polym. Bull., in the press. V. Percec, D. Tomazos and C. Pugh, Macromolecules, 1989, 22, 3259. 29 30 31 Series 435, Washington D.C., 1990, p. 308. V. Percec, Polymer Backbone Eflects, invited lecture presented at the 13th International Liquid Crystal Conference, Vancouver, July 22-27, 1990; Mol. Cryst. Liq. Cryst., to be published. (a) V. Percec, D. Tomazos and R. A. Willingham, Polym. Bull., 1989, 22, 199; (b)V. Percec and D. Tomazos, Polymer, 1990, 31, 1658. V. Percec and C. Pugh, in Side Chain Liquid Crystal Polymers, ed. C. B. McArdle, Chapman and Hall, New York, 1989, p. 30. Paper 1/006 1 1H; Received 1 1 th February, 1991

ISSN:0959-9428    

DOI:10.1039/JM9910100611

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

J. MATER. CHEM., 1991, 1(4), 621-627 621 Sol-Gel Synthesis of WO, Thin Films Patrick Judeinstein and Jaques Livage Laboratoire de Chimie de la Matiere Condensee, Universite Pierre et Marie Curie-4, Place Jussieu, 75252 Paris, Cedex 05, France Versatile molecular precursors for the sol-gel synthesis of tungsten oxide thin films can be obtained via the reaction of tungsten oxychloride WOCI, with alcohols. Oligomeric species [WOCl,-,(OR),], are formed. Their molecular structure is analysed by infrared spectroscopy, nuclear magnetic resonance (NMR)spectroscopy (lH, 13C, lE3W), X-ray absorption spectroscopy (EXAFS) and small-angle X-ray scattering (SAXS). Hydrolysis of these precursors leads to the formation of tungsten oxide colloidal solutions. They can be easily deposited by dip- coating.Thin films ca. 3000A thick are obtained. Their morphology depends on the nature of the alcohol. Homogeneous films are obtained with bulky alkyl groups such as Pr'OH. These amorphous tungsten oxide layers exhibit electrochromic properties and could be used for display devices or smart windows. They can also be easily transformed, at room temperature, into crystalline hydrates WO,*nH,O (n= 1 or 2) when left in a humid atmosphere. Keywords: Sol-gel processing; Tungsten chloroalkoxide; Electrochromism; Thin film 1. Introduction Electrochromic layers based on amorphous W03 have been extensively studied during the last decade. Optical switching from white to blue can be reversibly performed in electro- chemical cells and used for display devices or smart windows.' Amorphous W03 thin films are usually made by vacuum evaporation,2 ~puttering,~ spray deposition4 or anodic oxi- dati~n.~Electrochromic layers deposited from colloidal W03 solutions have been reported recently.These solutions are obtained via the sol-gel route from tungsten alkoxides W(OR)p8 tungsten oxyalkoxides WO(OR)4, and W02(OR)2,9 or tungstic acid aqueous solutions HxW03."-12 However, these precursors are not stable toward hydrolysis or condensation. They cannot be handled easily and have to be stabilized in order to avoid the rapid precipitation of hydrous tungsten oxides. This paper describes the synthesis and characterization of tungsten oxide (WO,) thin films obtained from chloroalkox- ides WOCl, -,.(OR),precursors.These molecular species appear to be versatile precursors for the sol-gel deposition of electrochromic W03 thin films by dip-coating. The mor- phology of these films depend on the nature of the alkoxy groups. Best results were obtained by dissolving W0Cl4 into isopropyl alcohol Pr'OH. 2. Results and Discussion 2.1 Spectroscopic Techniques The different chemical species formed during the sol-gel synthesis of tungsten oxide thin films were characterized using the techniques outlined in the following. Infrared spectra were recorded in the 4000-300 cm-' range with a 783 Perkin-Elmer spectrometer using KRSS discs. 'H and 13C NMR spectra were recorded on a Bruker AM250 spectrometer using standard procedures.Modified molecular precursors were first dried under vacuum in order to remove alcohol in excess. A brown powder is obtained which is then dissolved into CDC1, in order to avoid exchange reactions with the solvent. 183WNMR spectra were recorded on an MSL4OO Bruker spectrometer (9.6 T) at a frequency of 16.65 MHz. The length of the pulse was 10 ps (t9*=40 p) and the recycle time 2.0 s. About 10 h (104 scans) were required in order to obtain a good signal-to-noise ratio with 0.1 mol dm-3 solutions in CD2C12. Room-temperature X-ray absorption spectra at the tungsten LIII edge (EXAFS) were recorded at LURE, the French synchrotron radiation facility, using the EXAFS I spec-trometer. Operating conditions in the storage DCI were the following: positions at energies of 1.85 GeV and intensities of ca.150 mA. The two-crystal Si (31 l} monochromator was fixed at its maximum flux position. The photon flux was measured by two ionization chambers. The entrance slit was 0.5 mm wide. A 1 vm tungsten metallic foil was used for energy calibration. Energy was scanned by 2 eV steps over a 10 050-1 1 050 eV energy range. Accumulation time was 1.6 s per point. Solutions (0.1 mol dm-3) were sealed into the cells in order to avoid further hydrolysis. Cells were 2 mm thick with windows made of X-ray transparent Kapton. EXAFS modulations were analysed with standard method^.'^,'^ The continuous absorption background was estimated by fitting the spectrum before the edge by a Victoreen function.The main absorption, beyond the edge p(E), was fitted with an iterative procedure. Normalization of the EXAFS signal was achieved by using z(E)=[po(E)-p(E)]/[p(E)]. Structural information was extracted on the basis of the single-scattering theory. Tabulated values of Teo and Lee" were checked with reference compounds such as Na2W04, (BUqN)2W6019 and W03. The electron mean-free path A(k)was approximated by ;I=k/T,where r is a fitting parameter. Small-angle X-ray scattering (SAXS) spectra were recorded with the synchrotron radiation of the DCI storage ring (LURE) in order to take advantage of the intense X-ray beam associated with point collimation and the D22 bench device which allows very small angles to be reached. Collected data cover q values from 0.04 to 0.8 A-1 and from 0.003 to 0.1 A-', depending on the sample-detector distance. The scattering- vector amplitude q =2 sin 8/A,where 28 is the scattering angle and A the selected wavelength, is equal to 1.5 A.The analysed Bragg zone ranges between 5 and lo3 A. Samples were sealed into cells 1 mm thick. As a result of the narrow beam, the only correction was to subtract scattering arising from the solvent. Usual SAXS data analysis procedure was used in order to obtain the Guinier radius of gyration and the mass of the smallest particles.16 For larger particles, a log-log plot of the experimental scattering functions led to curves exhibit- ing one or two linear parts. The cross-over distance, B, is a characteristic length of ~oherence.'~ The slope of the linear parts give information on the compacity of aggregates. They can be analysed as more or less dense particles18 or fractal aggregates.'' 2.2 Synthesis and Characterization of Molecular Precursors Tungsten alkoxides W(OR)6 can be synthesized by reacting WCl, with an alcohol ROH as follows:2o WCl6 +ROH WCl6 -,(OR), +XHCl (1) This reaction does not go to completion unless a base is added in order to remove HC1." Moreover, Wv' is reduced into Wv species such as WCl,(OR)3, and the solution rapidly turns blue." Spontaneous oxidation then occurs when the solution is left in dry air and the colour turns yellow within a few days.Reduction can be avoided by using tungsten oxychloride instead of WC16.WOCl, (5 g) was dissolved in 100 cm3 of pure alcohol (isopropyl alcohol distilled on sodium) under a dry atmosphere. A violent exothermic reaction occurs while gaseous HCl evolves but the solution remains yellow. Accord- ing to similar experiments performed on MoOC~,,~~the overall reaction could be described as follows: WOCl, +xROH eWOCl,-,(OR), +xHCl (2) Alcohol in excess is removed upon heating the solution under vacuum at ca. 60 "C. A brown powder is obtained which can be dissolved in the parent alcohol or even in a neutral solvent such as CCl,. Concentrated solutions and powders are very sensitive towards moisture and therefore rather difficult to handle. Dilute solutions are much less reactive. They can be kept in a closed vessel and remain stable for months.Infrared spectra of concentrated solutions of the chloro- alkoxide in CC14 are shown in Fig. 1. All the bands typical of alkyl groups can be seen above 1OOOcm-'. The intensity of the v(C0) stretching vibration at 1130 cm-' increases significantly compared with pure isopropyl alcohol suggesting that isopropoxy groups are bonded to tungsten via the oxygen atom. Vibrations involving tungsten atoms are on the low- energy side. The sharp band close to 970 cm-' can be assigned to W=O double bonds.', Broad bands between 800 and 730 cm-' suggest the formation of W-0-W bridges" similar to those observed in polyanions. W-Cl vibrations can be seen at ca. 350 cm-'.26 'H and 13C NMR spectra of oxychloroalkoxide precursors were recorded in CDC13 in order to avoid exchange reactions between alkoxy ligands and the solvent.The 'H NMR spectrum exhibits several sets of broad peaks typical of isopropoxy groups. However, the corresponding chemical shifts are shifted towards low field C1.07 (CH3)2CHOH, 3.88 (CH3)2CHOH for isopropyl alcohol, 1.2-1.5 -n A7-'CH J. MATER. CHEM., 1991, VOL. 1 (CH3)2CHO-WW, 4.3 and 5.8 pprn (broad peaks) (CH3)2CNO-W]. Such unshielding effects could be due to the presence of tungsten atoms. Similar chemical shifts have already been observed in W(OPri)6.27 The 13C NMR spectrum of the same solution exhibits two groups of peaks as for pure Pr'OH [Fig. 2(a)]. The corresponding chemical shifts also suggest a strong unshielding due to the presence of heavy tungsten atoms.Three broad signals corresponding to different W-0-C bonds are observed at ca. 83,84 and 87 ppm. The large chemical shifts and the multiplicity observed for both 'H and 13C NMR peaks suggest that bridging and terminal alkoxy groups are simultaneously present in the molecular precursor. The rather broad linewidth could be due to some chemical exchange between these alkoxy groups. The 183W NMR spectrum was recorded with a solution of tungsten oxychloroalkoxide (0.2 mol dm-3) dissolved in a Pr'OH-CD2C12 mixture. It exhibits three signals [Fig. 2(b)] at ca. 36.2, -37.0 and -122.8 ppm. The 6 =O ppm reference corresponds to Na,W04 in D20 (pH 9). This suggests three different chemical environments for tungsten.Two processes could account for the observed chemical shifts which are small compared to WC16(2181 ppm) and W(OR)6 (ca. -480 ppm).28 (i) The substitution of OR groups by heavier and more electronegative C1 atoms changes the shielding effect. Similar features have been reported for WF6-,(OR),29 and for "V in the NMR spectra of VO(C1)3~,(OR),.30 In both cases chemical shifts larger than 200ppm are observed when x increases one unit. (ii) Axial distortion of local symmetry from a regular octahedron (0,) to a pyramidal C4"symmetry due to the W=O double bond also leads to an unshielding effect as was already reported in polyanion~.~~ Both processes could 100 80 60 40 20 I IIIIIIIIIIIIIIIIIIIIII 80 ' 50 20 -10 -40 -70 -100 -130 'w-0 6 (PPm) Ill I I I1; 111 I 1 I 4000 3000 2000 1000 500 Fig.2 (a) 13C-('H) decoupled NMR spectra of the WOC1,-Pr'OH v/cm -precursor in CDC1, (+, chemical shifts for Pr'OH; *, CDCl, signal); (b) lS3W NMR spectra of the WOC1,-Pr'OH precursor in Fig. 1 Infrared spectra of the WOC1,-Pr'OH precursor in CCl, CD2C12-Pr'OH J. MATER. CHEM., 1991, VOL. I actually occur simultaneously leading to chemical shifts in opposite directions. It is therefore difficult to claim which one is predominant. All resonance peaks are rather broad. This could be due to short relaxation items arising from the formation of oligomeric species and exchange reactions between different sites. Therefore a definite conclusion is not possible.NMR spectra could be assigned to a single oligomeric molecule as well as a mixture of different chemical species. EXAFS spectra were recorded with tungsten oxychloro- alkoxide solutions in isopropyl alcohol (0.2 mol dm-3). The radial distribution function of the EXAFS signal and the k-space filtered EXAFS spectrum are shown in Fig. 3. The EXAFS curve was fitted with the conventional single-scat- tering formalism. It was not possible to get a good fit unless both oxygen and chlorine atoms were included in the first- neighbour shell. The best fit then leads to the values reported in Table 1. The 0:C1 first neighbours ratio is close to 3 :2, suggesting the following chemical formula WO(OR)2C1,. A short tungsten-oxygen distance is observed at ca.1.76 A. It could correspond to the W=O double-bond contribution. However, the Debye-Waller factors associated with all W-0 distances are rather large, suggesting a broad distribution of these distances arising for instance from strongly distorted octahedra. W-C1 distances at ca. 2.36A are rather long compared with the mean W-Cl distances found in WC1, (2.24A, monomer), W0Cl4 (2.29 A) or W02C12 (2.31 A, poly-meric structure with W-0 bridging bonds).31 They could therefore be assigned to longer bridging C1 atoms rather than terminal ones. This would suggest the presence of oligomeric species. However W ... W correlations are not observed. This could be due to the small size or a linear structure of Y, Y 3 4 5 6 7 a 910 0123456 k/A-RIA h h Y 3 4 5 6 7 8 9 10 0123456 klk' RIA Fig.3 EXAFS spectrum of the WOC1,-Pr'OH precursor in isopropyl alcohol (a), (b)and the corresponding hydrolysed sample (h=20) (c), (d).(a),(c)k-space filtered EXAFS spectrum; (b),(d) radial distribution function oligomeric species leading to a small number of tungsten neighbours. Moreover a rather broad distribution of W-W distances together with the large extent of 5d tungsten orbitals could reduce significantly the intensity of such contributions. SAXS experiments were performed with tungsten oxychlor- ide solutions in isopropyl alcohol. Absolute intensity measure- ments suggest the presence of anisotropic molecular species with a Guinier radius close to 5 A. Scattering curves, expressed as In I =f(Q2) and In (IQ)=f(Q2)plots, can be assigned to oligomeric species ca.3-4 A in diameter and 10 A long. Molecular precursors are rather sensitive toward reduction. They become blue upon UV irradiation. Spontaneous reduction is even observed when WCI, is dissolved into Pr'OH. The electron paramagnetic resonance (EPR) spectra of such reduced solutions are similar. They are characteristic of Ws+ ions in an axial ligand field (Fig. 4).The frozen solution spectrum recorded at low temperature is well resolved. Satellite lines can be seen on each side of the spectrum. They are due to the hyperfine interaction of the electron spin (S= 1/2) with the nuclear spin of the lg3W tungsten nucleus (I = 1/2, natural abundance 14.8%).EPR parameters are gll= 1.81, g,= 1.78, All=.125 G and Al= 150G. They are close to those found in (WOC1s)2- or (WOC1,,H20)3-where W5 + is surrounded by oxygen and chlorine atoms.32 g, is smaller than gI1suggesting that the unpaired electron is in a d,, orbital. An EPR signal is still observed at room temperature. This is quite unusual in most tungsten-oxygen compounds in which electrons are so delo-calized that EPR signals can no longer be seen.33 This suggests that unpaired electrons remain rather localized even at room temperature. Such an electron localization could be due to the small size of the oligomeric species and the nature of W-OR-W and W-Cl-W bridges between W atoms. 2.3 Hydrolysis and Condensation Hydrolysis of precursor solutions in Pr'OH (0.2mol dm -3, was performed at room temperature by adding a mixture H20-Pr'OH (10% H20 in weight).The hydrolysis ratio is 3200 3450 3700 3950 4200 magnetic field/G Fig. 4 EPR spectrum of the WOC1,-Pr'OH precursor in Pr'OH reduced upon UV irradiation (recorded at 77 K) Table 1 EXAFS parameters of precursor compound (WOC1,-Pr'OH) and colloids obtained after hydrolysis (h=20) nature of neighbour RIA N Deb ye- Waller coeff. (6) AEIeV fitting coefficient WOC1,-Pr'OH 0 0c1 1.76 1.88 2.36 0.9 1.8 2.4 0.04 0.08 0.05 10.4 10.4 -0.6 r =0.7 R = 3% C 3.7 5.5 0.12 0.1 WOCI,-Pr'OH--20H20 0 0 1.76 1.92 1.1 3.5 0.06 0.06 10.4 10.4 r =0.7 W 3.50 5.0 0.10 15.2 R = 5% 0 3.10 6.8 0.15 10.4 624 given by h= H20:W.Transparent yellow colloidal solutions are obtained up to h=5, while white gelatinous precipitates are obtained beyond h= 10 (Table 2). Infrared spectra are progressively modified as hydrolysis proceeds (Fig. 5). Absorption bands at ca. 1100 cm-' disap-pear as well as the W-C1 bands at ca. 350cm-'. The IR spectrum between 500 and 1000 cm-' becomes more and more poorly resolved. A broad absorption with two maxima rising from the formation of an oxide network is seen when h=50 [Fig. 5(d)] with a small shoulder at ca. 980 cm-' corresponding to W =0 bonds. The Fourier transform of the EXAFS spectrum of hydro- lysed samples (h=20) exhibits two peaks corresponding to W-0 and W-W distances (Fig. 3). A careful analysis of the first-neighbours shell shows that all Cl atoms have been removed from the co-ordination sphere of tungsten, whereas W-0 distances are not significantly modified.The best fit (Table 2) leads to two W-0 distances (1.76 and 1.92 A) corresponding to a distorted octahedron with one short W=O bond. Tungsten neighbours can be seen in the second shell at 3.50A. Such a distance lies between those reported for edge sharing (W-W =3.3 A) and corner sharing (W-W =3.7 A) W06 ~ctahedra.~, It could therefore corre- spond to a random distribution of both species. Small-angle scattering curves of samples hdyrolysed with h= 10 recorded at different times after water has been added vco / vw-o Ill I I I Ill I I III 4000 3000 2000 1000 500 vfcm-Fig.5 Infrared spectra of the hydrolysis product of WOC1,-Pr'OH solutions with different hydrolysis ratio h=H20:W (recorded after I h): (a) 1, (b) 5, (c) 10, (d) 50 J. MATER. CHEM., 1991, VOL. 1 are shown in Fig. 6. For the first 2 h scattering curves In I= Aln Q) exhibit a single linear variation in the Porod region, with a slope of -1.4 suggesting the anisotropic growth of rod-like particles. A maximum can be seen at low Q values (Q=0.002 k').It does not vary with dilution and can be attributed to particles ca. 400 A long. Colloidal particles keep on growing as hydrolysis and condensation proceed further. Two linear regimes are observed on the scattering curve. The first one at high Q with a smaller slope close to -1.1 and the other one at low Q with a larger slope of ca.-2.5. These curves suggest some complex growth mechanism involving two different steps. The first regime could be assigned to an aggregation process arising from the diffusion of rod-like particles while the second one could correspond to the forma- tion of branched polymers arising from the cross-linking of small rods. The crossing point between these two curves gives an order of magnitude of the coherence length, B, of these rods. This coherence length decreases as the hydrolysis ratio, h, increases, from B =600 A for h= 10 to B =200 A for h= 20. It does not vary with time for a given hydrolysis ratio. More nuclei are formed but they do not grow as much. The growth process seems to stop when the hydrolysis ratio becomes too small (h~5).2.4 Thin-film deposition W03 thin films can be deposited easily from molecular precursor solutions via the dip-coating pr~cedure.~' A glass plate covered with a conducting transparent layer (antimony- doped tin oxide) is dipped into an alcoholic solution of WOCl, (0.lmol-0.3 mol dm-3). It is then pulled out at a controlled speed of ca. 1 cm s-'. The plate is left in air so that hydrolysis and condensation occur spontaneously in the ambient moisture. Evaporation of the solvent (alcohol) occurs simultaneously and a dry coating ca. 800 A thick is obtained O1 -2 OO-'I OUo0 -6 I I I I I I I 1 -6 -5.5 -5 -4.5 -4 -3.5 -3 -2.5 -2 In Q Fig. 6 SAXS curves of the hydrolysis product of WOC1,-Pr'OH + 10H20.Delay after hydrolysis, tfh: (+) 0.2, (+) 2, (0)10 Table 2 Hydrolysis of 0.1 mol dm-3 isopropoxy oxochloroalkoxide solutions in isopropyl alcohol hydrolysis ratio h=HzO:W final appearance of the compounds stability 0 1 2 5 10 20 50 yellow solution oligomers [WOC14-,(OR),], n I3 yellow solution yellow solution rod-like particles 25 A long white opaque solution polymeric particles diameter >600 A white gelatinous precipitate precipitate yellow scattering solution rod-like particles 100 long months months weeks weeks 5 days J. MATER. CHEM., 1991, VOL. 1 within 30 min. The process can be repeated several times in order to have thicker films. Up to six layers have been superimposed by successive coatings leading to films ca.5000 A thick. Three dips only were required in order to obtain films exhibiting good electrochromic behaviour. Rather uniform coatings are obtained as shown by scanning electron microscopy (Fig. 7). Their morphology mainly depends on the nature of the alcohol in which WOCl, is dissolved. Spherical particles embedded into a continuous film are obtained with methanol [Fig. 7(a)]. More uniform films are obtained with alcohols such as isopropyl alcohol or butanol [Fig. 7(b)].The surface of these films appears to be smoother when the size of the alcohol molecule increases. This is probably related to the slower hydrolysis rate of tungsten precursors and the slower evaporation rate of the solvent.Both of them are known to decrease when the steric hindrance of the alkoxy group increases. Fig. 7 Scanning Electron Microscopy of sol-gel W03 thin films: (a)dip coating in WOC1,-MeOH solution; (b)dip coating in WOCI,- BuOH solution X-Ray diffraction patterns of W03 layers are typical of amorphous samples. The infrared absorption spectrum of a thin film deposited onto a KRS5 plate is quite similar to that of WOC1,-PriOH+50H20 [Fig. 5(d)]. It exhibits a strong absorption below 1000 cm- due to the formation of an oxide network. These bands are very broad, suggesting a large distribution of W-0 distances and W-0-W angles in the amorphous oxide. The shoulder at 980 cm -could be assigned to a short W=O bond. Absorption bands arising from water molecules can be seen on the high-energy side of the spectrum at ca.3500 and 1600 cm-'. No band corresponding to alcohol molecules or alkoxy ligands can be seen showing that all organics have been removed upon hydrolysis and drying. DTA experiments show two phenomena: an endothermic process between 50 and 170 "C corresponding to the departure of water molecules, and an exothermic process at 350 "C corresponding to the crystallization of orthorhombic W03. When dried under ambient conditions, hydrated tungsten- oxide films correspond to amorphous W03 * 1 .8H20. They can be easily transformed into other tungsten oxide phases as follows. (i) Crystallization occurs upon heating at 350 "C leading to orthorhombic W03. (ii) Crystallization is also observed at room temperature when the film is left in a humid atmosphere.This leads to the formation of the well known dihydrate W03 *2H20.This hydrous oxide exhibits a layered structure. Water molecules are intercalated between the W03 layers and the basal distance corresponds to d=6.91 However, no preferential orientation of these layers is observed while anisotropic layers were deposited from aqueous solu- tions of colloidal tungstic acid.36 (iii) Dehydration of the crystalline dihydrate occurs when heated at 120 "C during 20 h. It leads to the monohydrate phase W03- 1H,O which also exhibits a layered structure with a basal distance d= 5.34 Scanning electron microscopy of these tungsten oxide layers shows that the coverage and adhesion of the oxide coating onto the substrate are not affected by crystallization when the amorphous film is heated or left in a humid atmosphere.However, electrochromic properties strongly depend on the water content and ~rystallinity.~~ 3. Conclusions This work shows that electrochromic W03 thin films can be synthesized easily via the hydrolysis and condensation of cheap and stable molecular precursors. The reaction of alcohol with WCl, or better WOCl, leads to the formation of WOC1,-,(OR), solutions, which can be kept without major transformation for more than 6 months in a closed vessel. Moreover these solutions remain reactive enough to give uniform coatings when deposited onto a glass substrate in the presence of a humid atmosphere.WOCl, -,(0Pri), precursors have been prepared by dissolv- ing tungsten oxychloride in isopropyl alcohol. The molecular structure of these chloride alkoxides is not easy to determine. No crystal can be extracted from the solution and an amorph- ous powder is obtained upon evaporation. Even the chemical composition is not obvious because it is almost impossible to remove all alcohol and HCl in excess. Therefore chemical titrations were not successful. Infrared spectra show that both W-C1 and W-OR bonds are present in the molecular precursor as well as short W =0 double bonds. These results agree with EXAFS data suggest- ing that tungsten is surrounded by both oxygen and chlorine atoms with a ratio 0:C1 close to 3 :2 suggesting the following composition WOC12(0Pri)2.Infrared bands in the low-fre- quency range and SAXS measurements suggest the formation of oligomeric species. However, W ... W correlations cannot J. MATER. CHEM., 1991, VOL. 1 be seen in the EXAFS spectrum. This does not preclude the presence of oligomers. It could be due to the large extension of 5d atomic orbitals and the small number of tungsten neighbours. 183W NMR spectra suggest the presence of three different tungsten species. They might be found in a single oligomeric molecule. It seems, however, more reasonable to suggest that several molecular species such as [WOCl, -x(OPri)x],, with nI 3 are in equilibrium. Their rela- tive abundance should depend mainly on tungsten concen- tration and the steric hindrance of the alkyl chain.Similar analyses (IR, EXAFS, SAXS) have been performed on WC1,-PriOH solutions. They give the same features as for W OC1,-Pr'OH precursors, i.e. an oligomeric structure with W-C1, W-0 and W=O bonds. It shows that alcoholysis of WCl, leads to a change of co-ordination around tungsten from an Oh(WX,) to a C4"(WOX,) symmetry. Such changes are energetically favoured by the formation of an 0x0 group bonded to tungsten.34 Such bonds involve a large overlap of the d,-p, orbitals in order to produce a strong double bond W=O. Moreover, they are reinforced by the proximity of more ionic W-C1 bonds. This could explain the differences observed between chlororalkoxides and oxyal- koxides of the WO(OR), series.39 According to literature, double W=O bonds are not observed in these solid com- pounds, except for large tert-butoxy derivatives.The corre- sponding vibration band at ca. 950 cm-' cannot be seen on infrared spectra. One could also suggest a modifica-tion of these structures from solid-state [poly-meric-W(OR),-0-W(OR),-] structure to solution in parent alcohol [O=W(OR),], as observed with WOCl,, by IR spectro~copy.~~ Hydrolysis of WOC1,-Pr'OH solutions leads to the forma- tion of yellow colloidal solutions or white gelatinous precipi- tates depending on tungsten concentration and the hydrolysis ratio. Infrared and EXAFS data show that both chlorine and alkoxy ligands are removed upon hydrolysis, leading to the formation of condensed species evidenced by W ...W corre- lations in the EXAFS spectrum. A mean distance W-W= 3.5 A is found suggesting that condensation occurs via both edge and corners sharing [WO,] octahedra. Two linear variations of In I=f(ln Q) are seen on the small-angle X-ray scattering curves. They show that condensation leads to the formation of large branched aggregates (diameter d >600 A) with a complex structure. It could be assumed that rod-like particles are first formed which then aggregate to give dense ramified polymers. The anisotropic growth of the colloidal particles should be related to the functionality of the molecular precursor. In the case of tungstic acid prepared upon acidification of tungstate aqueous solutions it was shown that large condensed species were formed from neutral precursors such as [H2 W04].40*41 Co-ordination expansion occurs via the nucleophilic addition of two water molecules leading to the formation of sixfold co-ordinated precursors with one short W=O bond and a long W-OH2 bond along the z axis.Four equivalent W-OH bonds are formed along x and y axis so that neutral precursors have a functionality of four in the xy plane. This leads to the anisotropic growth of platelet- like particles as observed by electron micro~copy.~~ Scattering curves, performed by SAXS, with such colloids give a single linear In I=Aln Q) plot.43 The corresponding slope of -2 agrees with an anisotropic two-dimensional growth process leading to the formation of platelets.Chloride alkoxides WOCl,-,(OR), should behave in a different way from the previous inorganic precursors. W=O bonds are conserved during the hydrolysis process. Both chlorine and alkoxy groups are removed. However, these ligands do not exhibit the same reactivity toward hydrolysis. W-C1 bonds are more ionic than W-OR bonds. They should be hydrolysed first. The first apparent functionality therefore decreases around 2 leading to the anisotropic growth of rod-like particles. These rods then condense together. Moreover, branched polymeric networks are also formed during the hydrolysis of WO(OEt), solution^.^ Depending on the hydrolysis ratio, h, alkoxide precursors lead to the forma- tion of more or less branched polymers with a small Porod exponent, -1.8 to -2 measured by SAXS.The functionality of these alkoxide precursors is rather high, and the reactivity of all hydrolysable functions are quite similar. This leads to species of complex geometry described in terms of fractals. Thin films of large area can be deposited easily by dip- coating. The morphology of these films mainly depends on the nature of the alcohol used for the sol-gel synthesis. It appears that more uniform films are obtained with bulky alcohols such as Pr'OH. This can be attributed to the lower reactivity of the corresponding alkoxy groups toward hydroly- sis and the slow evaporation rate of the solvent. Such morpho- logical variations reflect the differences observed on the growing mechanism of different WOC14-PrOHi precursors, especially the value of the coherence length of the particles.These differences appear to play an important role on the microstructure of tungsten oxide coatings. Films deposited from organic solutions [Fig. 7(a),(b)] are made of spherical particles embedded into a continuous film. They are similar to those obtained from polymeric gels. Films deposited from aqueous solutions of tungstic acid exhibit a layered structure made of the stacking of flat colloidal particles.,' Moreover, amorphous W03 1.8H20 thin films deposited via the sol-gel process appear to be very versatile precursors in order to obtain other W03 *nH20 coatings. The hydration state can be controlled by the relative humidity of the ambient atmos- phere and the temperature. Crystalline layers are formed upon heating at 350°C (orthorhombic W03) or even at room temperature in the presence of water vapour (W03.2H20 and W03 *H20).The sol-gel process therefore leads to a wide range of tungsten oxide thin films which could be used for making display devices or smart windows. The electrochromic proper- ties of these oxides can be tailored via the chemical control of parameters such as the nature of the alcohol, the OR :C1 ratio in the precursor, the hydrolysis ratio h= H20: W, the amount of water in the layer or even the crystalline state. Authors are very grateful to M. Verdaguer, A. Michalowicz and B.Cabane for their help in EXAFS and SAXS experi- ments. They appreciate the assistance of R.Thouvenot for la3W NMR spectra. References 1 K. Bange and T. Gambke, Adv. Muter., 1990, 2, 10. 2 S. K. Deb, Philos. Mag., 1973, 27, 801. 3 S. Beni, M. Manfredi and G.Salviati, Solid State Commun., 1980, 70, 203. 4 D. Craigen, A., Mackintosh, J. Hickman and K. Colbow, J. Electrochem. SOC., 1986, 1529. 5 L. D. Bruke and E. J. M. O'Sullivan, J. Electroanal. Chem., 1980, 111, 383. 6 K. Kamanaka, J. Appl. Phys., 1981, 20, L307. 7 H. Unuma, K. Tonooka, Y. Suzuki, T. Furusaki, K. Kodaira and T. Matsushita, J. Muter. Sci. Lett., 1986, 5, 1248. 8 N. R. Lynam, F. H. Moser and B. P. Hichwa, SPZE, Optical Materials Technology for Energy Eflciency Solar Energy Conver- sion VZ, 1987, 823, 130. 9 M. I. Yanovskaya, I. E. Obvintseva, V.G. Kessler, B. Sh. Galya- mov, S. 1. Kucheiko, R. R. Shifrina and N. Ya. Turova, J. Non-Cryst. Solids, 1990, 124, 155. J. MATER. CHEM., 1991, VOL. 1 627 10 11 12 13 14 A. Chemseddine, R. Morineau and J. Livage, Solid State Zonics, 1983, 9/10, 357. A. Chemseddine, M. Henry and J. Livage, Rev. Chim. Minhrale, 1984, 21, 357. T. Yamase, M. Matsuzawa and Y. Sasaki, Znorg. Chim. Acta, 1987, 127, L9. B. K. Teo, EXAFS: Basic Principles and Data Analysis, Inorganic Chemistry Concepts 9, Springer-Verlag, Berlin, 1986. A. Michalowicz, EXAFS pour le MAC, Ecole du CNRS Structures Fines d'Absorption X en Chimie, ed. H. Dexpert, A. Michalowicz and M. Verdaguer, CNRS Press, Paris, 1988. 27 28 29 30 31 32 D. C. Bradley, M. H. Chisholm and M. W. Extine, Inorg.Chem., 1977, 16, 1791. M. Minelli, J. H. Enemark, R. T. C. Brownlee, M. J. O'Connor and A. G. Wedd, Coord. Chem. Rev., 1985,68, 169. W. McFarlane, A. M. Noble and J. M. Winfield, J. Chem. SOC. (A), 1971,948. D. Rehder, Bull. Reson. Magn., 1982, 4, 33. D. L. Kepert, The Early Transition Metals, Academic Press, 1972, p. 278. J. T. C. Van Kemenade, Recueil SOC. Chim. Pays-Bas, 1973, 92, 1102. 15 16 17 18 19 20 21 B. K. Teo and P. A. Lee, J. Am. Chem. SOC., 1979, 101,2815. A. Guinier and G. Fournet, Small Angle Scattering of X-Rays, Wiley, New York, 1955. Sir G. Allen and J. C. Bevington, Comprehensive Polymer Science, Pergamon Press, Oxford, 1989, 1, 127. 0. Glatter and 0. Kratky, Small Angle X-Ray Scattering, Aca-demic Press, London, 1982. D. W. Schaeffer and K.D. Keefer, Phys. Reu. Lett., 1984, 53, 1383. R. A. Walton, Prog. Znorg. Chem., 1972, 16, 1. L. B. Handy, K. G. Sharp and F. E. Brinckman, Znorg. Chem., 1972, 11, 523. 33 34 35 36 37 38 39 A. Chemseddine, C. Sanchez, J. Livage, J.P. Launay and M. Fournier, Inorg. Chem., 1984, 23, 2609. M. T. Pope, Heteropoly and Zsopoly Oxometalates, Inorganic Chemistry Concepts 8, Springer-Verlag, Berlin, 1983. M. L. Freedman, J. Am. Chem. SOC., 1959,81, 3834. A. Chemsseddine, F. Babonneau and J. Livage, J. Non-Cryst. Solids, 1987, 91, 271. J. T. Szymanski and A. C. Robert, Can. Mineral., 1984, 22, 681. P. Judeinstein and J. Livage, SPZE, Sol-Gel Optics, 1990, 1328, 344. S. I. Kucheiko, N. Ya. Turova, N. I. Kozlova and B. V. Zhadanov, Koord. Khim., 1985, 11, 1521. 22 23 0.J. Klejnot, Znorg. Chem., 1965, 4, 1668. S. K. Anand, R. K. Multani and B. D. Jain, J. Znd. Chem. SOC., 40 J. Livage, M. Henry and C. Sanchez, Prog. Solid State Chem., 1988, 18, 259. 24 25 1968,45, 1130. C. Rocchioccioli-Deltcheff,R. Thouvenot and R. Franck, Spectro-chim. Acta, 1976, 32A, 587. K. F. Jahr, J. Fuchs and R. Oberhauser, Chem. Ber., 1968, 101, 41 42 M. Henry, J. P. Jolivet and J. Livage, Structure Bonding, in the press. J. H. L. Watson, W. Heller, W. Wojtowicz, J. Chem. Phys., 1948, 16, 997. 477. 43 C. Sanchez, M. Nabavi, P. Judeinstein and S. Doeuff, J. Chim. 26 L. G. Hubert-Pfalzgraf and J. G. Riess, Znorg. Chim. Acta, 1980, 41, 111. Phys., 1989,86, 1593. Paper 1/00758K; Received 18th February, 1991

ISSN:0959-9428    

DOI:10.1039/JM9910100621

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

J. MATER. CHEM., 1991, 1(4), 629-635 Effect of Composition of Polymer Backbone on Spectroscopic and Electrochemical Properties of Ruthenium(l1) Bis(2,2'-bipyridyl)- containing 4=Vinylpyridine/Styrene Copolymers Dona1 Leech, Robert J. Forster, Malcolm R. Smyth and Johannes G. Vos* School of Chemical Sciences, Dublin City University, Dublin 9,Ireland A series of copolymers of 4-vinylpyridine and styrene and their corresponding ruthenium(i1) bis(2,2'-bipyridyl) metallopolymers has been prepared. The polymers have been characterised using IR UV-VIS and emission spectroscopy, differential scanning calorimetry and by microanalysis. The rate of charge transport through films of the metallopolymers on glassy carbon electrodes has been determined using chronoamperometry and cyclic voltammetry.The values obtained by these techniques for the charge-transport diffusion parameter and for the activation parameters are discussed in relation to the effect of styrene content in the copolymer backbone on the rate-determining step for charge transport. Keywords: Ruthenium; Modified electrode; Charge transfer; Polymer backbone Redox polymer-modified electrodes have been the object of active investigation in recent years because of their potential application in catalysis,'.2. photoelectrochemistry3~4 and mol- ecular electronics.* The main area of interest in these elec- trodes is as redox catalysts, where the polymer layer offers several advantages over monolayers, such as high local con- centration of catalytic sites and a three-dimensional reacting layer for homogenous catalysis.The applicability of the poly- mer-modified electrode will be limited by the rate of charge propagation through the film. This rate will be controlled by one of three processes: the intrinsic barrier to self-exchange for electron hopping between redox centres; counterion diffusion into/out of the film to maintain electroneutrality; polymer-chain juxtapositioning in order to allow electron exchange to occur. Ruthenium- and osmium-containing metal- lopolymers have received particular attention because of their potential photoelectrochemical and catalytic applications.6 -'' However, relatively little attention has been paid thus far to the effects of systematic structural variations in the polymer backbone on the charge-propagation rate and mechanism within these coatings.14, l9 In this paper we describe the synthesis and characterisation of a series of 4-vinylpyridine/styrene copolymers containing +[Ru(bipy),Cl] moieties (bipy =2,2'-bipyridyl) bound to the pyridine unit of 4-vinylpyridine. The structure of these metallopolymers, [Ru(bipy),(po1),,C1]C1, where pol refers to one monomer unit of the polymer, is depicted in Fig. 1. A recent study of the analytical application of the [R~(bipy)~(PVP),C11Cl polymer films as electrocatalysts [where PVP is poly(4-vinylpyridine)] found that these films were not sufficiently stable in the flowing solutions used (half- life of ca. 8 h),20 and further attempts to improve this stability, with coatings of both conducting and non-conducting poly- mers on top of the film, proved relatively unsuccessful.21 The incorporation of the styrene into the PVP backbone is expected to alleviate this problem.The analytical application and enhanced lifetimes of these copolymer-modified electrodes will be dicussed in a subsequent paper.22 In this study the charge-transport process for the Ru"/"' oxidation is investi- gated for glassy-carbon electrodes modified with thin films of these metallopolymers. Both potential-step techniques and cyclic voltammetry (CV) are utilised to probe the charge- transport process thoughout the whole of the layer, rather than just close to the electrode/layer interface. Activation parameters for the charge-transport process are also deter- r 1 r 1 I F' W Fig.1 Structure of the metallopolymers studied in this paper, where n+m=9 to give a polymer unit :metal centre ratio of 10: I mined to help distinguish between the different limiting pro- cesses. The results obtained are discussed with particular reference to the composition of the polymer backbone. Experimental Physical Measurements UV-VIS spectra were recorded on a Hewlett-Packard 342A diode array spectrophotometer. Emission spectra were recorded using a Perkin-Elmer LS-5 luminescence spec- trometer equipped with a red-sensitive Hamamatsu R928 detector. Spectra were recorded using an emission slit width of 10 nm at room temperature and 2.5 nm at 77 K, and are not corrected for photomultiplier response.Infrared spectra obtained as either KBr discs or from thin films on NaCl discs were recorded using a Perkin-Elmer 599 spectrophotometer. Differential scanning calorimetry (DSC) was performed on vacuum-dried samples using a Stanton Redcroft CPC 706 Temperature Programmer, a DSC lineariser and a d.c. ampli- fier. Experiments were carried out in static air at a heating rate of 5 "C min-' up to a temperature of 350 "C. Electrochemical measurements were performed using an EG and G PAR 273 potentiostat/galvanostat. Glassy-carbon electrodes of 7 mm diameter were modified by pipetting the required quantity of a 1% (m/v) solution of the metallo- J. MATER. CHEM., 1991, VOL. 1 Table 1 Microanalytical results for the synthesised copolymers theory found abbreviation abbreviation for PVP (Yo) for copolymer c (Yo) H (Yo) N ('10) c ("0) H ("10) N ('10) Ru-polymer 75 pvp75 78.03 7.17 9.42 67 pvp67 79.52 7.23 8.43 50 pvp50 82.60 7.34 6.42 33 pvp33 85.70 7.45 4.35 polymer directly onto their surface.The solvent was then allowed to evaporate slowly in a solvent-saturated chamber followed by air drying. All potentials are quoted versus a potassium saturated calomel electrode (SCE) without regard for liquid junction potentials. Surface coverages of ruthenium for the metallopolymer films were determined by the integration of the charge under slow scan (1 mV s-') cyclic voltammetric peaks. Surface coverages of 4-8 x lo-* mol cm-, were utilised throughout this study in order to minimise variations caused by varying coverages.Potential-step chronoamperometry was used to determine RUII/III charge-transport rates by stepping from 0.0 to 1.2 V (ca. 500 mV past the formal potential of the redox couple). These transient-current measurements were made over the timescale of 0-10 ms with a Philips 331 1 digital storage oscilloscope interfaced to a BBC microcomputer for data interrogation, and allowed signal-averaged results to be obtained. The experimentally determined parameter obtained in the study of charge-transport process is D:i2C (where D,, is the apparent charge-transfer diffusion coefficient and C is the concentration, in mol cm-3, of electroactive species on the electrode surface).This concentration may be estimated from polymer density measurements, determined by flotation in non-swelling solvents (in this case, carbon tetrachloride and hexane). Concentrations of 0.82, 0.81, 0.80, 0.79, and 0.78 mol dm-3 were estimated for Ru-PVPloo, RU-PVP~~, RU-PVP67, Ru-PVPSO and Ru-PVP~~,respectively (see Table 1 for abbreviation). Light was excluded from the cell for all of the electrochemical experiments to prevent the photosubstitution of the co-ordinated C1-with H20.I6 Elemental analyses were carried out at the Microanalytical Department of University College Dublin. Materials Bulk polymerisation of the polymers was carried out without regard for reactivity ratios in order to obtain higher-molecu- lar-weight materials.Poly(4-vinylpyridine) was prepared by bulk polymerisation of freshly distilled 4-vinylpyridine under nitrogen atmosphere using 2,2'-azoisobutyronitrile as initiator at 70 "C. The resulting polymer was fractionated by partial precipitation from methanol solution and the precipitant dried at 60 "C in uucuo overnight. Copolymers of 4-vinylpyridine and styrene were prepared as above with the monomers mixed in the desired molar ratio. The resulting polymers were purified by repeated precipitation in diethyl ether from 2-methoxyethanol and dried in uucuo (60 "C) overnight. Microanalytical data for the series of copolymers are given in Table 1. Theoretical values allowing for 0.5H20 molecules per pyridine unit were used, as samples of PVP have been known to retain water even upon rigorous drying.' [Ru(bipy),Cl,] .2H20 was prepared as described pre-viou~ly.~~ For [Ru(bipy),(pol)loCl]Cl, the copolymer (100 mg, 1 mmol) and the required weight of cis-[Ru(bipy),C1,]-2H20 77.69 6.82 9.66 Ru-PVP, 5 79.63 6.91 8.91 RU-PVP67 82.38 7.13 6.88 Ru-PVP,~ 85.91 7.33 4.63 Ru-PVP,, (51 mg, 0.1 mmol) were heated at reflux for 72 h with the course of the reaction being monitored by UV-VIS spec-troscopy and cyclic voltammetry. The product was isolated by precipitation into hexane and purified by repeated precipi- tation into diethyl ether from dichloromethane.Results and Discussion The synthetic approach used was to react the preformed copolymers with [Ru(bipy),Cl,].This allowed the metallo- polymers to be characterised both in solution and as thin films on the electrode surface. The formation of the metal- lopolymers is based on the well documented lability of the chloride ions in the [Ru(bipy),Cl,] complex. Extensive studies have shown that the first chloride ion is easily removed by refluxing in methanol or ethanol leading to substitution by polymer molecules [Ru(bipy),Cl,] +S-+[Ru(bipy),(S)Cl] + +C1-(1) [Ru(bipy),(S)Cl]+ +p~l,+[Ru(bipy)~(pol)~Cl]+ +S (2) (S =~olvent).'~*'~*~~,~~The solvent used in this study was 2-methoxyethanol, which is a better solvent for the copolymers than the lower alcohols.25 The metal loading of the metallopolymers is based on the quantity of starting material employed, assuming complete reaction.The validity of this assumption is supported by continuous monitoring of the reaction using spectroscopic and electrochemical techniques. A constant loading of 10 :1 polymer unit:metal centre (where the polymer unit can be either that of styrene or 4-vinylpyridine) was used to maintain an approximately constant mean metal-to-metal intersite dis- tance (assumed to be 2.5 nrn).', A list of the polymers exam- ined is presented in Table 1, together with the respective abbreviations for the free copolymers and their ruthenium complexes. The characterisation of this series of copolymers and their respective metallopolymers was accomplished by applying several techniques to each sample.Glass-transition Temperature The glass-transition temperatures (T,) of the polymers are presented in Table2. Thoroughly dried PVP has a Tp of ca. 142 "C, while polystyrene has a T, of ca. 100 "C, both of Table 2 Glass-transition temperatures determined by differential scan- ning calorimetry for the copolymers and their respective metallopo- lymers T,/ "C polymer polymer metallopolymer 140( 2) 224(5) 133(4) 216(5)131(3) 210(5) 126( 3) 209(5) 119(3) 208(5) J. MATER. CHEM., 1991, VOL. 1 which are largely independent of molecular weight.26 The linear increase in the Tp values of the copolymers with increasing 4-vinylpyridine content supports and reflects the experimentally determined copolymer composition. The much higher glass-transition temperatures for the met- allopolymers most likely reflect the greater difficulty in obtaining fluid-like motion because of a decrease in mobility upon co-ordination of the bulky metal centres.This increase in Tp upon metallation has also been reported for osmium analogues of the PVP homopolymer.'8 The trend in the Tp values for the metallopolymers reflects the trend associated with increasing 4-vinylpyridine content observed for the free copolymers. The small differences in transition temperatures determined for the metallopolymers (+25 "C) are an indi- cation of a similar metal-to-polymer subunit loading for all of the polymers. This is compared to the large differences in (180-230 "C) obtained for the osmium polymers with metal- to-polymer subunit loadings ranging from 1 :25 to 1 :5.18 IR Spectroscopy The spectra of the copolymers were a combination of all the bands present in poly-4-vinylpyridine and polystyrene spectra, and the presence of the chloride ligand in [Ru(bipy),(pol),,C1]C1 was confirmed by an M-Cl stretch-ing vibration at 330 cm-' for all of the metallopolymers.Absorption and Emission Spectroscopy The electronic spectra in 2-methoxyethanol for the homopoly- mers of styrene and 4-vinylpyridine both show strong overlap- ping n+n* transitions at ca. 220 and 256nm. The ratio between the absorbance of the copolymers at 256 nm and the absorbance obtained at this wavelength upon protonation of the pendant pyridines with hydrochloric acid gives an estimate of the 4-vinylpyridine content with the results obtained in good agreement with the expected 4-vinylpyridine content.The electronic spectra obtained for ruthenium 2,2'-bipyridyl containing polymers have been well doc~mented~.'~ and have proved to be a useful tool in the characterisation of such compounds. In particular, the energy of the lowest absorption maxima and the wavelength of emission are characteristic for a particular ruthenium moiety. According to the literat~re,,~ the visible absorption spectrum of Ru" diimine compounds is assigned to a spin-allowed metal-to-ligand charge-transfer transition ('MLCT) from the metal d orbitals to the n* orbital of the bipyridyl ligand while the emission is thought to occur from a ligand-based 3MLCT orbital. The metallopolymers studied here have two intense absorptions in the visible region, at 356 and 502nm.The UV-VIS spectrum of a model monomeric compound, [Ru(bipy),(py)Cl)Cl (where py rep- resents a pyridine nit)^^^^ shows two absorptions, at 352 and 496nm in ethanol. The slight red shifts for the polymeric compounds can be attributed to various factors, such as the steric effect of the polymeric ligand having a methylene chain as the backbone or solvent effects. A molar absorption coefficient of 8500 dm3 cm-' mol-' has been given for the lowest energy 'MLCT band of the model monomeric com- pound. The metallopolymer extinction coefficients are all approximately equal to this value, confirming the metal-to- polymer subunit loading of 1 : 10 calculated from the initial reaction concentrations of polymer and [Ru(bipy),Cl,]. The presence of an emission maximum for the metallopoly- mers, upon excitation at 500 nm, at 712+9 nm at room temperature and at 671 f:3 nm at 77 k are consistent with the presence of the [RuN,Cl] moiety.,'+-631 Electrochemistry The formal potentials for the RuI1/Ir1 oxidation in perchloric acid and sodium nitrate (pH 3.0) electrolytes (0.1 mol dm-3 and 1.0 mol dm-3 concentrations) were determined from the slow sweep (1 mV s-') cyclic voltammetric waves.For a glassy carbon electrode coated with the Ru-PVPloo metallopolymer formal potentials of 655 and 600mV were obtained in 0.1 mol dm -and 1.O mol dm -HC104, respectively, whereas potentials of 690 and 655 mV were obtained in 0.1 and 1.0 mol dm-3 NaNO,.These formal potentials were shifted positively in all electrolytes upon incorporation of styrene in the polymers to a maximum shift of 30 mV for the Ru-PVP,~ metallopolymers. This shift can be explained by the increas- ingly hydrophobic nature of the films with increasing styrene content. All of the above is evidence that the polymer backbone may be altered by the addition of styrene with little subsequent variation in the spectroscopic data or electrochemical formal potential for the metallopolymers. Charge-transfer Processes This section of the paper involves the systematic investigation of some parameters that affect the rate of charge propagation through thin films of the materials prepared on electrode surfaces.Factors such as electrolyte type and concentration and temperaure, as well as the 4-vinylpyridine mole ratio in the polymer backbone, have been varied. These variations are such that they enable information to be obtained as to the charge-transport mechanism and the factors which affect its rate. Potential-step chronoamperometry (CA) and cyclic voltam- metry (CV) were used to estimate DcI,the apparent charge- transfer diffusion coefficients for the Ru"/"' oxidation. In general, for the CA experiments, linear Cottrell plots were obtained for times up to ca. 10 ms. For these short timescales values determined for the charge-transfer diffusion coefficient reflect charge injection and ion motion in the initial 10-20% of the polymer layer.In the CV experiments sweep rates between 100 and 500 mV s-' were utilised giving linear peak current versus square root of sweep rate plots. This allows the use of Randles-Sevcik analysis for semi-infinite diffusion for the evaluation of Dc1.30 Charge-transfer diffusion coefficients were also estimated by CV using a theoretical approach devised by Aoki et al., which combines both surface Since the results and semi-infinite diffusional beha~iour.~' obtained for D,,from the two approaches were equal, within experimental error, only those obtained by Randles-Sevcik analysis will be presented. Longer timescale potential-step methods for the determination of D,,yielded results similar to those obtained by CV methods, indicative of the interfacial ion transport required upon oxidation of large proportions of the film.The evaluation of activation parameters for the charge transport process can be helpful in the identification of the rate-determining step in redox polymer films.12*32*33The activation energies for the [RU(~~~~)~(PVP)~CI]C~ homopoly-mer have been reported as 40 kJ mol-' for ion movement within the film, while for segmental polymer-chain motion activation energies of ca. 120 kJ mol-' are e~pected.'~,~~ These values are, however, electrolyte dependent. Charge-transfer diffusion coefficients and activation param- eters for the charge-transfer process have been determined in both 0.1 and 1.0 mol dm-3 sodium nitrate (PH 3.0) and per- chloric acid.Perchloric acid as an electrolyte is of interest because of the insolubility of the perchlorate salts of the metallopolymers and also because of the reported ion- pairing associations of the perchlorate ion, which can act as a cross-link, in poly~inylferrocene~~-~~and [0~(bipy)~(PVP)~~Cl]Clcoated17 electrodes. A study of charge transport in sodium nitrate was undertaken because of the use of this electrolyte, in previously published analytical work, on the [R~(bipy),(PVP)~C11Cl modified electrodes in flowing solutions.20*2' Sodium Nitrate Electrolyte Sodium nitrate electrolytes were adjusted to a pH of 3.0 to protonate the pendant pyridine units (pK, 3.3) and enhance ion transport.Interestingly, a polymer 'break-in' effect was observed for the low styrene content polymers. Fig. 2 shows repeated cyclic voltammograms (1 min delay between scans) for the Ru-PVPloo metallopolymer in 0.6 mol dmP3 NaN03. This effect is asumed to be due to polymer swelling upon protonation because it is not observed for the Ru-PVP~~ metallopolymer, and, indeed, the time required for stable cyclic voltammograms to be observed decreases with increas- ing styrene content. All measurements of charge transfer f ,I 0.0 0.6 1.2 potential/V Fig. 2 Repeated cyclic voltammograms of the Ru-PVP,,, metallo-polymer in 0.6moldm-3 NaNO, (pH 3.0). The arrows show the direction of growth. Delay between repeated scans= 1 min, with a scan rate of 100 mV s-.Ruthenium surface coverage=4.5 x lo-* mol cm-2 J. MATER. CHEM., 1991, VOL. 1 diffusion coefficients and activation parameters are therefore performed on films that have been previously 'broken-in'. The effect of electrolyte concentration and styrene content on the magnitude of D,,evaluated from both potential step and cyclic volammetry is presented in Table 3. The difference, by at least an order of magnitude, between the charge-transfer values obtained by potential-step and cyclic voltammetric methods most likely arises from the different timescales of the experimental techniques emp10yed.l~ In the CA experiment, ion motion within the film will be 'short range' with the charge-compensating counterion motions being localised in that region of the film at that experimental timescale.For the CV experiment interfacial charge transfer across the film/ electrolyte boundary will occur, and thus ion motions will be of a 'long-range' type, owing to the extent of the redox reaction. Thus, if ion motion and/or availability is the rate- limiting process for charge transport, the charge-transport diffusion coefficient determined by the short-timescale poten- tial-step technique would be expected to be higher than that determined by CV methods. In general, the errors in the charge-transfer diffusion coefficient values and in the activation parameters for the charge-transport process are quite large. The large errors arise, possibly, because of the inhomogeneity of the polymer films. This problem is further aggravated by a photosubstitu- tion process at the electrodes.Photochemical investiga-tions of glassy carbon electrodes modified with [Ru( bipy),( PVP)5 C1 ]Cll * 38 and [Ru( bipy),( PVP)5] C1238 films have reported the exchange of the co-ordinated C1- (or PVP in the case of the bis polymer) with H20: [Ru(bipy),(PVP),Cl] + + H20+ [Ru(~~~~),(PVP),(H,O)]~ + + Cl-[R~(bipy),(PVP),1~' + H,O+[Ru(bipy),(PVP) (H20)]2f+ PVP It is to be expected that the presence of even small amounts of the aqua complex will seriously affect D,,values. Values obtained for D,, are, however, still useful to study the trends observed when different parameters are varied. Effect of Electrolyte Concentration The effect of changing the sodium nitrate concentration from 0.1 tol.0 mol dmP3 on D,,is shown in Table 3.The relatively constant charge-transfer diffusion coefficients obtained in the 0.1 mol dm-3 electrolyte, by both CV and CA techniques for the metallopolymers, possibly reflects ion-transport limi-tations within the film at this low electrolyte concentration as the values are too low for electron hopping to be the rate- determining process.' For all of the polymers studied, a sharp increase in the charge transfer diffusion coefficient is obtained Table 3 The effect of electrolyte concentration and styrene content on the charge-transfer parameter of the metallopolymer-modified electrodes in sodium nitrate electrolyte (pH 3.0) polymer concentration/mol dm- 1O1'DcI/cm2s-l' 1O9DC,/cm2s-Ru- PVP 100 0.1 3.4(1.4)c l.O(O.4)c 1.o 14.3(5.0) 3.2( 1.2) Ru-PVP,, 0.1 3.8( 1.8) l.O(O.5) 1.o 25.1( 1.3) 1343.1) Ru-PVP~~ 0.1 3.5( 1.2) 1.2(0.5) 1.o 30.0( 10.1) 12.5(4.5) Ru-PVPSo 0.1 3.8(2.1) 1.5(0.9) 1.o 22.6(6.6) 10.1(5.8)Ru-PVP,~ 0.1 4.1(0.9) 1.7(0.5) 1.o 12.q5.0) 7.1(3.1) a Evaluated from cyclic voltammetry using the anodic peak current and the Randles-Sevcik equation; evaluated from chronoamperometry using the Cottrell equation; errors calculated as standard deviation from the mean of, usually, determinations on three different films.J. MATER. CHEM., 1991,VOL. I 633 using 1.0 mol dm- NaNO,. Since Di(2C is the experimentally determined parameter, with C equal to the fixed site concen- 50 5tration in the film, the increase observed for D,,with increased electrolyte concentration may reflect a variation in either the charge transport rate or the fixed site concentration.In the 7 40t T NaN0, electrolyte, at pH 3.0, the increased electrolyte con- centration is not expected to increase the fixed-site concen- tration because of electrostatic repulsion between the protonated unbound pyridine units in the film. Thus, it seems likely that the increase in D,,seen with increasing electrolyte concentration reflects an increased charge-transport rate. As 1the charge-transport rate is governed by counterion diffusional 10 -1 limitations in these polymers (see next section) any increase 7-in counterion concentration is likely to result in an increase 9 I I I 1 I in the experimentally determined charge-transfer diffusion coefficient. Activation parameters for the charge-transport process in NaNO, electrolytes are presented in Table 4.Activation ener- Fig. 3 Variation of the charge-transfer diffusion coefficient, Dct,with gies of 17-30 kJ mol-' together with negative entropy values mole fraction of PVP in the metallopolymer films as determined from are obtained in 0.1 mol dm-, NaN03 electrolyte. The nega- the anodic peak currents of the cyclic voltammograms and the tive entropy values are thought to be an indication of the Randles-Sevcik equation. a, 0.1 rnol dm-, NaNO,, pH 3.0; A, ordering of the polymer film upon transport of charge-1.0mol dm-, NaNO,, pH 3.0.Ruthenium surface coverages are 4-compensating counterions and solvent molecules.This trans- 8 x lop8rnol cmV2 port requires a certain degree of expansion of the polymeric lattice, which will result in a decreased polymeric chain data obtained in the 1.0 mol dm-, NaNO, electrolyte (Fig. 3) mobility and, by implication, an increase in the order of the show that a maximum D,, value is obtained for a styrene system. The activation parameters determined in content of ca. 35%. The initial increase in D,,upon addition 0.1 mol dm- electrolyte are consistent with ion-transport of styrene to the polymer backbone may be the result of an control within the [Ru(bipy),(pol),Cl]Cl film^.'^,^^ The acti- increase in the fixed-site concentration in the film.This vation energies obtained in 1.0 mol dm-, NaNO, for the possible increase is attributed to both a decrease in protonated metallopolymers will be discussed in the following section. pyridine sites and an increase in the film hydrophobicity associated with the styrene moieties, resulting in a more Eflect of Styrene Content compact film. The subsequent decrease in D,,observed with The effect of styrene content in the polymer backbone on D,, the higher styrene content polymers could be a reflection of for the CV and CA experiments is shown in Fig.3 and 4, film compaction to such an extent that ion transport across respectively. At 0.1 mol dm-3 NaNO, not much effect of the film/electrolyte interface becomes hindered. styrene content is observed, as discussed previously.The CV The activation parameters obtained from the CV experi- Table 4 Activation parameters for charge transport through the metallopolymer films in sodium nitrate (pH 3.0)and perchloric acid electrolyte polymer E,"/kJmol-' ASfa/J mol-' K-' AG"/kJ mol-' E,b/kJ mol-' ASfb/J mol-' K-' AG#b/kJmol-' NaNO, (0.1mol drn-,) Ru-PVP 100 -158(l0)C 58(2)' -109( 10)' Ru-PVP75 -140(11) 59(2) -88( 18) Ru-PVP,, -114(8) 59(2) -65(9) Ru-PVPso -105(12) 57(2) -103(8) Ru-PVP,~ -115(9) 57(3) -87(9) Ru- PVP 100 -1 l(9) -8(10) Ru-PVP,, -33(7) 11(8)Ru-PVP,, 20(8) -40(6) Ru-PVP~O 7(9) -93(11) Ru-PVP~~ 4(16) -86(9) Ru-PVPlO, 230(4) -91(4) Ru-PVP~, -15(3) -1 12( 1) RU-PVP67 -77(4) -116(4) Ru-PVPso -93(3) -106(5) Ru- PVP 33 -87(4) -119(5) Ru-PVPioo 20(2) 35(4)Ru-PVP~, -65(1) -20(3)RU-PVP6, -41(6) -58(6) Ru-PVP,~ -1 17(4) -88(2) Ru-PVP,, -102(2) -115(4) a Evaluated from cyclic voltammetric results; evaluated from chronoamperometric results; 'errors (+) calculated from the standard deviation from the slope and intercept of the best-fit line of all points.634 I T I O.i)rC, Q.2C 0.40 0.60 0.80 1.00 mole fraction of PVP Fig. 4 Variation of the charge-transfer diffusion coefficient, D,,,with the mole fraction of PVP in the metallopolymer films as determined from the Cottrell equation. W, 0.1 mol dm-, NaNO,, pH 3.0; A, 1.0 mol dmP3 NaNO,, pH 3.0. Ruthenium surface coverages are 4-8 x rnol cm-2 ments (Table4) are slightly higher than those obtained for [R~(bipy)~(PVP)~C11Clpolymers in HCl and HC104 electro- lytes.12 The activation energy values are thought to reflect interfacial ion-motion limitations, with counterion transport into the film being hindered.Given the errors associated with the determination of these parameters, no significant trend in those values in 1.0 mol dm-, NaNO, electrolytes determined by CV method are observed. Activation energies obtained from the potential-step results are lower than the CV values. This is most likely as being a result of the experimental timescale, with the short-range localised ion motion of charge- compensating ions on the potential-step timescale being more facile than the interfacial ion transport of counterions required in the CV experiments. Also the extent of reaction for the CV methods will be much greater than for the potential-step methods, thereby requiring a much greater flux of counterions to maintain electroneutrality.The activation energies deter- mined by potential-step methods decrease with increasing styrene content. This is possibly a reflection of the reported coupling of the NO, ion with protonated pyridine group- ing~,~~thus making the short-range ion motion more facile within films of the polymers with higher styrene content. The variation of D,,with styrene content for the 1.0 mol dmP3 NaN03 (Fig. 4) seems to reflect a trade-off between the two factors involved in ion-motion restriction, viz. film compaction with increasing styrene content because of increased hydro- phobicity and ion-motion restriction at higher 4-vinylpyridine content because of ion coupling of the nitrate with the protonated pyridine ions. Perchloric Acid Electrolyte As mentioned previously, the perchlorate salts of the ruthenium metallopolymers are less soluble in aqueous solu- tions than the chlorides, and the perchlorate ion has also been reported to act as a cross-link within polyvinylferro- cene33-35 and [Os(bipy)2(PVP)loC1]C117films because of ion- pair associations.D,,values determined in perchloric acid show little variation with changes in the polymer-backbone styrene content. Average values of (2.6k1.2) x10-l' cm2 s-l and (1.6_+0.8)~10-" cm2 s-l, by CV methods, and cm2 s-l and (3.9&2.0)~10-~(1.2_+0.5)~10-~ cm2s-', by potential-step methods, were obtained in 0.1 mol dm-, and 1.0 mol dm-3 HC104, respectively.For the CV results, the activation energy values determined show a decrease with J. MATER. CHEM., 1991, VOL. 1 increasing styrene content for both electrolyte concentrations. This decrease is most likely a reflection of an increase in the ease of ion-permeation into and motion within the film. This is explained by the fact that there are less protonated pyridine sites available for ion-pair formation and hence less cross- linking of the polymer, giving the styrene-containing metallo- polymers a less rigid conformation. The high activation energy value for the Ru-PVPloo metallopolymer films in 0.1 mol dm -HC104 is associated with an increasing disorder within the film (hence the positive entropy terms) and probably represents polymer-chain motion.12 It seems that the film is so compact, from cross-linking, that only by polymer chain motion can counterions be transported across the film/electro- lyte interface to maintain electroneutrality upon oxidation of the ruthenium redox centres.Activation parameters determined for the potential-step method indicate that in HC104 ion transport is the rate- determining step for these metallopolymer films at the CA timescales. The activation energies determined by CA decrease, in both concentrations of electrolyte, with increasing styrene content. This decrease is most likely a reflection of increasing ease of ion motion in the film because of the more open structure of the film.Thus, it is possible that the effect of styrene in perchlorate electrolyte is to improve ion transport in the films. Conclusions The synthetic approach used in this paper allows the prep- aration of various polymer backbones and their functionalis- ation with ruthenium redox centres which can be readily characterised. Physical measurements confirm that the excited state and redox properties of the metal centres are maintained upon polymer attachment. Charge-transfer diffusion coefficients and activation parameters for the charge-transport process within the metallopolymer films on electrodes vary with electrolyte composition and concentration, as well as the polymer backbone composition, with ion-transport limitations being the controlling factor of the process.The results obtained are of fundamental importance in establishing a link between D,, and the physical attributes of the polymer film. The successful sensor application of these metallopolymers as modifiers of electrodes is also dependent on the optimisation of the charge-transport rate and process. The investigation of charge transport through the osmium analogues of the poly- mers in this study is in progress. The reproducibility problems associated with the ruthenium polymers are not expected to occur with these osmium analogues. D. Leech gratefully acknowledges a Department of Science and Technology (Ireland) scholarship. EOLAS, the Irish Sci- ence and Technology Agency, is also thanked for financial support.Johnson Matthey are thanked for a generous loan of RuC13.xH20. References C. P. Andrieux, 0. Haas and J. M. Saveant, J. Am. Chem. SOC., 1986, 108. 8175. W. J. Albery and A. R. Hillman, J. Electroanal. Chem., 1984, 170, 27. D. A. Buttry and F. C. Anson, J. Am. Chem. SOC., 1982, 104, 4824. I. Rubinstein and A. J. Bard, J. Am. Chem. SOC., 1981, 103, 5007. C. E. D. Chidsey and R. W. Murray, Science, 1986, 231, 25. M. Kaneko. T. Okada, S. Teratani and K. Taya, Electrochem. Acta 1987, 32, 1405. K. Murao and K. Suzuki, J. Chem. SOC., Chem. Commun., 1984, 238. J. MATER. CHEM., 1991, VOL. 1 635 8 9 10 11 12 13 14 15 16 K. T. Potts and D. A. Usifer, Macromolecules, 1988, 21, 1985.J. M. Clear, J. M. Kelly and J. G. Vos, Makromol. Chem., 1983, 184, 613. J. M. Clear, J. M. Kelly, C. M. O’Connell and J. G. Vos, J. Chem. Res., (M), 1981, 3039. 0.Haas and J. G. Vos, J. Electroanal. Chem., 1980, 113, 139. M. E. G. Lyons, H. G. Fay, J. G. Vos and A. J. Kelly, J. Electroanal. Chem. 1988, 250, 207. J. M. Calvert and T. J. Meyer, Inorg. Chem. 1981, 20, 27. S-M. Oh and L. R. Faulkner, J. Am. Chem. Soc., 1989,111,5613. S. M. Geraty and J. G. Vos, J. Chem. Soc., Dalton Trans., 1987, 3073. 0.Haas, M. Kriens and J. G. Vos, J. Am. Chem. SOC., 1981,103, 26 27 28 29 30 31 32 33 Polymer Handbook. ed. J. Brandup and E. H. Immergut, Wiley, New York, 1975. F. E. Lytte and D. H. Hercules, J. Am, SOC. Chem., 1969, 91, 253. T.Shimidzu, K. Izaki, Y. Akai and T. Iyoda, Polym. J., 1981. 13, 889. B. A. Moyer and T. J. Meyer, Inorg. Chem., 1981, 20,444. A. Sevcik, Collect. Czech. Chem. Commun., 1948,44, 327. K. Aoki, K. Tokuda and H. Matsuda, J. Electroanal. Chem., 1983, 146,417. J. Q.Chambers and G. Inzelt, Anal. Chem., 1985,57, 1117. P. Daum, J. R. Lenhard, D. Rolison and R. W. Murray, J. Am. Chem. SOC., 1980, 102, 4649. 17 18 1318. R. J. Forster, A. J. Kelly, J. G. Vos and M. E. G. Lyons, J. Electroanal. Chem., 1989, 270, 365. R. J. Forster and J. G. Vos, Macromolecules, 1990, 23, 4372. 34 35 C. P. Andrieux. in Electrochemistry. Sensors and Analysis, ed. M. R. Smyth and J. G. Vos. Analytical Symposia Series Vol. 25, Elsevier, Amsterdam, 1986, p. 235. E. F. Bowden, M. F. Dautartas and J. F. Evans, J. Electroanal. 19 20 21 22 23 K. Sumi and F. C. Anson, J. Phys. Chem., 1986, 90, 3845. J. N. Barisci, G. G. Wallace, E. A. Wilke, M. Meaney, M. R. Smyth and J. G. Vos, Electroanalysis, 1989, 1, 245. G. G. Wallace, M. Meaney, M. R. Smyth and J. G. Vos, Electroanalysis, 1989, 1, 357. D. Leech, M. R. Smyth and J. G. Vos, to be submitted. B. P. Sullivan, P. J. Salmon and T. J. Meyer, Inorg. Chem., 1978. 17, 3334. 36 37 38 39 Chem., 1987, 219, 49. M. F. Dautartas, E. F. Bowden and J. F. Evans, J. Electroanal. Chem., 1987, 219, 71. E. F. Bowden, M. F. Dautartas and J. F. Evans, J. Electroanal. Chem., 1987,219, 91. 0.Haas, H. R. Zumbrunnen and J. G. Vos, Electrochim. Acta., 1985,30, 1551. P. Ferruti and R. Barbucci, Adv. Polym. Sci., 1984, 58, 55. 24 E. M. Kober, J. V. Casper, R. S. Lumpkin and T. J. Meyer, J. 25 Phys. Chem., 1986, 90, 3722. R. C. Sutton, L. Thai, J. M. Hewitt, C. L. Voycheck and J. S. Tan, Macromolecules, 1988, 21, 2432. Paper 1/00883H; Received 25th February, 1991

ISSN:0959-9428    

DOI:10.1039/JM9910100629

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

J. MATER. CHEM., 1991, 1(4), 637-642 Transformation of Schoepite into Uranyl Oxide Hydrates of the Bivalent Cations Mg2+,Mn2+ and Ni2+ Renaud Vochten," Laurent Van Haverbekeb and Roger Sobry" a Laboratorium voor chemische en fysische mineralogie, Universiteit Antwerpen (RUCA), Middelheimlaan I, B 2020 Antwerpen, Belgium Laboratorium voor anorganische scheikunde, Universiteit Antwerpen (RUCA), Groenenborgerlaan 171, 82020 Antwerpen, Belgium lnstitut de physique, Universite de I'etat de Liege, Sart Tilman, B 4001 Liege, Belgium The magnesium, manganese and nickel uranyl oxide hydrates have been synthesized by the reaction of the appropriate metal salts with schoepite at elevated temperature. After separation from the reaction mixture, the chemical composition of the different compounds was determined, and the chemical formulae were calculated based on different structural hypotheses.X-Ray powder diffraction data were obtained from each of the three species. After indexing and unit-cell parameter determination, the most plausible space groups were proposed, and compared with the known space groups of existing uranyl oxide hydrates. Thermal analysis was performed to determine the character and the number of water molecules in the structures. The solubility and solubility products were determined, and the influence of the different uranyl species was evaluated. Finally, to examine the presence of a charged double layer, the zeta potential was determined. Keywords: Uranyl oxide hydrate; Schoepite; X-Ray diffraction; Powder diffraction For the past few years, we have been studying the formation and the transformation of secondary uranium minerals in nature. In order to understand these phenomena more clearly, we have performed several syntheses of these minerals in the laboratory.Previously, a number of uranyl oxide hydrates (uranates) of the bivalent cations of calcium, barium, lead and strontium were synthesized in aqueous media at room temperature or elevated temperature. For these syntheses, the reader is referred to the papers of Potdevin and Brasseur,' Peters,2 Pr~tas,~ Bignand' and Brindley and Gillard and P~tdevin,~ Bastovanov.6 In a previous study,7 schoepite H30+[U020(OH)] was transformed directly into ca[(uo&o4(oH)6] 8H20 becquerelite, Ba[(U02),04(0H)6]*8H20 billietite and PbO*2U03 *2H20 wolsendorfite by treating with the appro- priate cations.In the present study, the uranyl oxide hydrates of Mg2+, Mn2+ and Ni2+ were synthesized in the same way by adding the cation solution to schoepite. The exchange of the H30+ ions by the bivalent cations was performed in order to give us a positive confirmation of the oxonium hypothesis of Brasseur.* It also indicates that this hypothesis is not restricted to naturally occurring minerals, but may be extended to chemical compounds that have not (yet) been found in natural formations. Experimental Schoepite was synthesized either by the reaction of C02,O2 and H20 on U3o8' or by the hydrolysis of uranyl acetate' in aqueous solution at 373 K.All other compounds were commercially available reagent-grade chemicals. The relative amounts of the bivalent cations were deter- mined by atomic absorption spectrometry. For this purpose, a Pye Unicam PU-9200 atomic absorption spectrometer was used, equipped with, respectively, an Mg, an Mn or an Ni lamp. Uranium was determined spectrophotometrically by means of Arsenazo 111. Measurements were performed with a Pye Unicam SP8- 100 UV-VIS absorption spectrometer at a wave- length of 662.5 nm. Thermal analysis was carried out using a DuPont 990 thermal analysis controller equipped with either a DSC-9 10 differential scanning calorimeter or a TG-95 1 thermogravi-metric analyser. A heating rate of 5 K min-' was used in combination with a nitrogen flow of 20 cm3 min-'.The densities were determined by measuring the buoyancy in toluene using a Cahn RG electrobalance. X-Ray powder diffraction data were obtained by means of a Guinier-Hagg camera, with a diameter of 100mm, using Cu-Kcr, radiation, 1= 1S406 A and silicon powder NBS-460 as an internal calibrant. Relative intensities were determined with a Zeiss MD-100 microdensitometer. Measurement of pH values was carried out with a Radi- ometer PHM-62 pH meter equipped with a standard com- bined glass electrode. The electrophoretic mobility of the particles in suspension was measured in a Rank Brothers Mark I1 cylindrical micro- electrophoretic cell with an applied magnification of 200.We used platinum electrodes, for which the polarity was reversed after each measurement. The calibration was carried out by means of a standard quartz suspension, with an accuracy of 1 mV. Results and Discussion The uranyl oxide hydrates were easily obtained by treating 1 g of schoepite with 100 cm3 of an 0.5 mol dm-3 solution of MgS04, MnSO, and NiS04, respectively, at 333 K over 2 weeks. The remaining solids were filtered off, and dried under atmospheric conditions; non-fluorescent powders were obtained for the magnesium (lemon-yellow), manganese (orange) and nickel (ochre) uranyl oxide hydrates. The results of the chemical analyses on the three compounds are given in Table 1, together with their formulae, their densities and the number of molecules in the unit cell. The formulae have been deduced from the oxidic composition, J. MATER.CHEM., 1991, VOL. 1 Table 1 Chemical composition of synthetic Mg, Mn and Ni uranyl oxide hydrates composition (wt.%) composition (XO :U03:H20) z value measured density X XO U03 H,O total calculated ideal cell volume/A3 (25 "C)/g ~rn-~ calculated ideal MgMn 2.10 6.40 86.00 83.00 11.60 10.50 99.70 99.90 1.03:5.98:12.79 0.94:3.02:6.03 1:0.6: 13 1 :3:6 260 1 .O 2593 5.128 5.298 4.03 7.98 4 8 Ni 6.70 82.96 9.63 99.29 0.94:3.02: 5.60 1 :3:6 2555.8 5.310 6.16 6 taking into account the structural formulae of protasiteg and that can be represented as [(U02)604(OH),]~n- and of billietite" and becquerelite.' [(U02)303(OH)2]in-.Brasseur' postulated the presence of According to Brasseur' and Sobry,12 the general oxidic oxonium and hydroxyl ions in uranates, and proposed a formula of the uranates can be expressed as: general formula for uranates mXO-2U03-(4-2m)H20 Oimi 2 X2+(H30+)2 -mC(UO2)202+,(OH), -mI 0 5m 5 2 Based on the oxidic composition, the studied uranyl oxide in which the H30+ ions are balanced by OH ions. hydrates can be represented as The presence of oxonium ions in the structure of hydrated +XO-2U03 *23H20 (X =Mn, Ni) uranates was clearly demonstrated uia NMR spectroscopy by Sobry.l2 Simple ion-exchange experiments proved that biva- and lent cations of uranates can easily be substituted by univalent fMgO.2U03 *3+H2O cations.Taking into account the structure of schoepite proposed by which results in the overall formulae Sobry,14 the formation of uranates of bivalent cations can be XO-3U03*4H20 (X=Mn, Ni) written as and MgO*6U03 10H20 The results from the chemical analyses in Table I show a higher water content. This may be regarded as zeolitic water. The results of the X-ray powder diffraction data show the The Mn and Ni compounds can be represented as remarkable fact that the nickel uranyl oxide hydrate is charac- X2 [(U02)303(OH)2]2-5H20 (X2+ =Mn2+,Ni2+) terized by fewer reflections than the magnesium and manga- + nese homologues. The first 20 diffraction lines of the Mn, Ni whereas the Mg compound can be represented as and Mg uranyl oxide hydrates are given in Table 2, together with their relative intensities and hkl values.From these data Mg2f[(U02)604(OH)6]2-10H20 it is clear that the three synthetic compounds can be easily These formulae meet the prediction of Evans13 who suggested distinguished from one another. that the uranyl oxide hydrates would show mainly pentagonal Since no single-crystal measurements could be carried out, co-ordination around the uranyl ion, with structures the unit-cell parameters were calculated from the powder [(UO2)O2(0H),] and [(UO2)O3(0H),]. These polyhedra give diffraction data using the method of Cox,l5 resulting in a rise to the formation of sheets" with a polymeric structure pseudohexagonal orthorhombic sublattice with approximate Table 2 X-Ray powder data of synthetic uranyl oxide hydrates of bivalent cations (Mn, Ni, Mg)" +Mn0.2U03 .23H20 +Ni0.2UO3 .25H20 3Mg0 .2U03 .3+H20 dabs hkl 1/10 dabs hkl 1/10 dabs hkl 1/10 7.5 1 002 65 7.5 1 002 65 7.5 1 002 50 6.16 210 15 3.872 031 10 6.22 020 10 5.76 112 25 3.746 131 65 5.7 1 21 1 50 4.718 103 10 3.606 104 25 4.790 212 10 4.6 13 013 5 3.487 230 90 3.872 213 10 3.872 213 20 3.237 402 30 3.734 123 30 3.734 004 50 3.156 232 100 3.585 032 100 3.565 400 95 2.553 234 100 3.486 132 10 3.536 230 100 2.494 225 40 3.220 402 100 3.393 223 25 2.402 342 15 3.095 420 10 3.220 313 100 2.333 610 10 3.021 42 1 50 3.188 024 100 2.03 1 515 80 2.846 3 14 40 3.065 420 15 2.019 623 85 2.797 240 10 3.036 040 40 1.977 451 I0 2.773 510 10 2.846 115 35 1.946 254 75 2.636 242 10 2.821 042 35 1.876 055 30 2.583 520 50 2.76 1 205 10 1.775 346 65 2.533 34 1 10 2.625 333 10 1.746 643 60 2.475 006 10 2.573 404 85 1.697 536 60 2.384 151 10 2.489 006 25 1.652 626 60 2.34 1 530 30 +67 additional lines +I7 additional lines +80 additional lines a Cu-Ka, radiation (A= 1.5406 A), 40 kV,20 mA.J. MATER. CHEM., 1991, VOL. 1 values of 0.701, 0.405 and 0.745 nm for a', b' and c', respect-ively, (a'xb'J3). The real lattice parameters a, b and c, corresponding to a larger orthorhombic cell, can be obtained by multiplying a' by 2, b' by 3 and c' by 2. The obtained values are summarized in Table 3. These cell parameters are in good agreement with the values obtained for other uranates by Sobry.I2 When these parameters were used in the computer program of Visser,16 all lines could be indexed with an accuracy of 0.5 pm.The obtained hkl values for the three species agree with the reflection conditions of the space groups Pnma for the manga- nese and nickel compounds and Pbn2, or Pb21 for the magnesium uranyl oxide hydrate. Based on these results, we assume that the compounds belong to these space groups. The dehydration of the three compounds was studied by thermogravimetry and differential scanning calorimetry. The combination of the two techniques shows for the three com- pounds four well separated endothermic dehydration steps. Table4 lists the four temperature ranges and the number of water molecules lost in each range for the three species.The dehydratation mechanism is extensively discussed by Sobry.12 Taking into account the general formula mXO .2U03 '(4 -2rn)H20 and the total amount of water molecules lost, it is obvious that some water molecules are not structurally bound. They may be regarded as surface- bound or zeolitic-bound water molecules. This zeolitic water in the magnesium and manganese compounds is lost below 400 K and in the nickel compound below 425 K. The remain- ing dehydration steps agree within 25% with the temperature intervals proposed by Sobry.12 The solubility of the three compounds in aqueous medium was measured after equilibration of the solid phases at different pH values at 303 K over 2 weeks. The pH was adjusted to the desired value with nitric acid or sodium hydroxide.After equilibration, the suspensions were centrifuged, the pH meas- ured accurately and the metal-ion concentrations determined. From the metal-ion concentrations, the solubilities of the three compounds were calculated. The results are given in Table 5, together with the pH values of the solutions. Taking into account the oxidic formulae of the compounds, the dissolution can be expressed as +nX, + +pUO:+ +2(n +p)OH- leading to a solubility product which can be expressed as In order to calculate the solubility product, the exact concen- tration of each ion must be known accurately. There are no problems for the metal ions because they exist only as free ions in solution. This is also the case for the hydroxyl ions, since the pH value is known precisely.However, this is difficult for the uranyl ions, because of the various complexes that are formed between uranyl and hydroxyl ions. The formation of these complexes can be described generally as pUO$++qH20-+(U02)p(OH)~P-4)++4H+ These reactions have an associated formation constant which can be expressed as [(U02)p(OH)fP-q)+][H 'Iq PP, = [uo;'3" The most relevant uranyl-hydroxyl formation constants have been subtracted from Sillen and Martell," Perrin" and Hogfeldt." Using these values, we can express the total uranyl concentration as follows 843[U0,]=3 +4+ 853 [uof+]3i -1 CH 1 CH+I5 Table 3 Unit cell parameters in %i of synthetic uranyl oxide hydrates of Mn, Ni and Mg" m xo n a Aa b Ab C Ac unit cell volume/A3 3 MnO 84 14.2 198 0.0078 12.1880 0.0075 14.9623 0.0085 2593.1 3 NiO 37 14.2860 0.0 165 12.0283 0.0 154 14.8879 0.01 70 2555.8 3 MgO 88 14.2818 0.01 15 12.2647 0.0107 14.8610 0.01 17 2603.0 Aa, Ab and Ac represent the maximum deviations of the cell parameters; n =number of reflections used in lattice parameter calculations.Table 4 Dehydration data of synthetic Mg, Mn and Ni uranyl oxide hydrates Mg uranate Mn uranate Ni uranate temp. range/ "C moles H20 lost temp. range/ "C moles H20 lost temp. range/ "C moles H20 lost <125 3.06 <125 1.78 <175 1.92 125-250 4.38 125-150 2.45 175-250 2.02 2 50- 3 60 2.74 150-450 1.53 250-400 1.38 350-600 2.6 1 450-600 0.27 400-600 0.28 Table 5 Solubilities at 25 "C of Mg, Mn and Ni uranyl oxide hydrate at different pH values Mg uranyl oxide hydrate Mn uranyl oxide hydrate Ni uranyl oxide hydrate PH solubility/10-5 mol dm-3 PH solubility/10-5 mol dm-j PH solubility/10-5 mol dm-3 3.91 254.00 4.25 78.23 4.10 160.00 4.40 43.13 4.35 30.20 4.27 64.70 4.73 4.93 5.30 6.98 5.88 9.20 5.52 2.46 6.19 2.00 5.83 2.2 1 6.07 1.06 6.5 1 0.78 6.28 0.50 6.43 0.20 6.61 0.47 6.60 0.20 Since the total uranyl concentration can be derived from the metal-ion concentration, the concentration of the free uranyl ions can be determined at a given pH.After introducing these values in the expressions of the solubility product, we obtained 156.3, 86.2 and 87.7 for the pK,, values of the magnesium, manganese and nickel uranyl oxide hydrate, respectively, with an accuracy of 4.Since the number of ions in the expression of the solubility is very large, it is clear that the pK,, values must be interpreted with some caution. For example, the magnesium uranyl oxide hydrate has a solubility product which can be expressed as K,, =[Mg' '1 [UO? l6 [OH-]'4+ J. MATER. CHEM., 1991, VOL. 1 Indeed the solubility product is largely influenced by the factor [OH-]. If one assumes pH 7, it is clear that in this case the solubility product is influenced by since [OH-] must be written as [OH]'4. Therefore, these extremely low solubility products may well be responsible for relatively moderate solubilities. Inverse calculations were also performed.Given the pK,, of the compound, the relative contribution of the different uranyl species to the total uranyl content is determined in the pH range 2-12. Fig. l(a), (b)and (c) show these contri- butions for the magnesium, manganese and nickel uranyl oxide hydrates, respectively. All three graphs show the same 7 14 PH 7 14 PH Fig. 1 J. MATER. CHEM., 1991, VOL. 1 641 0 7 14 PH Fig. 1 Distribution of the different uranyl species between pH 2 and 12 for a saturated solution at 25 "C for (a) magnesium, (b)manganese and (c) nickel uranyl oxide hydrates tendencies. Above pH 9, all three compounds show a domi- nant presence of (UO,)(OH)+. In the pH range 2-8, the presence of UO; +,(UO,),(OH);+ and (UO,),(OH); varies according to the sequence Mg-Ni-Mn.From these graphs, it is clear that the large difference in solubility product does not cause large differences in the uranyl species distribution, but is mainly a consequence of the chemical composition. In order to determine the variation of the surface charge of these compounds us. the pH, the zeta potential was meas- ured. In order to avoid the formation of negatively charged uranyl carbonato complexes, the suspensions of very-fine powder particles (35-75 pm) were shaken over a period of 1 week in a nitrogen atmosphere. The measurements of the zeta potential were carried out also in a nitrogen atmosphere. The zeta potential (expressed in mV) of an electrical double layer is 112890 D +Iu with D the relative permittivity of the solvent, +I the viscosity (in lo-' Pa s) of the solvent and u the electrophoretic mobility of the particles in the suspension (in pm cm s-' V-I).Fig. 2 shows the variation of the zeta potentials of the three species us. the pH. It is clear that the surface of the uranates becomes more negatively charged with increasing pH, prob- ably caused by an increased static presence of hydroxyl ions at the surface of the particles. The variation of the zeta potential as function of pH is more pronounced for the nickel and manganese compounds than for the magnesium one. The points of zero charge have been determined at pH values 4.10, 4.25 and 4.40 for the magnesium, nickel and manganese compounds, respectively. Conclusions The fact that the magnesium, nickel and manganese uranyl oxide hydrates are synthesized by simple exchange reactions between schoepite and the appropriate metal ions indicates that in schoepite a positively charged ion other than uranyl -30--10-0-L PH Fig.2 Variation of the zeta potential as function of the pH at 25 "C for (a)nickel, (b)manganese and (c) magnesium uranyl oxide hydrates is present. This can be seen as a confirmation of the oxonium hypothesis of Brassew.* Because the chemical composition, the unit-cell parameters, the space groups and the thermal behaviour all agree very well with bivalent uranyl oxide hydrates occurring in natural environments, it may be concluded that they are structurally comparable. The solubility product of the magnesium com- pound is of the same order of magnitude as those of the calcium and barium uranyl oxide hydrates. The solubility products of the transition metals nickel and manganese are much higher.This can be explained by the different chemical composition of the latter compounds. Because of the anal- ogous uranyl distribution pattern, it may be concluded that this is not an indication of different behaviour. There is a striking similarity between all uranyl oxide 642 J. MATER. CHEM., 1991, VOL. 1 hydrates studied in this and previous papers. Therefore, it is very remarkable that in the series of uranyl oxide hydrates that we synthesized, the magnesium, manganese and nickel uranyl oxide hydrates have not yet been found in nature.Our experiments do not indicate any reason why these compounds 4 5 6 7 S. Gillard and H. Potdevin, Bull. SOC. R. Sci. Liige, 1959, 28, 222. C. Bignand, Bull. SOC.Fr. Mineral. Crist., 1955, 78, 1. G. W. Brindley and M. Bastovanov, Clay Clay Minerals, 1982, 30, 135. R. Vochten, E. De Grave and H. Lauwer, Mineral. Petrol., 1990, do not occur in a natural environment. Since magnesium ions in particular are present in a number of uranium deposits, the question why these uranyl oxide hydrates do not occur remains open. 8 9 10 41, 247. H. Brasseur, Bull. SOC.Fr. Mineral. Crist., 1962, 85, 242. M. K. Papagoa, D. E. Apleman and J. M. Stewart, Mineral. Magaz., 1986, 50, 125. J. Piret-Meunier and P. Piret, Bull. Mineral., 1982, 105, 606.11 M. K. Papagoa, D. E. Appleman and J. M. Stewart, Am. Mineral., The authors wish to thank the National Fund for Scientific Research (Belgium) for financial support. They are grateful 12 13 1987, 72, 1230. R. Sobry, J. Znorg. Nucl. Chem., 1973, 35, 1515. H. T. Evans Jr., Science, 1963, 141, 154. for the technical assistance of Mr. K. Van Springe1 and Miss A. Vanysacker. 14 15 R. Sobry, Am. Mineral., 1971, 56, 1065. A. A. Cox, A Program for Least-Squares Rejnement of Unit Cell Dimensions, The City University of London, 1967. 16 J. W. Visser, J. Appl. Crystallogr., 1969, 2, 89. 17 L. G. Sillen and A. E. Martell, in Stability Constants of Metal- References 18 ion Complexes, The Chemical Society, London, 1964, pp. 50-51. D. D. Perrin, in Ionisation Constants of Inorganic Acids and Bases 1 H. Potdevin and H. Brasseur, Bull. Acad. R. Belg. Cl. Sci., 1958, 44, 874. 19 in Aqueous Solution, Pergamon Press, Oxford, 1982, pp. 122-124. E. Hogfeldt, in Stability Constants of Metal Zon Complexes- 2 J. M. Peters, Bull. SOC.Sci. Liige, 1965, 34, 308. 3 J. Protas, Bull. SOC. Fr. Mineral. Crist., 1956, 79, 350; J. Protas, C.R. Acad. Sci. Paris, 1957, 244, 91; J. Protas, Bull. SOC. Fr. Part A: Inorganic Ligands, Pergamon Press, Oxford, 1982, pp. 40-4 1. Mineral. Crist., 1959, 82, 239. Paper 1/01897C; Received 25th February, 1991

ISSN:0959-9428    

DOI:10.1039/JM9910100637

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

J. MATER. CHEM., 1991, 1(4), 643-647 643 Synthesis, Characterization and Biodegradation Test of Nylon 2/6 and Nylon 2/6/6 K.E. Gonsalves,a X. Chen" and T. K.Wongb a Polymer Science Program, Institute of Material Science and Department of Chemistry, U-60, University of Connecticut, Storrs, CT 06269, USA Enzymatics lnc., Horsham, PA 19044, USA Initial results obtained from the characterization and biodegradation tests on poly(6-glycylaminohexanoic acid) (nylon 2/6) and poly[hC(6-glycylaminohexyl)adipamicacid] (nylon 2/6/6) are presented. Nylon 2/6 was synthesized via a condensation reaction and nylon 2/6/6 by interfacial polymerization. Their intrinsic viscosities were 0.03 and 0.09 dm3 g-', respectively. Thermal analysis and X-ray diffraction methods were applied to characterize these two nylons.The biodegradation of nylon 2/6 and nylon 2/6/6 films was observed by using a qualitatively fast screening test, which showed that nylon 2/6/6 is more readily degradable than nylon 2/6 under attack by fungi. Keywords: Nylon; Biodegradation; Intrinsic viscosity 1. Introduction A series of glycine-containing nylons, such as poly(6-glycylaminohexanoic acid) (nylon 2/6), poly[N-(6- glycylaminohexy1)adipamic acid] (nylon 2/6/6) and poly [N-(6-glycylaminohexylaminocarbonylmethyl)adipamicacid] (nylon 2/6/2/6), were reported to be biodegradable based on the measurement of the release of carbon dioxide from slurry samples under attack by fungi and bacteria.'v2 It therefore appears that these biodegradable nylons have the potential to be used in replacing commercial nylons in packaging and marine applications.Hence, it was considered necessary to conduct investigations on the thermal and physical character- istics of these polymers to assess their processability, as well as to determine their degradability in end products such as films and fibres. We reported previously an alternative syn- thesis of nylon 2/6.3 In this paper, we report a modified synthesis of nylon 2/6/6, the thermal and X-ray diffraction characterizations of nylon 2/6 and nylon 2/6/6, as well as a fast screening test to ascertain the biodegradability of films of these gl ycine-containing polymers. 2. Results and Discussion 2.1 Synthesis 2.1.1 Nylon 2/6 (fNHCH2CONH(CH2),C~,) As described previ~usly,~ nylon 2/6 was made uia conden-sation of a pentachlorophenol (PCP) active ester.Its molecular weight can be inferred to be low because of the low intrinsic viscosity [q] of 0.029 dm3 g- in 90% formic acid. Note that in the earlier synthesis of nylon 2/6,4i5 low [q] values were also reported. In order to increase the molecular weight of nylon 2/6, different substituted phenols were used in the process of active-ester formation. As shown in Table 1, all of the three active esters gave polymers with similar values of [q1.It can be qualitatively concluded that only relatively low-molecular- weight nylon 2/6 can be obtained uia these condensation polymerizations. 2.1.2 Nylon 2/6/6(fHNCH2CONH(CH2)6NHCO(CH2)4 CWn) This polymer was first synthesized by Wu6 with an intrinsic viscosity of 0.086 dm3 g-' in dichloroacetic acid at 298.2 K by the following method (Scheme 1, note tert-BOC is the tert-butyloxycarbonyl group).By applying a dicyclohex ylcarbodiimide (DCC) peptide-linking method instead of the slow (1 week) aminolysis reaction, we synthesized nylon 2/6/6 with [q] of 0.093 dm3 g-' in 90% formic acid at 298.2 K as shown in Scheme 2. 2.2 Characterization 2.2.1 Diflerential Scanning Calorimetry, Thermogravimetric Analysis and Thermal Chromatography-Mass Spectrometry For comparison, poly[N-(6-aminohexyl)adipamic acid] (nylon 6/6) was prepared by the same method used for making nylon 2/6/6 described in section 3.3. The <(glass transition) and T, Table 1 Intrinsic viscosities [q] of nylon 2/6 condensed from different active esters phenol [ttiw3g -0.029 ~ ~ ~ ~ ~ ~0.020 $ ~ ~ pen tafluorophenol 0.034 a In 90% formic acid.rerr-BOC azide ClCH COCI H2N(CH2),NH2 -tert-BOC-NH(CH2),NH2 NHJH20 conc. HC1 tert-BOC-NH(CH,),NHOCCH2cl-tert-B0C-NH(CH2),NHCOCH2NH2- J. MATER. CHEM., 1991, VOL. 1 di-fert-butyl pyrocarbonate H2N(CH 2) 6NH 2 + tert-BOC-NH(CH2),NH2HC1 tert-BOC glycine, DCC ' tert-BOC-NHCH2CONH(CH2)6NH-tert-BOC HCl/acetic acid ClH,NCH2CONH(CH2)6NH3CI (melting point) of the former were shown to be 338 and 538 K, respectively. As seen in Table2, the Tg and T, of nylon 2/6 are close to those of nylon 6/6, whereas those of nylon 2/6/6 are apparently lower.It is shown in Table2 that the onset decomposition temperatures of both nylon 2/6 and nylon 2/6/6, measured by thermogravimetric analysis (TG), are 50 K higher than their melting points. To elucidate further the thermal stabilities of these two nylons, the thermal chromatog- raphy-mass spectrometry technique (TC-MS) was applied. Details of the TC-MS instrument have been described pre- vio~sly.~Here, only the results from level I-FID analysis of 3000 2 2000 i:0 LL 1000 -/,L0-'-1 ', TC-MS (i.e. total gas production us. temperature) are dis- cussed. The experimental runs were conducted in helium (20 cm3 min-') at a heating rate of 30 K min-' from 303 to 873 K. As shown in Fig. 1, the onset temperatures for gas product formation from nylon 2/6/6 and nylon 2/6 are ca.620 and 570 K, respectively. Nylon 2/6 shows a small decompo- sition peak also at ca. 420 K. Both TG and TC-MS results, based on the measurements of weight loss and gas product formation, respectively, indicate -303 423 573 723 873 that these nylons decompose only at temperatures higher than 570 K. As mentioned above, the melting points of nylon 2/6 and nylon 2/6/6 are 549 and 501 K, respectively. Therefore, it should be possible to process these two nylons by regular approaches such as melt-moulding or melt-spinning. However, nylon 2/6 was discoloured after melt-moulding at 550 K in N2. The film and fibre samples of nylon 2/6/6 were also slightly discoloured upon moulding and melt-spinning at 500 K.Further studies on the thermal decomposition modes of these glycine-containing nylons are therefore necessary and will be reported separately. 2.2.2 X-Ray Difruction Preliminary powder X-ray diffraction patterns of nylon 2/6 and nylon 2/6/6 samples, crystallized from formic acid-n- butanol at 368 K, are shown in Fig. 2. A strong reflection at spacing 4.15A was observed, which is the same as that reported by Puigalli et al.' for nylon 2/6. This result probably indicates that both of these nylons have hexagonal crystal structures as proposed by Puigalli et aL5 Table 2 Physical properties of nylon 216 and nylon 21616 property nylon 216" nylon 21616 method IIr11/dm3g-0.029 0.093 in 90% formic acid, 298.2 K TB/K 333 327 DSC, 15 K min-' TmIK 549 501 DSC, 10 K min-' Gb/K in N, 599 611 TGA, 20 K min-in air 600 601 a Nylon 216 was synthesized from PCP ester; Td is the onset decomposition temperature measured by TGA.L g 2000 -1000 . i Oh ' ~303 423 573 723 873 l------cell temperature/K cool pyrocell Fig. 1 FID response of nylons by TC-MS analysis: (a)nylon 216, (b) nylon 21616 2.3 Biodegradation Test Poly[oxy( 1-oxohexamethylene)] film was tested as a positive control in this experiment. After 35 days of incubation, it became so fragile that it could hardly be picked up from the Petri dish for cleaning and SEM observation. The degradation on nylon films was not detectable visually. Biodegradation invovles the loss of structural characteristics and mass of a material as C02 and water-soluble components.There are many methods to test the biodegradation of syn- thetic Most measurements are based on C02 production, O2 consumption, weight loss and changes in structural characteristics. As pointed by Wool and Cole" most of these methods are technically difficult and may yield inconclusive or misleading results. In this experiment, we estimated the surface erosion of nylon films by using a qualitative scanning electron microscopy method. As seen in Fig. 3, nylon 6 maintained a smooth surface after 14 days of incubation, while obvious degradation occurred on the surface of nylon 2/6/6 after 31 days of incubation. These results confirmed the fact that the commercial nylons like nylon 6 are non-biodegradable, and also indicated that nylons con- J.MATER. CHEM., 1991, VOL. 1 (a) Fig. 2 Powder X-ray diffraction patterns of nylon (the innermost ring corresponds to the spacing 4.15 A): (a)nylon 2/6; (b)nylon 2/6/6 (a1 (b) Fig. 3 SEM pictures of nylon film surface, x 10 000: (a)nylon 6 after 14 days of incubation; (b)nylon 2/6/6 after 31 days of incubation taining glycine units in the backbone could be potentially biodegradable. Fig. 4 and Fig. 5 show the time progression of surface erosion on nylon 2/6 and nylon 2/6/6 films caused by fungal attack. From these SEM micrographs it was observed that nylon 2/6/6 degrades more readily than nylon 2/6 under the conditions mentioned above. 3. Experimental 3.1 Synthesis of Nylon 2/6/6 3.1.1 Protection of Hexanediamine with tert-BOC Group A solution of di-tert-butyl pyrocarbonate (43.6 g, 0.2 mol) in dioxane (70 cm3) was added slowly to a solution of hexanedia- mine (75 g, 0.65 mol) and NaOH (8 g, 0.2 mol) in 1 : 1 H20-dioxane (220 cm3) held in an ice-water bath while stirring.After ca. 3 h addition, the solution was allowed to warm to room temperature and was left overnight. Dioxane was removed by vacuum evaporation, and the residue solution was saturated with NaCl and then extracted with EtAc. The oily material, obtained by removing EtAc from the organic phase, was dissolved in H20 (100 cm3) and acidified to pH 3-4 with 1 mol dmP3 HCl. A white precipitate was formed on the top of the solution after it was saturated with NaCl.The solid was dissolved in hot ethanol. Three times its volume of ether was added to this solution and a white solid was formed. 21 g (83.1 mmol, 41.3%) of solid tert-BOC-HN(CH2),NH3C1 (N-tert-BOC-hexanediamine hydrochloride, I) was obtained upon vacuum drying, m.p. 433-435 K. 3.1.2 N-Glycylhexanediamine Dihydrochloride Solid I (21 g), tert-BOC-glycine (14.6 g, 83.5 mmol), freshly distilled dicyclohexylcarbodiimide (DCC) (1 7.4g, 83.5 mmol) and 1,4-diazobicyclo[2.2.2]octane (7.9 g, 70 mmol) were dis- solved in 400cm3 of dry EtAc in an ice-water bath while stirring. The mixture was stirred at room temperature over- night followed by addition of 5 cm3 glacial acetic acid to decompose the excess of DCC.After cooling in ice water for 30 min, the mixture was filtered. The filtrate was washed thoroughly with 40 cm3 cool 1 mol dmP3 HCl, water, satu- rated NaHCO, and water. The solution was dried with Na,SO, and the solvent removed by rotory evaporation under reduced pressure. To the oily residue of tert-BOC-HN(CH2),NHOCCH2NH-tert-BOC (11), 120 cm3 7% HC1-acetic acid was added to remove the blocking tert- BOC groups. A white precipitate was formed after adding 400 cm3 acetone. The solid was recrystallized from 1 : 1 meth-anol-acetone giving 15.7g (63.6 mmol, 76.6%) C1H3NCH2CONH(CH2),NH3C1 (N-glycylhexanediamine dihydrochloride, 111), with m.p. 433-434 K. 3.1.3 Nylon 21616 via Interfacial Polymerization In a 1 dm3 capacity blender, 15.7 g of solid I11 and 10.4 g (0.26 mol) of NaOH were dissolved in 400 cm3 H20.To this solution, 150 cm3 of a solution of 9.45 cm3 (65 mmol) adipoyl chloride in CCl, were added slowly with vigorous stirring. The nylon 2/6/6 solid was washed successively with many portions of H20, 1 :1 EtOH-H20 and EtOH, and then dried at 363 K in a vacuum oven overnight. 8.15 g nylon 21616 (yield 45.2%) were obtained. Its intrinsic viscosity in 90% formic acid at 298.2 K was 0.093 dm3 g-'. Elemental analysis results for the polymer are as follows. Found: C, 57.79; H, 9.22; N,14.36%. Calc. for [C14H25N303]n: C, 59.34; H, 8.89; N, 14.83 and for [C14H25N303-H20]n: C, 56.47; H, 9.01; N,14.11%. 3.2 Characterizations The intrinsic viscosities [q] of nylon 2/6 and 2/6/6 were measured in 5 g dm-, solution of 90% formic acid at 298.2 K.No change in [q] was observed after 1 week for both nylons. Samples for powder X-ray diffraction were prepared as fol- Fig. 4 SEM pictures of biodegraded nylon 2/6 film, x 5000: (a) before test; (b)7 day test; (c) 31 day test n-butanol was added to an equal volume of a 0.2% polymer solution of 90% formic acid at 368 K followed by cooling slowly to room temperature overnight. The crystals were separated by centrifugation, washed twice with n-butanol, and finally dried in a vacuum for 2 days. Powder X-ray diffraction diagrams were recorded using Cu-Ka radiation with an Ni filter at 30 kV and 20 mA. Perkin-Elmer DSC-2 and Perkin-Elmer TGS-2 instruments were used for DSC and TG measurements, respectively.TC-MS runs were made in helium using the Pyran System (Ruska Labs).’ 3.3 Biodegradation 3.3.1 Samples and Cultures The cultures used were Aspergillus niger, Aspergillus Jauus, Penicillum notatum, Saccharomyces cereuisiae, Sordaria Jimic- ola and Schizosaccharomyces octospores. J. MATER. CHEM., 1991, VOL. 1 Fig. 5 SEM pictures of biodegraded nylon 2/6/6 film: (a) before test, x 10 000; (b)after 7 day test, x 5000; (c) 14 day test, x 5000 Films of nylon 2/6/6 and nylon 6, with a thickness of ca. 0.1 mm, were moulded at 500 K by fast heating and cooling within 1 min. Nylon 2/6 was moulded at 550 K and turned yellow. Poly[oxy( 1-oxohexamethylene)], which is a well known biodegradable polymer, was moulded at 340 K to be used as a positive control, and nylon 6 film as a negative control.3.3.2 Test Procedures After each fungal culture was grown in its own medium in agar for 1 week, one loopful was transferred to a tube containing 1 cm3 of liquid nutrient medium. This mixed fungal culture was incubated at 298 K for 1 day. Liquid nutrient agar was poured into a Petri dish and cooled to a semi-solid state, and a piece of sample film (ca. 0.5 cm2) was placed onto the agar <0.5 mm below the surface. After the agar was solidified, the mixed cultures were laid directly on top of the film area. Incubation was performed at J. MATER. CHEM., 1991, VOL. I 298 K in a humidified environment for 7, 14 and 31 days. During the incubation, the growth of the cultures was observed by optical microscopy.After incubation, each film was subjected to soaking in 5moldm-3 NaOH and clorox solution, and then washed thoroughly with distilled water. Scanning electron microscopy (SEM) was used for testing the surface erosion by the cultures in order to assess qualitatively the degree of biodegradation. 4. Conclusions Only a low molecular weight of nylon 2/6 can be obtained from polycondensation reactions of azide or active ester, whereas, by interfacial polymerization, nylon 2/6/6 with a moderately high molecular weight can be made. Here, a fast method for making the diamine monomer of nylon 2/6/6 was developed. Although the thermal analysis (TG and TC-MS) showed that nylon 2/6 decomposes only at temperatures 50 K beyond its melting point, the nylon 2/6 sample was noticeably degraded after melt-moulding.Nylon 2/6/6 also showed slight discolouration after melt-moulding or melt-spinning at 500 K. The difference of thermal stability between nylon 2/6 and nylon 2/6/6 upon melt processing can be explained from the fact that both contain a glycine group which is assumed to be relatively thermally unstable and the melting temperature of nylon 2/6/6 is lower than that of nylon 2/6 (see Table 2). Results from the fast screening test of biodegradation described here show that nylon 2/6 and nylon 2/6/6 are biodegradable. This method can be applied as a general approach to measure the biodegradability of other polymer fiims and fibres qualitatively.Refinement of the tests are underway and will be reported subsequently. References I D. Ennis and A. Krammer, J. Food Sci., 1975,40, 181 2 W. J. Bailey and B. Gapud, Ann. N.Y. Acad. Sci., 1985, 446,42. 3 K. E. Gonsalves and X. Chen, Polym. Commun., 1990, 31, 3 12. 4 W. J. Bailey, Y. Okamoto, W. Kuo and T. Narita, Proc. 3rd Znt. Biodegradation Symp., Applied Science Publishers, Barking, 1976, p. 765. 5 J. Puiggali, S. Munoz-Guerra and B. Lotz, Polymer, 1987, 28, 209. 6 Y. S. Wu,Ph.D. Thesis, University of Maryland, 1977. 7 K. E. Gonsalves, S. S. Stivala, L. Reich, S. H. Patel and D. H. Trivedi, Polym. Muter. Sci. Eng., 1990, 63, 962. 8 Standard Practice for Determining Resistance of Synthetic Poly- meric Materials to Fungi, G21-70, Annual Book of ASTM Standards, American Society of Testing and Materials, 1985. 9 S. H. Huang, in Comprehensive Polymer Science: Synthesis, Characterization, Reactions and Applications of Polymers, Perga-mon Press, Oxford, 1989, vol. 6, ch. 21, p. 597. 10 R. P. Wool and M. A. Cool, ASM Engineering Materials Hand- book, Vol. 2: Engineering Plastics, Plenum Press, New York, 1988, p. 783. Paper 1/00964H; Received 1st March, 1991

ISSN:0959-9428    

DOI:10.1039/JM9910100643

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

J. MATER. CHEM., 1991, 1(4), 649-653 The Bi,O,-Sm,O, System: Phase Diagram and Electrical Properties Pierre Conflant,* Claudine Follet-Houttemane and Michel Drache Laboratoire de Cristallochimie et Physicochimie du Solide, URA CNRS 0452,ENSCL et USTLFA, BP 108, 59652 Villeneuve d'Ascq Cedex, France The (Bi203), -x-(Sm203)x system (0Ixl0.65) was investigated by means of X-ray diffractometry, differential thermal analysis, impedance spectroscopy and e.m.f measurements. The corresponding equi Iibrium solid-phase part of the diagram is proposed. The most striking feature is the existence of a very large domain of a 6 Bi20, f.c.c.-type solid solution. By air-quenching from this domain two different phases were obtained: for 0.035~10.07, the room-temperature quenched phase exhibits tetragonal symmetry, changing to cubic for greater values of x.Both these phases are metastable and decompose into equilibrium phases on annealing. This transformation is always accompanied by a significant decrease in the electrical conductivity. Keywords: Bismuth-samarium-oxygen system; Oxide conductor; Solid electrolyte The high-temperature 6 modification of Bi203, which is stable between 730 "C and its melting point, 825 "C, is a well known and very attractive oxide ion conductor.' This behaviour has been related to the existence of a highly disordered oxygen- deficient fluorite-type structure.2 Partial substitution of Bi3 + by a large variety of cations allows the preservation of the 6 structure down to room temperature. This can be achieved, in particular, using trivalent cations such as Y3+ or the lanthanide~.~-~Among them, the solid solution obtained with Sm203 was shown to have the largest domain of quench- ability: 0.10 IX ~0.40,~assuming an import- ant stability extent.By annealing quenched samples, Takah- ashi and co-workers came to the conclusion that the 6 structure was stable from room temperature up to 900 "C only for those compositions very close to Bi,.,Sm,.,03 (x= 0.40). These conclusions are contrary to those previously reported by Levin and Roth,, in which the domain of stability of the 6 phase was shown to be limited to temperatures higher than 640 "C and x values lower than ca. 0.12. Moreover, in a general study of the stability of the &substituted phases, Watanabeg came to the conclusion that the 6 structure can never be stabilized at room temperature and that all the reported phases are, in fact, metastable.Taking into account these contradictory results, we have undertaken a systematic study of the bismuth-rich part (0Ix 50.65) of the Bi203-Sm203 system. The electrical properties of the various phases were carefully examined in relation to the oxygen deficient character of the structures. Experimental Powder samples were synthesised by solid-state reaction between Bi203 and Sm203 oxides, using pure commercial grade (>99Oh) raw materials. Bi203 was first heated at 600 "C in order to eliminate any traces of carbonate and moisture. Whatever the source of the product the so-called Sm203 'commercial' oxide was shown, by X-ray analysis, to be mainly Sm(OH),; it was therefore heated at 700 "C for a few hours in order to decompose to Sm203.The calcined oxides were then thoroughly mixed and ground in an agate mortar in the required proportions. Alumina or gold crucibles were used as containers for the synthesis. The thermal treatment consisted of a heating cycle around 800 "C for 2 or 3 days. This process was interrupted several times for intermediate grinding. The samples were then annealed for 24 h at different temperatures, in the range 600-1200 "C, and finally air-quenched. The products were analysed by means of X-ray diffraction using a Guinier-De Wolff camera. Investigation of the thermal behaviour was carried out using high-temperature X-ray diffractometry (HTXRD) with a Guinier-Lenne camera (heat- ing rate 20 " h-') and differential thermal analysis (DTA) (heating rate 300 O h-').Accurate cell-parameter variation was measured using KCl (Guinier-De Wolff camera) or gold (Guinier-Lenne camera) as the internal standard. Investigations of the electrical properties were carried out on sintered materials. Cylindrical pellets (diameter 5 mm; thickness ca. 3 mm) were obtained using a conventional press and used for conductivity studies. For oxygen transport measurements, the powdered electrolyte was first isostatically pressed into a cylindrical mould of 20 mm diameter and then machined in order to get 10 mm thick pellets.In both cases, samples were sintered for 15 h at ca. 25 "C below their melting temperature. After air-quenching the degree of compaction was in all cases >80%. For conductivity studies, two samples of each composition were used. The presence of the pure 6 phase was monitored by X-ray examination of the first sample. Gold electrodes were vacuum-deposited on both flat surfaces of the second pellet. Conductivity measurements were carried out by impedance spectroscopy, in the range 1-lo6 Hz using a Solartron 1170 frequency-response analyser. Each set of values was recorded at a given temperature after a 1 h stabilisation time. Oxygen transport measurements were performed by meas- uring the e.m.f of an oxygen concentration cell in which air and pure oxygen were used, respectively, as anode and cathode gases.Results and Discussion Equilibrium Phase Diagram Description Part of the Bi2 -2xSm2x03 solid phase diagram 0 5x I0.65 has been deduced from examination of room-temperature X-ray diffraction patterns and HTXRD and DTA of quenched samples (Fig. 1). It exhibits four solid-solution domains. a Solid Solution A monoclinic a-Bi203-type solid solution is observed in a narrow range of composition below 730 "C; it exhibits its maximum solubility (0I x I 0.02) near 690 "C. p2 Solid Solution The p2 solid solution is stable in a large domain of compo- sition from room temperature (0.17 5x s0.32)up to ca. 800 "C 650 A 1200 --I I I I I I / / / 0 oio 0.40 0.60 X Fig.1 (Bi203), -x-(Sm203)x solid phase diagram (05x 10.65) (0.1851~10.31). It is of the same form as the rhombohedra1 solid solution already reported in several Bi203-M0 or Bi203-M203 binary systems.8-" Fig. 2 represents the vari- ation of the lattice parameters for samples quenched from 780 "Cin the range 0.18 Ix I0.35. It exhibits a linear variation of a and c (hexagonal notation) inside the limits 0.20 Ix I0.3 1, which represents the range of stability of fi2 at this tempera- 0.396 tI 2.75 2.73 1Lt 1%I 0.10 0.20 0.30 0.40 X Fig. 2 x dependence of the lattice parameters of b2 solid solution J. MATER. CHEM., 1991, VOL. 1 ture. Above and below these limits, X-ray investigations confirm the presence of biphasic samples.Table 1 lists calcu- lated and observed diffraction-angle values and relative inten- sities of reflections for the composition Bil .4Sm0.603 (x = 0.30). fil Solid Solution At temperatures between 760 and 810 "C, the fi2 phase transforms into the fil phase. This transformation, evidenced by HTXRD, is accompanied by a sudden variation of a and c parameters. Fig. 3 presents the temperature dependence of a and c for a sample having the composition Bi,.4Smo.,03. In the Bi,03-alkaline-earth this transition has been attributed to an order-disorder structure rearrangement. It is associated with an endothermic effect. This transition Table 1 X-ray powder diffraction data for Bi1.,Sm,.,O3 h k I 20 obs. 20 calc. A (20) I/lmax(Yo) 0 0 6 19.4700 19.4761 -0.006 1 9 1 0 1 26.0650 26.0924 -0.0274 28 0 1 2 26.6900 26.7068 -0.0168 41 1 0 4 29.0400 29.0443 -0.0043 100 0 0 9 29.3700 29.3950 -0.0250 61 0 1 5 30.6800 30.6901 -0.0101 53 1 0 7 34.6900 34.7352 -0.0452 10 0 1 8 37.0400 37.0650 -0.0250 19 0 1 11 44.9700 44.9798 -0.0098 9 1 1 0 45.6600 45.6510 0.0090 47 1 1 3 46.7200 46.7862 -0.0662 4 1 1 6 50.I800 50.0732 0.1068 9 1 0 13 50.8800 50.8714 0.0086 35 0 1 14 53.9600 53.9742 -0.0142 5 0 2 4 55.0500 55.0213 0.0287 28 1 1 9 55.2300 55.2300 0.0000 57 0 2 7 58.6400 58.6 127 0.0273 7 0 0 18 60.9700 60.9857 -0.0157 6 0 2 13 70.7300 70.7257 0.0043 11 2 1 4 74.2100 74.1989 0.01 11 16 1 1 18 79.3800 79.3940 -0.0140 9 L=0.154056 nm (Cu-Ka,)-refined parameters: a =0.3971 3(8) nm c = 2.7324(7) nm.,A' A2.78 I / A' E A 1 2.76 .-2.74 .. 0.41 i 0.39 1 I I*--4 250 500 750 106 T/"C Fig. 3 Temperature dependence of Bil.,Sm,.6 0, cell parameters J. MATER. CHEM., 1991, VOL. 1 was shown to be highly reversible, and quenching of samples from the PI domain always leads to X-ray patterns and unit- cell parameters identical to those obtained when quenching from p2. 6 Solid Solution This solid solution is directly related to the Bi203 cubic modification. Its domain of existence is strongly dependent on temperature. It varies from 01~10.08 at 730 "C to 0.301~~0.55at 1200 "C.Two eutectoids are observed at 690 "C for x=0.07 and at 770 "C for x= 0.43, respectively. For x values greater than the upper limits of stability of the cubic phase, a rhombohedra1 LaOF-type phase emerges in the X-ray patterns. Air quenching of the cubic 6 phase leads to different results depending upon the x value. For ~10.01, the quenched samples appear monoclinic. For 0.03IxI0.07, a monophasic domain is obtained which exhibits tetragonal symmetry (T). Such a structural type has already been reported in bismuth- based system^^*'^ and is related to the structure of P-Bi203. theFor ~~0.08 quenched materials exhibit the 6 cubic structure 6,. Fig. 4 represents the variation of the tetragonal and cubic unit-cell parameters as a function of x for samples quenched from 780 "C.The tetragonal domain is characterized by an increase of the unit-cell volume with increasing x, OS7 t 0.56 E 6 6i-1 Y 0.55 0.54 1 ~~ 0 0.05 0.1 0.15 X Fig.4 Dependence of the lattice parameters of T and 6, quenched phases 0 I -.-.250 < 500I-750 651 whereas a slight decrease is observed through the cubic domain. Thermal stability of the quenched T and 6 phases was examined by HTXRD. For x<0.4, irrespective of the struc- tural type of the starting phase, a decomposition occurs (Fig. 5 and 6). For larger values of x, no decomposition can be detected by this technique (Fig. 7). However, room-tempera- ture X-ray analysis of samples annealed for 24 h at a tempera- ture of ca. 650 "C clearly shows that decomposition also affects (at least partly) samples with x20.4.Conductivity Measurements Previous electrical investigations performed by Iwahara et aL5 were carried out without any precise prior knowledge of the equilibrium phase diagram and without evaluation of the thermal stability of 'stabilised' fluorite samples. According to our HTXRD investigations, it appears that a decomposition occurs when materials with the fluorite structure are heated; this decomposition is complete for samples with x=0.20 and 0.30 and is at least partial for x =0.40. As a consequence, we performed a new set of electrical measurements and the corresponding results were analysed in connection with the DTA and HTXRD experiments.This led to a precise correlation between the electrical behaviour and the nature of the phase(s) present in the samples because of their thermal history. Investigations were performed on pellets quenched from the high-temperature 6 domain, i.e. on the tetragonal phase for 0.03Ix10.07 and on the cubic phase for 0.081~10.45. Conductivity measurements were made in the temperature range 300-760 "C for x10.1 and 300-850 "C for x>O.l, during two heating-cooling cycles. Fig. 8 represents the Arrhenius plot of two representative samples [x =0.05 (a) and x=0.30 (b)] in each domain. For x=O.O5 during the first heating run, the initial linear part is characteristic of the tetragonal phase.The decrease of conductivity, which occurs at 540 "C, corresponds to its decomposition into a mixture of a and P2.The jump at 750 "C accompanies the move to the 6 domain. On cooling, only one transition is present and corresponds to the 6,+a +p2 transformation, which occurs with significant hysteresis of ca. 100". The subsequent heating or cooling cycles reproduce this behaviour. Fig. 8(a) and (b)shows significant differences between the 40 100 . Fig. 5 HTXRD pattern for quenched Bi,,,Sm,,,O, J. MATER. CHEM., 1991, VOL. 1 Fig. 6 HTXRD pattern for quenched Bi,,,SmO,,O3 Fig. 7 HTXRD pattern for quenched Bi,,,Sm,,,O, two domains. In the first, the conductivity of the quenched phase (T) is lower than that of the equilibrium phases (a +p2).This quite surprising effect can be explained by the presence of the a phase which is known to be an electronic conductor. The situation is reversed in the second domain (x=0.30) and the conductivity of 6, is about one order of magnitude larger than the conductivity of the equilibrium /j2 phase. Fig. 9 shows the variation of log 0 for different x values. The substitution of Bi3+ by Sm3+ is accompanied by a linear decrease of log 0, for example from -1.5 to ca. -3 when x increases from 0.10 to 0.40 at 440 "C [Fig. 9(b)].It must be remembered that the stability of the quenched phase increases with the Sm3+ content. E.m.f. measurements performed on the S,, and p2domains always lead to to2-ratios of the measured e.m.f.to the theoretical value ranging from 0.8-1. However, a signifi- cant decrease in the time necessary to reach the equilibrium potential in the S, domain was observed. Additional investi- J. MATER. CHEM., 1991, VOL. 1 t I 1t- I 1 I +-I1.11.1 1.31.3 1:51:5 1:1 1.3 1.5 lo3 KIT 103 KIT Fig. 8 Arrhenius plots of conductivity for the first heating and cooling cycles for quenched Bi,,,Smo.,03 (a) and Bi,,,Sm0~,O3 (b)samples 7°C 441 393 352 -1 0.10 0 0.20 0 0 0 0 0 0 0 0.30 f 0. A+ 0..0.35 00.40 oA+ OA:+oo0 0. + -4 .-oA + Or 0 I II L 1t-: -L-: . > 1.4 1.5 1.6 1.7 0.10 0.20 0.30 0.40 103 KIT X Fig. 9 (a) Arrhenius plots of conduction for 6, solid solutions at various x; (b) variation of isothermal 0 versus x (T=440 "C) gations are currently in progress to confirm these observations and to explain the origin of this phenomenon.Conclusion This study has enabled the (Bi203)1-x-(Sm,03), (0 5x 50.65) equilibrium solid-phase diagram to be established. It has been shown unambiguously that the f.c.c. phase is only stable at high temperatures. However, air quenching from this domain leads to two metastable phases: a tetragonal symmetry phase for 0.031~10.07 and an f.c.c. phase for 0.08Ix I0.45. Metastable and equilibrium phases both exhi-bit oxygen-ion conduction. By annealing, the quenched phases go back to equilibrium phases whatever the Sm203concen-tration. This transition is always accompanied by a significant decrease of electrical conductivity.The quenched cubic phase exhibits, at moderate temperature, the most attractive ionic conductivity values and investigations are in progress in order to check its stability under long-term operating conductions. The authors would like to thank G. Hanekamp for useful collaboration on this work. References 1 T. Takahashi, H. Iwahara and Y.Nagai, J. Appl. Electrochem, 1972, 2, 97. 2 H. A. Harwig, Z. Anorg. Allg. Chem., 1978,444, 151. 3 T. Takahashi, H. Iwahara and T. Arao, J. Appl. Electrochem., 1975, 5, 187. 4 T. Takahashi, T. Esaka and H. Iwahara, J. Appl. Electrochem., 1975, 5, 197. 5 H. Iwahara, T. Esaka, T. Sato and T. Takahashi, J. Solid State Chem., 1981,39, 173. 6 M. J. Verkerk and A. J. Burggraaf, Solid State Ionics, 1981, 314, 463. 7 A. Watanabe and T. Kikuchi, Solid State Ionics, 1986, 21, 287. 8 E. M. Levin and R. S. Roth, J. Res. Natl. Bur. Stand., 1964, @A, 2,200. 9 A. Watanabe, Solid State Ionics, 1990, 4/41, 889. 10 P. Conflant, J. C. Boivin and D. Thomas, J. Solid State Chem., 1976, 18, 133. 11 R. Guillermo, P. Conflant, J. C. Boivin and D. Thomas, Rev. Chim. Mineral., 1978, 15, 153. 12 P. Conflant, J. C. Boivin, G. Nowogrocki and D. Thomas, J. Solid State Ionics, 1983,9-10, 925. 13 P. Conflant, J. C. Boivin and D. Thomas, J. Solid State Chem., 1980, 35, 192. Paper 1/01005K; Received 4th March, 1991

ISSN:0959-9428    

DOI:10.1039/JM9910100649

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

J. MATER. CHEM., 1991, 1(4), 655-661 Aromatic Ether-Ketone-'X' Polymers Part 2.-EK-lmide Copolymers Christopher J. Borrill and Richard H. Whiteley* Raychem Ltd., R & Q Faraday Rd., Dorcan, Swindon, Wiltshire SN3 5HH, UK A modified Friedel-Crafts synthesis has been used to make two series of aromatic ether-ketone-imide copolymers. All the copolymers were semicrystalline and, in some cases, useful combinations of glass-transition temperature and melt temperatures were found. Some properties of the copolymers are reported and discussed. Keywords: Poly(ether-ketone); Friedel-Crafts; Aromatic polymer; Polyimide In the first paper of this series the synthesis and properties of a variety of high-molecular-weight aromatic ether-ketone-')(' (EKX) polymers were reported.' The polymers were made using a modified Friedel-Crafts reaction.2 One of the X groups that could be incorporated was the imide group, and because of the potential utility of EK-imide polymers as high Tg thermoplastic materials with very good thermo- oxidative stability, we have carried out further studies of this class of materials.Experimenta1 Two series of copolymers were made. One, series A, was made by polymerizing the imide monomer EIKIE (I) with mixtures of isophthaloyl chloride (IPC) and terephthaloyl chloride (TPC) to give polymers with the general structure 11. -1" The other series of copolymers, series B, was made by poly- merizing terephthaloyl chloride with mixtures of the ether- ketone monomer EKE I11 and the imide monomer I to give copolymers which contained the structural units IV and V.t IV J The ether-ketone polymer, PEKEKK, obtained by poly- merizing 111 with TPC is now being made commercially, and is marketed by BASF under the registered trademark 'ULTRAPEK'.2-4 Reagents and Solvents Terephthaloyl chloride (TPC), isophthaloyl chloride (IPC), aluminium chloride, 4,4'-diphenoxybenzophenone, and 4-phenoxybenzophenone were all supplied from the research laboratories of Raychem Corporation, Menlo Park, Cali- fornia. The imide monomer I was made as described pre- viously.' N,N-dimethylfomamide (DMF) and benzoyl chloride were obtained from Aldrich. Dichloromethane (DCM), 1,2-dichloroethane (DCE), sulphuric acid and meth- anol were obtained from Fisons.Synthesis and Characterization of Polymers The polymers were all made using the general procedure as described previously. Series A copolymers were made using a stoichiometric excess of diacid chloride(s) with a correspond- ing amount (i.e. twice the molar excess) of 4-phenoxybenzo- phenone as an end-capper, and with DCE as the solvent. With the exception of polymer 1, series B copolymers were made using a stoichiometric excess of dinucleophilic mono- mer(s) with a corresponding amount of benzoyl chloride as an end-capper7 and with dichloromethane as the solvent. DMF was used as the Lewis base in all the polymerizations except in the case of polymer 8 when dimethyl sulphone was used. The amounts of reagents used are listed in Table 1.Polymerization yields were between 93 and 100%. Polymer structures were confirmed using high-resolution NMR spec- troscopy, and detailed assignments for polymers 7 and 8 have been reported.' Because the general method of synthesis was similar for all the polymers, full details are given below for only two represen tat ive examples. Polymer 4 DCE (90cm3) was stirred and cooled under nitrogen at -30 "C in a 250 cm3 resin kettle. Aluminium chloride (24.6 g, 0.184 mol) was added, followed by DMF (4.71 cm3, 0.0612 mol), keeping the temperature below -10 "C. EIKIE (9.850 g, 0.01 50 mol) and 4-phenoxybenzophenone (0.165 g, 0.0006mol) were then added, followed by a mixture of IPC (1.553 g, 0.00765 mol) and TPC (1.553 g, 0.00765 mol), all added at a temperature of ca.-30 "C. 30 cm3 of DCE were used to ensure complete addition of the reagents. The mixture was stirred and allowed to warm to -5 "C over 2 h. A tough rubbery gel separated from a clear, dark brown-yellow liquid. The mixture was left to stand for 15 h at 8 "C,then the liquid was decanted (ca.80 cm3) and the gel was worked up in two portions, each with 400 cm3 of methanol in a Waring blender. The combined decomplexed polymer portions were then blended again with a further 400cm3 of methanol. The resulting lemon-yellow fibrous polymer was dried at 140 "C for 2 h under vacuum. The yield was 12.02 g (100%). J. MATER. CHEM., 1991, VOL.l Table 1 Amounts of reagents used in polymerisation reactions" polymer solven t/cm EKE/mmol EIKIE/mmol TPC/mmol IPC/mmol AlCl,/mmol ~~ 1 340 - 137.3 133.3 - 1627 2 110 - 30 26.52 4.68 375 3 120 - 15 10.71 4.59 184 4 120 - 15 7.65 7.65 184 5 120 - 15 4.59 10.71 184 6 800 - 100 15.30 86.70 1228 7 120 - 15 - 15.30 184 8 330 309.1 - 303.0 - 1658 9 375 18'4.0 20 200.0 - 1641 10 400 164.0 40 200.0 - 1729 11 475 120.0 50 166.6 - 1514 12 500 103.3 66.7 166.6 - 1588 13 510 95.0 75 166.6 - 1624 14 620 86.7 83.3 166.6 - 1661 15 520 56.9 100 153.8 - 1635 16 520 31.4 1 14.3 142.9 - 1612 a Dimethyl sulphone was used as the Lewis base for the preparation of polymer 8.The mole ratio of dimethyl sulphone to TPC was 1.5: 1.DMF was used for the preparation of all the other polymers with a mole ratio of DMF to diacid chloride(s) of 4: 1. The stoichiometric imbalance was ca. 2%, except for polymers 1 and 2 when it was 3 and 4%, respectively. Polymer 11 DCM (300 cm3) was cooled under nitrogen to -20 "C in a 1 dm3, jacketed reaction vessel. The vessel was cooled using circulating ethylene glycol-water from a thermostatically con- trolled reservoir. Aluminium chloride (202 g, 1.5 14 mol) was added to the stirred solvent, followed by DMF (51.3 cm3, 0.666 mol) at such a rate that the reaction mixture was maintained at a temperature below -10 "C. TPC (33.824 g, 0.1666 mol) and benzoyl chloride (0.956 g, 0.0068 mol) was then added using 80 cm3 of DCM to ensure complete addition.The reaction mixture was cooled to below -15 "C and a mixture of 4,4'-diphenoxybenzophenone (43.97 g, 0.120 mol) and EIKIE (32.833 g, 0.050 mol) was added, again using DCM (95 cm3) to ensure complete addition. The nitrogen gas inlet was replaced with an outlet to a bubbler to monitor the evolution of hydrogen chloride, and the reaction mixture was warmed to room temperature over a period of ca. 1 h. Stirring was stopped when the solution became viscous, and the mixture was then left overnight at room temperature. The polymer gel was removed from the vessel as a 'lollipop' on the stirrer rod [the internal diameter of the vessel (100 mm) was the same as the flange opening], and was then cut up into four pieces.These were each worked up separately in 1 dm3 of cold (-18 "C) methanol in a 4 dm3 Waring blender. The batches of decomplexed polymer were then filtered and blended again, in one batch, with the minimum amount of cold methanol required to cover the polymer in the blender. The polymer was again filtered, then left to soak in methanol overnight before being boiled under reflux for 3 h. Finally, the polymer was filtered, washed with methanol and dried under vacuum at 200 "C overnight. The yield of fibrous yellow polymer was 97 g (98%). Equipment and Procedures Inherent viscosities were measured at 25 "C on 1.0 g dm-3 solutions of polymer fluff in 98% analytical-grade sulphuric acid. All other measurements were made on samples cut from compression-moulded plaques.DSC measurements were made on a DuPont 1090 instru- ment using a 910 cell. Indium and zinc were used for tempera- ture calibration. Scans were made in nitrogen at 10 "C min-'. 'H and 13C NMR spectra were recorded using a Bruker AM300 Fourier transform instrument. Solutions were obtained by soaking ca. 100 mg of polymer in 2 cm3 of CDC13 for 30 min, followed by the addition of trifluoroacetic acid dropwise, with stirring, until the polymer dissolved. TG measurements were made using a Perkin-Elmer TGS- 2 instrument with a System 4 controller. A scan rate of 10 "C min-' was used. Dynamic mechanical thermal analysis (DMTA) measure- ments were made using a Polymer Laboratories Mark 1 machine in single cantilever and auto-strain modes.A scan rate of 4 "C min-' was used, at an oscillation frequency of 1 Hz. Unless stated otherwise, the glass-transition temperature of a given polymer corresponds to the maximum in the DMTA plot of loss modulus uersus temperature. Wide-angle X-ray diffraction spectra were recorded using a Philips PW1050 goniometer using Cu-Ka radiation with a nickel filter. Crystallinity was calculated using the formula where A, is the area under the I uersus 0 diffraction scan due to amorphous scattering, and A, is the area due to crystalline scattering, i.e. the total area minus A,. [N.B. This method yields an approximate value. For more accurate values s21 uersus s scans should be used, where s =2 sin0/A.'] Tensile testing was done according to BS2728 at a strain rate of 1 mm min-'.BS type 2 dumbell test samples were cut from 2mm thick plaques using a router. Values quoted are the averages of four tests. Volume resistivities were measured according to BS2782 method 202B. Permittivities and loss factors were evaluated according to BS2782 method 206B at 15.92 Hz. Smoke measurements were made using a small-scale dynamic smoke-measuring instrument developed at Raychem, Swindon.6 Samples of 0.2 g were used. Smoke values were evaluated by integration of absorbance uersus time plots, and values quoted are the averages of three measurements. Average total smoke values (s)were obtained from integration of smoke uersus temperature plots over the temperature range 473-1 173 K, using the formula (1173 K) (473 K) Limiting oxygen index measurements were made according to ASTM D2863-74.The solutions used for testing hydrolytic stability were 0.01 mol dm- (as.) HC1 (pH 2), 0.025 mol dm-3 (aq.) Na2HP04-0.025 mol dm-3 (aq.) KH2P04 (pH 7) and J. MATER. CHEM., 1991, VOL.l 0.01 mol dm-3 (aq.) NaOH (pH 12). The pH of each of the solutions was checked at regular intervals during the testing and adjusted when necessary. Results and Discussion Series A Copolymers Table 2 lists the copolymers that were made together with their solution inherent viscosities (IV), and some differential scanning calorimeter (DSC), dynamic mechanical thermal analysis (DMTA), and thermogravimetric analysis (TG) data. Polymer 1, made from EIKIE and TPC, was semicrystalline with a Tgof 247 "C and a T, of 435 "C.Polymer 7, made from EIKIE and IPC, was also semicrystalline with a Tgof 218 "C and a T, of 358 "C. These polymers were described in an earlier paper,' and it was partly because of their interesting properties that we extended our studies to the copolymers discussed in this paper. Polymers with nominally the same repeat unit as 1 and 7 have also been made by the amic acid route.7 We expected that the incorporation of mixtures of isoph- thaloyl and terephthaloyl units would disrupt the structure of the resulting polymers to such an extent that crystallization would be suppressed. However, to our surprise all the copoly- mers were semicrystalline.Fig. 1 shows the effect of the changing ratio of isophthaloyl to terephthaloyl units on both Tgand T,. With the exception of polymer 2 the Tgvalues increased with increasing terephthaloyl content. However, the changes in T, were more complex. The general trend was an increase in T, with increasing terephthaloyl content, but replacement of 15% of isophthaloyl units with terephthaloyl units (1 5 :85, TPC :IPC) caused a slight reduction of T, from 358 to 354 "C which was contrary to the trend. Also at 85: 15 TPC :IPC (polymer 2) two melting points were observed. Fig. 2 shows the DSC scan of a quenched (amorphous) sample of the polymer. In addition to the Tgtransition at 231 "C and a crystallization exotherm at 288 "C, there is a major melting endotherm at 395 "C and a minor melting endotherm at 432°C.The higher melting point corresponds with that of the 1OO:O TPC:IPC polymer and is presumably associated with blocks of that structure. The anomalous Tg of this polymer is probably related to its unusual morphology. Fig. 3 shows a wide-angle X-ray diffraction scan of an annealed sample of polymer 4 which contained a 50 :50 ratio of isophthaloyl and terephthaloyl units. There are two strong diffraction peaks corresponding to d spacings of 3.96 and 4.75 A. Integration of the areas under the crystalline peaks and under the amorphous halo indicated an approximate degree of crystallinity of 34%. All the copolymers showed good thermo-oxidative stability as measured by TG.A 1% mass loss occurred at an average temperature of 478 "C in air, and at an average temperature of 505 "C in nitrogen. However, the melt stability of the copolymers was not good. It was possible to make com-I I I 200 300 400 T/ "C Fig. 2 DSC scan of polymer 2 10 20 30 40 201" Fig. 3 Wide-angle X-ray diffraction scan of polymer 4 440 r p320 2801I 200 1OO:O Fig. 1 polymer I I I I I 7-I 80:20 60:40 40:60 20:80 0:lOO TPC:IPC ratio Series A copolymers; T, and T,,,data Table 2 Series A copolymers TPC:IPC ratio IV/ cm3 g- ' TI%(air)/ "C 100 :0 121 247 435 480 472 85: 15 132 227 395,432 487 543 70 :30 71 233 381 465 502 50 :50 109 229 382 469 507 30 :70 168 227 37 1 490 520 15:85 100 220 354 494 508 0: 100 112 218 358 459 486 J.MATER. CHEM., 1991, VOL.1 Table 3 Series B copolymers polymer EKE :EIKIE ratio IV/cm3 g-' Tg/"C TJC AHJJ g-I Tl%(air)/ "C Tl%(N2)/"C 8 1oo:o 115 170 38 1 75 495 520 9 90.2:9.8 108 180 369 65 470 505 10 80.4:19.6 134" 186 357 35 405 470 11 70.6:29.4 127 186 347,372 34 445 495 12 60.8 :39.2 141 196 33 5,3 72 32 475 495 13 55.9:44.1 173 204 33 1,374,406 29 450 515 14 5 1 .O:49.0 163 208 377,409 27 455 480 15 36.3 :63.7 152 220 382,420 17 465 500 16 21.6: 78.4 139 237 37 5,427 52 465 520 1 0: 100 121 247 435 57 480 472 " Some undissolved gel was filtered from the solution.pression-moulded plaques of fair quality by pressing at a temperature slightly above the T,, but attempts to measure the melt viscosity of the polymers by capillary rheometry showed that rapid and large increases in viscosity occurred when the polymers were held for short times above their melting points. We intend to report further work concerning melt stability in due course. Series B Table 3 lists the copolymers that were made together with IV, DSC, DMTA and TG data. Polymer 8 is the aromatic ether- ketone polymer PEKEKK which has a T, of 170 "C and a T, of 381 "C.* We have studied this series of copolymers in greater depth than the series A copolymers and, in addition to examining their thermal properties, we have investigated some mechan- ical, electrical, flammability, and hydrolytic properties.Thermal Properties As with series A we expected that the incorporation of monomer I into the backbone of the polymer would disrupt the structure and produce amorphous polymers. However, again we were surprised to find that all the copolymers were semicrystalline. Furthermore, although incorporation of I caused the T, of the polymers to rise steadily, from 170 to 247 "C, the effect at low levels of imide incorporation was to depress the melting point rather than to increase it as we had expected. We thus had the attractive opportunity of being able to increase the T, of the basic PEKEKK polymer whilst simultaneously lowering the T, and so potentially increasing the facility with which the polymer could be melt processed.Fig.4 shows the effect of increasing imide content on T, and T,. It can be seen that between ca. 30 and 80mol% imide? more than one melting point was observed, indicating the presence of different crystalline morphologies. At 40 mol% imide the T, has increased to 196°C whilst the higher T, was still slightly lower than for PEKEKK at 372 "C. The melting behaviour of the copolymer series was most unusual and suggested that three different morphologies could occur, represented by lines 1, 2 and 3 in Fig. 4. The line 1 melting points are presumably associated with a disrupted PEKEKK structure. The line 2 melting points are all similar and average 376 "C. This suggests that they are associated with a basically undisrupted PEKEKK structure. The line 3 melting points are all above 400 "C and are presumably associated with a disrupted PEIKIEKK structure.Fig. 5 shows the changing nature of the DSC endotherms with increasing imide content. The intensity of the melting endotherms decreased as the imide content increased, falling 7 By 30 mol% imide we mean a 70:30 mol ratio of monomer 111 (EKE): monomer I (EIKIE). 450 II I I I 400 350 9,300I--250 200 1OO:O 80:20 60:40 40:60 20:80 0:lOO EKE:EIKIE ratio Fig. 4 Series B copolymers; Tgand T,,,data from 75 J g-' for polymer 8, PEKEKK, to a minimum of 17 J g-' with polymer 15 which contained 63.7 mol% imide. As the imide content increased further the endotherms increased in intensity, reaching 57 J g-' with polymer 1, PEIKIEKK.Thus, although the incorporation of the imide unit into the PEKEKK structure did not completely destroy the crystallinity of the polymer, it did reduce it considerably. All the DSC scans showed a small endotherm at ca. 280 "C. We have found that all annealed samples show this minor melting endotherm, typically between 10 and 20 "C above the annealing temperature. Similar behaviour has been noted in PEEK.' It should be appreciated that our analysis of the melting behaviour is mainly conjectural at this stage and is no doubt an oversimplification. The relative intensities of the melting endotherms observed in DSC scans are sensitive to the thermal history of the polymers, and the morphology of the copolymers will be affected by how random or blocky the polymers are, and this in turn may be affected by the polymerization conditions.Work in this area is continuing. The copolymers showed good thermo-oxidative stability as measured by TG. 1% mass loss in air was at 461 "C, and in nitrogen was at 497 "C.However, as with the series A copoly-mers, melt stabilities were not particularly good, and they became worse with increasing imide content. J. MATER. CHEM., 1991, VOL.1 Table 4 Mechanical properties of series B copolymers ~ ~ ~~~~ polymer EKE :EIKIE ratio modulus/MPa tensile strength/MPa" EtJ (%) 8 1oo:o 3730 103.4 3.5 9 90.2 :9.8 3720 106.4 4.5 10 80.4:19.6 3760 96.1 3.3 11 70.6:29.4 3770 108.7 6.6 12 60.8:39.2 3600 103.4 6.5 13 55.9:44.1 3650 102.9 4.9 14 51.O :49.0 3655 64.4 2.0 15 36.3 :63.7 3575 109.1 8.3 16 21.6 :78.4 3700 63.0 1.8 1 0: 100 4205 63.2 1.8 The data given are ultimate tensile strengths except for polymers 11, 12 and 15 when they are yield stresses.in Table 5. The relative permittivity was unaffected by imide content and was 3.08 k0.09. The loss factor (tan 6) decreased with increasing imide content from 0.0024 at Omol% imide to 0.0018 at 100mol% imide. This was contrary to our expectations as the EIKIE unit contains more polar groups than the EKE unit and therefore we expected power loss to increase. Flammability and Smoke Production Fig. 6 shows the limiting oxygen index (LOI) of the copolymers as a function of imide content.All the copolymers were inherently fire resistant with LO1 values between 35.5 and 44.5. The slight fall in LO1 as the imide content increased to ca. 30 mol% may be related to the decrease in the melting temperature of the copolymers. Fig. 7 shows the plots of smoke production as a function of temperature for five of the copolymers. No smoke was detectable below 525 "C,but above this temperature smoke production rose rapidly until the self-ignition temperature (Tip)was reached at ca. 630 "C.Above Ti, smoke production was lower and approximately constant for a given polymer. Increasing imide content has a pronounced smoke suppressing effect, and Fig. 8 clearly illustrates this showing how the average total smoke production, S, fell from 549 cm2 g-' at 0 mol% imide to 175 cm2 g-' at 100 mol% imide.Hydrolytic Stability In our previous paper we suggested that the method of synthesis of these EK-imide polymers should result in good hydrolytic stabilities.' To test this hypothesis we boiled poly- mer fluff samples in distilled water and in pH 2, 7, and 12 aqueous solutions, and we monitored the changes in inherent viscosity of the polymer with time. Table 6 shows the results obtained for polymer 8, PEKEKK, which contained no imide groups, and Table 7 shows the data for polymer 11 which contained 29.4 mol% of EIKIE. In the case of PEKEKK all the measured inherent viscosities were in the range 108-121 cm3 g-' and they showed no significant Table 5 Electrical properties of series B copolymers 0% imide 200 300 /I560 9.8% 19.6% 29.4% 39.2% 44.I Yo 49.0% 63.7% 78.4% 100% 2z 300 400 500 Fig. 5 Series B copolymers; DSC scans Mechanical Properties Initial modulus, tensile strength, and elongation-to-break (Eb) data are listed in Table4. Because of the relatively poor quality of the compression-moulded test specimens, many of the samples broke before they yielded. However, data from the better quality specimens suggested that both modulus and tensile strength were unaffected by imide content. Tensile strengths were ca. 105 MPa and moduli were ca. 3.7 GPa; similar to the values quoted for both PEEK and PEKEKK.*,'' The elongation to yield of the better quality polymer EKE :EIKIE ratio 8 1oo:o 9 90.2 :9.8 10 80.4:19.6 11 70.6:29.4 relative permittivity loss factor x I04 3.08 24 3.14 20 3.19 22 2.97 20 3.18 20 3.13 21 3.03 19 3.10 17 3.04 18 2.89 18 specimens was ca.6.5% 12 60.8:39.2 13 55.9:44.1 Electrical Properties 14 51.O :49.0 All the copolymers were good insulators with resistivities in 15 36.3 63.7 excess of 1.4 x 1OI6 0 cm (the maximum measurable on our 16 21.6 :78.4 1 0: 100equipment). Relative permittivities and loss factors are listed 660 48 I I I I 44 2 40 36 0 32 I I I I 0 20 40 60 80 100 imide (mol%) Fig. 6 Series B copolymers; LO1 us. imide content +-19.6% imide +-44.1% imide 1 +--78.4% imide +-100%imide 1 0' II I I 500 600 700 800 900 T/ "C Fig.7 Series B copolymers; smoke us. temperature. Imide content (Yo):0,0; 0, 19.6; A, 44.1; 0,78.4; 0, 100 trends with time. If one assumes therefore that no degradation took place, a mean IV of 114.5 cm3 g-' and a standard deviation of 3.4 cm3 g- ' can be calculated. This gives a useful measure of the precision of the IV determination. In the case of polymer 11 there were no significant changes in IV in distilled water, pH 7, and pH 2 solutions up to 167 h. After 982 h in distilled water and in the pH 2 solution the IV values had fallen from 113 cm3 g-' to 106 cm3 g-' and 105 cm3 g- ' rzspectively. These are small but possibly signifi- J. MATER. CHEM., 1991, VOL.l 100 0 20 40 60 80 100 imide (mot%) Fig.8 Series B copolymers; s us. imide content Table 6 Inherent viscosities of polymer 8 (PEKEKK) after boiling in water time/h distilled water PH 2 pH7 pH 12 0 115 115 115 115 1 116 118 115 116 7 108 117 109 113 31 116 119 116 121 167 114 116 114 115 479 112 114 108 113 982 110 116 109 115 Table 7 Inherent viscosities of polymer 11 after boiling in water IV/cm3 g- time/h distilled water PH 2 pH7 pH 12 0 113 113 113 113 1 I18 115 112 114 7 115 113 113 112 31 115 113 114 109 167 115 116 112 107 479 109 110 105 101 982 106 105 99 92 cant changes. After the same time in the pH 7 phosphate buffer solution the IV had fallen by slightly more to 99 cm3 g-'.Degradation was more evident in the pH 12 solution where there was a steady fall in IV with time to 92 cm3 g-' after 982 h. In general, as we had predicted, the hydrolytic stability of the EK-imide polymer was good. We believe that these results support the case that the reported hydrolytic instability of some polyimides made by the amic acid route is because of defects in the polymer chain rather than because of any inherent instability of the imide bond.' '-14 J. MATER. CHEM., 1991, VOL.l 661 We thank Bruce Fox for X-ray diffraction measurements, Tim Harding for the testing of tensile properties, John Harris for DSC, DMTA and TG measurements, Pat Horner and Ian Towle for their helpful discussions, and Raychem Ltd. for permission to publish this paper.5 6 7 8 J. F. Rabek, Experimental Methods in Polymer Chemistry, Wiley-Interscience, Chichester, 1980, p. 507. R. H. Whiteley, Int. Wire Cable Symp. Proc., 1982, 427. P. M. Hergenrother, N. T. Wakelyn and S. J. Havens, J. Polym. Sci., Polym. Chem. Ed., 1987, 25, 1093. BASF product brochure B607e, ‘ULTRAPEK’, 1989. 9 S-S. Chang, Polym. Commun., 1988, 29, 138. References 10 T. W. Haas, in Handbook of Plastic Materials and Technology, ed. I. Rubin, Wiley, New York, 1990, p. 287. P. J. Horner and R. H. Whiteley, J. Muter. Chem., 1991, 1, 271. V. Janson and H. C. Gors (Raychem Corp.), PCT Znt. Appl., WO 84 3891, 1984. P. Becker, P. J. Horner, S. Moore, L. J. White, L. M. Edwards, B. Macknick and R. J. Mosso (Raychem Corp.), Eur. Pat. Appl., 11 12 13 14 C. E. Sroog, J. Polym. Sci., 1965, 3, 1373. D. R. Askins, Hydrolytic degradation of Kapton film, Air Force Wright Aeronautical Laboratories, AFWAL-TR-83-4125, 1983. J. H. Hodgkin, J. Appl. Polym. Sci., 1976, 20, 2339. D. Kumar, J. Polym. Sci., Polym. Chem. Ed., 1980, 18, 1375. EP 314384, 1989. J. Koch, W. Stegmaier and G. Heinz, BASF AG, DE 3829520 Al, 1990. Paper 1/01065D; Received 6th March, 1991

ISSN:0959-9428    

DOI:10.1039/JM9910100655

出版商:RSC    

年代:1991

数据来源: RSC