摘要:

Transition-metal stannides with high tin content: Os,Sn,,, RhSn,, RhSn, and IrSn, Arne Lang and Wolfgang Jeitschko Anorganisch-Chemisches Institut, Universitat Miinster, Wilhelm-Klemm-Str. 8, 0-48149 Miinster, Germany The title compounds have been prepared in well crystallized form by lengthy annealing of the elemental components with the atomic ratios varying between 1:3 and 1:10. The tin-rich matrix was dissolved in dilute hydrochloric acid. The crystal structures of these stannides were determined from single-crystal X-ray data. Os,Sn,, crystallizes with a new structure type: Pbcrn, a=694.9( 1) pm, b= 1428.1(3)pm, c= 1921.1(4) pm, Z=4, R=0.023 for 1573structure factors and 102 variables. RhSn, is isotypic with P-CoSn,: 141/acd, a=632.4( 1)pm, c= 3412( 1) pm, Z= 16, R=0.016 for 231 F values and 21 variables.RhSn, has a IrGe,-type structure: P3,21, a= 677.4(2)pm, c= 861.4(2)pm, Z= 3, R=0.018 (1285 F values and 26 variables). IrSn, is isotypic with RhSn, and IrGe,: P3,21, a =679.1 (2) pm, c =857.5(3)pm, Z =3, R =0.018 (1165 F values and 26 variables). In all of these structures the transition-metal (T) atoms have well defined coordination polyhedra. The coordination polyhedron of the Rh atom in RhSn, is formed by one Rh and eight Sn atoms; in the other compounds (with higher tin content) the T atoms have (only) eight or nine Sn neighbours with T-Sn distances covering the relatively narrow range 267-303 pm. In contrast, the coordination polyhedra of the Sn atoms are difficult to define. In addition to the four (one Sn site in RhSn,) or two T neighbours (all others), each Sn atom has between ten and twelve Sn atoms in its coordination shell at distances almost continuously covering the range 286-408 pm.As a consequence the structures contain relatively large voids. Magnetic susceptibility measurements carried out with a SQUID magnetometer indicate Pauli paramagnetism for all four compounds, in agreement with the metallic conductivity found by four-probe electrical conductivity measurements for Os,Sn,, superconducting down to 2 K. In the course of our investigations of transition-metal stannides with high tin contents we have reported on MoSn,,' VSn,,, NbSn,2 and CrSn,., Recently we reinvestigated the tin-rich side of the cobalt-tin binary system, where we characterized two modifications of the new compound COS~,.~ The low- temperature a-phase was found to be isotypic with PdSn,;, for the high-temperature P-phase a new structure type was established.We have subsequently searched for isotypic stan- nides in related binary systems and found the compounds Os,Sn,, and RhSn,, which, to our knowledge, have not been reported previously. The other two stannides discussed in the present paper, RhSn, and IrSn,, have already been mentioned in the literature. A compound with the tentative composition RhSn, was observed to form a eutectic with tin, but the compound was not characterized f~rther.~ IrSn, has been investigated by "'Sn Mossbauer spectro~copy~.~ and appar- ently its trigonal crystal structure had been determined.' However, it seems that this structure has never been published.Another modification of IrSn, has been prepared at high pressure, which crystallizes with the orthorhombic PtSn,-type structure.' In the present paper we report the crystal structures and some physical properties of Os,Snl,, RhSn,, RhSn, and the ambient-pressure modification of IrSn,. Experimental Sample preparation Starting materials were powders of osmium (Ventron: 99.9%), rhodium (Matthey: 'reinst'), iridium (Ventron: 'reinst') and small granules of tin (Merck: 99%). Well crystallized samples of 0s4Snl7, RhSn, and IrSn, were obtained by annealing the transition metals with tin in the atomic ratio 1:10 in evacuated silica tubes at 550°C for 2 days, followed by 7 days at 240°C (0s4Snl7) or 5 days at 300°C (RhSn, and IrSn,) and sub- sequent quenching in water.The sample of RhSn, was prepared in a similar way. The atomic ratio was 1:3 and the annealing was at 550°C for 2 days, followed by 7 days at 330°C. The tin-rich matrix of the samples with the starting ratio 1:10 was ,RhSn, and IrSn,. The compounds do not become dissolved in moderately dilute (1:1) hydrochloric acid, which attacks the binary stannides at a slower rate. The sample of RhSn, was also treated with hydrochloric acid to separate the crystals; grinding or crushing is not well suited, because this compound has some ductility. Energy-dispersive X-ray analy- ses of the four stannides in a scanning electron microscope did not show any impurities with an atomic mass greater than that of sodium.X-Ray diffractometry Guinier powder diagrams of the samples were recorded with Cu-Ka, radiation using a-quartz (a=49 1.30 pm, c =540.46 pm) as an internal standard. The lattice constants were refined by least-squares fits. To ensure the proper assignment of the indices the observed intensities were compared with the calcu- lated" ones, ultimately using the positional parameters of the refined structures. Single crystals suitable for the collection of the intensity data were selected on the basis of Laue patterns. The data were determined with an Enraf-Nonius four-circle diffrac- tometer using graphite-monochromated Mo-Ka radiation and a scintillation counter with pulse-height discrimination. The scans were along 8 with background counts on both ends of each scan.Empirical absorption corrections were made on the basis of psi-scans. Further details of the data collection are summarized in Table 1. RhSn, was recognized to be isotypic with P-CoSn,, from the Guinier powder diagrams. The structures of the other stannides were determined from Patterson and difference Fourier syntheses using the program package SHELXL86.l' They were refined by a full-matrix least-squares program using the atomic scattering factors provided by the Enraf-Nonius SDP programs.12 The weighting schemes accounted for the counting statistics and parameters correcting for isotropic secondary extinction were optimized as least-squares parameters.The space group of RhSn, (14,lacd) is unique. The structure of Os,Sn,, was refined in the space group Pbcrn, the group J. Muter. Chern., 1996,6( 12), 1897-1903 1897 Table 1 Crystal data for Os,Sn17 , RhSn, ,RhSn, and IrSn,' 0s4Sn 17 RhSn, RhSn, IrSn, structure type 0s4Sn17 CoSn, IrGe, IrGe, Pearson symbol oP84 t164 hP15 hP15 formula mass 2778 5 458 98 577 7 667 0 space group Pbcm (no 57) 14,lacd (no 142) P3121 (no 152) P3,21 (no 152) alpm 694 9( 1) 632 4( 1) 677 4(2) 679 l(2) blpm 1428 l(3) clpm 1921 l(4) 3412( 1) 861 4(2) 857 5(3) v/nm3 1906 5 1364 6 342 3 342 4 z 4 16 3 3 calculated densitylg cm-, 9 68 8 94 8 41 9 70 crystal dimensions/pm3 20 x 20 x 40 lox 10x20 40 x 40 x 40 30 x 30 x 40 rdnge in h, k, I +11, +22, -30, +18 & 10, f10, +50 k12, f12, +15 +12, +12, f15 0120 scans up to 29 =70" 28 =70" 20 =80" 28=80" total number of reflections 17167 6097 8426 8414 highest/lowest transmission 135 145 122 1 80 unique reflections 4727 964 1421 1421 inner residual, R, 0 050 0 042 0 030 0 035 reflections with I, >3o(I,) 1573 23 1 1285 1165 number of vanables 102 21 26 26 highest residual peakle A 22 0 56 24 45 conventional residual, R 0 023 0 016 0 018 0 018 weighted residual, R, 0 023 0 015 0 022 0 020 ~~~~~~ ~~~ "Standard deviations in the place values of the last listed digits are given in parentheses throughout the paper with the highest symmetry compatible with the space group R, =O 089 for the incorrect one, even though the positional extinctions The situation was more complicated in the case of parameters were the same within three standard deviations the trigonal structures of RhSn, and IrSn, In these cases the As is usually the case for solid-state compounds, it can be structure refinements eventually resulted in the non-centrosym- expected that the samples of RhSn, and IrSn, contained metric enantiomorphous space groups f'3,21 and P3221, crystals of both space groups, P3121 and P3221, in equal respectively To facilitate comparisons, the dca of IrSn, were amounts The positional parameters were standardized using transformed to the other handedness (hkl-thkl) and therefore the program STRUCTURE TIDY l3 The final residuals, both structures are now described in P3121 The differences in atomic parameters and interatomic distances are listed in the residuals between the refinements with the wrong and the Tables 1-5 Anisotropic thermal parameters have been correct handedness were AR=0 001 and AR, =0 002 for deposited at the Cambridge Crystallographic Data Centre RhSn, For IrSn, the differences were much larger R =O 018 (CCDC) See Information for Authors, J Muter Chem, 1996, and R, =O 028 for the correct handedness and R =O 076 and Issue 1 Any request to the CCDC for this material should quote the full literature citation and the reference number Table 2 Atomic parameters of 0s4Snl7, RhSn,, RhSn, and IrSn,' 1145/17 atom X Y Z Be, Electrical conductivity and magnetic measurements Pbcm Single crystals of 0s4Snl7, RhSn, and IrSn, with the largest 8e 0 02744( 7) 0 2441 l(4) 0 11942(2) 0 323(6) dimension of approximately 0 5 mm were contacted with four 8e 0 48741(8) 0 00303(4) 0 12445(2) 0 360(6) copper filaments using a silver epoxy cement The electrical 8e 0 1353(2) 006061(7) 0 07301 (5) 0 64( 1) resistivities of these crystals were measured in the temperature 8e 0 1675(2) 041750(6) 0 09010(5) 0 59( 1) 8e 0 2087( 1) 0 61322(7) 0 17243(5) 0 68( 1) range 4-300 K using the van der Pauw technique 8e 0 3737(1) 0 18674( 7) 0 17172( 5) 0 63( 1) The magnetic susceptibilities of the binary stannides 8e 0 5589(1) 0 09903 (6) 0 00346( 6) 0 62( 1) Os,Sn,,, RhSn,, RhSn, and IrSn, were determined for poly- 8e 0 6832( 2) 0 33370( 7) 0 15041(6) 0 76( 1) crystalline samples, which did not show any impurity phases 4d 0 0480( 2) 0 3079( 1) 114 0 74( 2) on the Guinier powder diagrams A SQUID magnetometer 4d 0 2236( 2) 0 0063( 1) 114 0 90( 2) was used for these investigations at temperatures between 2 4d 0 3705(2) 0 4362( 1) 114 0 69(2) 4c 0 2598(2) 114 0 0 56(2) and 300 K with magnetic flux densities of up to 5 T 4c 0 8302( 2) 114 0 0 58(2) I4,lacd Results and Discussion 16d 0 114 0 33313(3) 0469(7) 16f 0 1599( 1) 0 4099 118 0 492( 7) The four stannides are stable in air for long penods of time 3% 0 1760( 1) 0 4270( 1) 0 54088(1) 0 667( 8) The crystals of Os,Sn,, have the shape of elongated prisms, P312I those of RhSn, and IrSn, have the form of approximately 3b 031119(6) 0 516 0 410(6) equidimensional polyhedra and the crystals of RhSn, are 6c 0 23646(4) 0 50036(4) 0 42990( 3) 0 672( 4) platelike As is usually observed, the preferred growth direc- 3a 009119(6) 0 113 0 664( 6) tions of these crystals are those with the short translation 3a 0 63043 (6) 0 1/3 0 685(6) penods All of these compounds are somewhat ductile, never- P3121 theless, they can be ground to a fine powder at room 3b 0 31220(4) 0 516 0 265(3) temperature6c 0 23334( 6) 049821(6) 0 43296(4) 0 523( 5) The well crystallized samples of the four stannides have tin- 3a 0 08982(8) 0 113 0 534(8) like metallic lustre, and as might be expected from their 3a 0 631 18(8) 0 113 0 523(8) composition, they show metallic conductivity (Fig 1) The "The equivalent isotropic thermal parameters Be, are listed in units room-temperature resistivities were found to vary between 30 of lo4pm2 and 70 p.sZ cm for Os,Sn,, and RhSn,, respectively However, 1898 J Muter Chem, 1996, 6(12), 1897-1903 Table 3 Interatomic distances in the structure of Os4SnI7" 267 3 267 4 268 8 272 0 273 4 277 7 280 7 286 8 273 9 274 4 276 4 277 1 277 6 279 0 287 5 288 7 303 0 276 4 286 8 295 2 309 6 272 0 277 1 295 2 302 1 315 7 322 4 326 8 339 1 354 6 362 0 375 5 377 0 392 1 395 2 268 8 279 0 298 0 308 6 314 3 315 1 322 4 273 4 288 7 300 8 303 3 308 6 309 6 316 1 322 2 351 1 369 8 386 7 392 1 274 4 287 5 286 5 294 7 299 6 322 6 326 8 327 9 344 6 273 9 277 7 303 3 318 7 319 7 322 6 324 5 326 5 329 0 375 5 378 2 382 6 395 2 267 3 289 5 319 7 322 2 340 4 354 6 362 4 303 0 Sn(9) 20s(2) Sn(7) Sn(8)2Sn( 3) 2Sn (6) 2Sn(2) 2Sn (4) Sn( 10) 20s( 1) Sn(l1) 2Sn(5) 2Sn(2) 2Sn( 1) 2Sn(4) Sn(l1) Sn( 11) 20s( 1) 2Sn( 5) Sn( 10) 2Sn(6) 2Sn( 1) 2Sn(2) 2Sn(3) 277 6 289 5 299 3 314 3 324 5 339 1 386 7 280 7 298 6 299 6 302 1 316 7 351 1 396 4 267 4 286 5 298 6 329 0 371 2 377 0 385 5 315 1 326 5 362 0 299 3 Sn( 10) 396 4 315 7 362 4 363 1 316 1 316 7 363 1 369 8 318 7 327 9 368 5 340 4 344 6 385 5 354 1 354 1 368 5 371 2 378 2 379 5 "All distances shorter than 415 pm are listed Standard deviations computed from those of the lattice parameters and the positional parameters are all 0 2 pm or less Table 4 Interatomic distances in RhSn," 13 pi2 cm depending on the direction," since higher resistivities can be expected for intermetallic phases than for a pure metallic Rh 1Rh 285 7 Sn(2) 1Rh 273 7 element The relative resistivities are more reliable They2Sn (2) 273 7 1Rh 274 4 2Sn( 2) 274 4 lSn(2) 294 3 decrease systematically with the temperature, as is typical for 4Sn( 1) 277 3 lSn(2) 315 7 metallic conductors The decrease for IrSn, is rather small and Sn(1) 4Rh 277 3 1Sn( 1) 322 2 this may be due to minor amounts of impurities, as is also 1Sn( 1) 286 I 2Sn(2) 329 4 suggested by the magnetic measurements 2Sn(2) 322 2 2Sn(2) 329 8 The magnetic susceptibility data of all four compounds 4Sn( 1) 336 1 1Sn( 1) 357 2 indicate Pauli paramagnetism (Fig 2) The susceptibilities at 2Sn(2) 357 3 1Sn( 1) 357 6 2Sn(2) 357 6 lSn(2) 368 6 room temperature (not corrected for the core diamagnetism) are very small They vary between 0 2 x and 3 5 x lop9m3 "All distances shorter than 420 pm are listed Standard deviations are formula unit (f u)-l The data were fitted to the modified all 0 1 pm or less Curie-Weiss law x=xo+C/(T-@) This resulted in the tem- perature-independent xo values of 3 0 x 0 06 x lop9, Table 5 Interatomic distances in RhSn, and IrSn4" 03 x and 0 8 x m3 (f u)-' for Os,Sn,,, RhSn,, RhSn, and IrSn,, respectively The moments calculated from Rh/Ir 2Sn(l) 270 11270 5 Sn(2) 2Rh/Ir 286 21286 4 the temperature-dependent terms varied between 0 11 and 2Sn( 1) 273 31274 6 2Sn( 2) 306 41304 7 0 48 pB (f u)-' for RhSn, and IrSn,, respectively Thus, they 2Sn(3) 273 81273 9 lSn(3) 312 11311 4 are much smaller than the value of 1 73 pB expected for one 2Sn( 2) 286 21286 4 2Sn( 1) 313 31313 4 unpaired electron per formula unit, and for this reason the Snl 1Rh/Ir 270 11270 5 2Sn( 1) 315 71313 5 lRh/Ir 273 31274 6 1 Sn( 3) 365 31367 6 corresponding upturns of the magnetic susceptibilities at low lSn(3) 298 61296 2 2Sn(3) 365 31364 5 temperatures should be ascribed to paramagnetic impurities lSn(2) 313 31313 4 2Sn( 1) 405 51407 6 or to paramagnetic surface states The temperature dependence lSn(2) 315 71313 5 Sn(3) 2Rh/Ir 273 81273 9 is largest for IrSn, This compound also showed a rather small lSn(3) 319 91323 0 2Sn( 1) 298 61296 2 temperature dependence of the electrical conductivity This lSn(1) 332 31332 0 lSn(2) 312 11311 4 sample may therefore have been contaminated by the homo- 2Sn( 1) 333 61333 9 2Sn( 1) 3 19 91323 0 2Sn( 1) 377 31380 1 lSn(2) 365 31367 6 geneous inclusion of an unknown impurity element lSn(3) 388 81388 1 2Sn (2) 365 31364 5 The magnetic susceptibilities were also determined with lSn(3) 401 41402 6 2Sn( 1) 388 81388 1 small magnetic flux densities, in particular at temperatures lSn(2) 405 51407 6 2Sn( 1) 401 41402 6 down to 2 K, the lowest temperature attainable with our instrument None of the samples showed highly negative "All distances shorter than 450 pm are listed Standard deviations are susceptibilities, as would be observed for a superconductor in all 0 1 pm or less a SQUID magnetometer, due to the Meissner-Ochsenfeld effect Slight discontinuities in the susceptibilities at ca 4K these values are affected by large errors of up to a factor of were ascribed to impurities of elemental tin, which becomes two, because of the difficulty in estimating the geometry of the superconducting below 3 7 K l6 small crystals and the size of the contacting areas Nevertheless, The most interesting results of the present investigation are these values compare well with those found for the tetragonal, the crystal structures of these tin-nch compounds Os,Sn,, metallic P-modification of tin, which vary between 9 and crystallizes with a new structure type The structure is rather J Muter Chem , 1996, 6(12), 1897-1903 1899 r I I II I I I I I I I 0 100 200 300 0 loo 200 300 TIK Fig.1 Relative electncal resistivities of (a) 0s4Snl7, (b) RhSn, and (c) IrSn4 between 4 and 300 K complicated with 84 atoms in the orthorhombic cell It may be visualized as consisting of several atomic layers, although the bonding within and between the layers (as can be concluded from the interatomic distances) is of comparable strength In Fig 3 we have emphasized (somewhat arbitrarily) three differ- ent kinds of layers, which we designate with the letters A, B and C Layer B is rather densely populated by a puckered network of four osmium and ten tin atoms Layer A contains onIy eight tin atoms, and only six tin atoms are sltuated on the mirror plane, which constitutes layer C Intermetallic phases are frequently characterized by close packing of all atoms If the atoms are of equal size, the structures usually contain both octahedral and tetrahedral voids, as is well known for the structures of the metallic elements, eg fcc and hcp packing The ‘tetrahedrally close- packed structures’, sometimes also called ‘o-phase related’ or ‘Frank-Kasper 22 contain only tetrahedral voids This is possible when the compounds consist of atoms with differing space requirements, resulting in different coordination numbers, as is well known for the Laves phases with the three prototype structures MgNi2, MgCu, and MgZn, The struc- ture of Os,Sn17 has a rather high tin content and therefore we cannot expect it to contain only small tetrahedral voids, however, it contains rather large voids We have not made an extensive search for such voids in the structure of Os4Sn17, however, as examples we show the positions of the five voids 1900 J Mater Chern, 1996, 6(12), 1897-1903 i I I I I I I 50 100 150 200 250 300 I4--a--1.5 1.o 0.5 1 I I I 1 I I 50 100 150 200 250 300 I I 1 I I I I 50 100 150 200 250 500 TIK Fig.2 Temperature dependence of the magnetic susceptibilities of (a) 0s4Snl7, (b) RhSn,, (c) RhSn, and (d) IrSn, measured with a magnetic flux density of 5 T Fig.3 The crystal structure of Os4Sn,,. A projection of the whole structure along the x direction is shown on the lower left-hand side. The structure may be visualized as consisting of the atomic layers A, B, C, B’, A’, B”, C’ and B”‘.Several of these layers, viewed along the translation period z, are shown on the right-hand side. Only the 0s-Sn bonds are indicated in the lower left-hand side projection. All Sn-Sn distances within the layers, shorter than 400 pm are indicated in the projection on the right-hand side. There are equally strong bonds between the layers; one such interface between the layers A, B”’ and B is shown in the upper left-hand corner. The positions of the voids V( 1)-V( 5) are indicated in the layers A, B and C. V( 1)-V( 5) (Fig. 3) and we list their positions and ‘coordi- nations’ (Table 6). The trigonal-prismatic void V( l), formed by six tin atoms at 245 pm is particularly large. It is even larger than the largest void in p-tin, which is ‘coordinated’ by four tin atoms at 225.7 pm and two more at 318.7 pm in a rather irregular arrangement, and it is almost as large as the tetrahedal void in the a(diamond) modification of tin, formed by four tin atoms at a distance of 281.0 pm.Sometimes such voids are filled by interstitial atoms; however, our final differ- ence Fourier analysis did not give any indication of this. We also found no related ‘filled’ structures in searching Pearson’s handb~ok.’~ The osmium atoms of Os,Sn,, occupy two different sites. The Os(1) atoms have eight tin neighbours in a distorted square-antiprismatic arrangement at distances varying between 267.3 and 286.8 pm with an average distance of 274.3 pm.The Os(2) atoms are situated in a monocapped square antiprism of tin atoms with a larger range of Os(2)--Sn distances extending from 273.9 to 303.0 pm. Both the larger coordination number and the larger spread should resuIt in larger Os(2)-Sn distances, and this is indeed the case with an average of 282.0 pm. There are eleven different tin sites in Os,Snl,. All tin atoms have two osmium and between ten and twelve tin neighbours (Table 3, Fig. 4). However, in contrast to the well defined coordinations of the osmium atoms, the coordination J. Muter. Chem., 1996,6( 12), 1897-1903 1901 Table 6 Location and coordination/pm of unoccupied sites [voids V( 1)-V(5)] in the structure of Os,Sn,, Pbcm X Y Z V(1)V(2) 4d 8e 0 6396 0 0929 0 2281 0 5916 114 0 9863 V(3) V(4)V(5) 4d 4d 8e 0 4840 0 9575 0 0760 0 7895 0 6442 0 9315 114 114 0 1645 V(1) 2Sn(3) 2454 V(3 2Sn(4) 2322 V(5) lSn(6) 2196 2Sn(4) 2454 2Sn(6) 2326 lSn(8) 2212 2Sn(6) V(2) lSn(2) 2455 2334 V(4 lSn(9) lSn(8) 2326 2337 lSn(2) lSn(7) 2224 2561 lSn(l1) 2339 lSn(1) 2342 2Sn(3) lSn(7) 2338 2338 lSn(1) lOs(1) 2581 2899 lSn(5) lSn(1) 2445 2472 2Sn(4) 20s(l) 2816 2888 Fig.4 Near-neighbour environments in the structure of Os,Sn,, The site symmetries are given in parentheses polyhedra of the tin sites are not easy to define The Sn-Sn distances almost continuously cover the range between 286 5 [Sn(5)-Sn(ll)] and 3964 pm [Sn(lO)-Sn(ll)] Thus, the shortest Sn-Sn distance of 286 5 pm is only slightly greater than the two-electron bond distance of 281 0 pm in the 1902 J Muter Chew, 1996, 6(12), 1897-1903 diamond (a)modification of elemental and the great variety of Sn-Sn bond lengths in Os,Sn,, is reminiscent of the Sn-Sn interactions in the P-modification of tin, where each tin atom has 4 +2 +4 Sn neighbours at 302 2, 318 1 and 376 8 pm, respectively 24 RhSn, crystallizes with a structure type which was deter- mined first for the high-temperature (p) modification of CoSn, , While there are no 0s-0s bonds in Os,Sn,,, the higher transition-metal content of RhSn, is reflected in the coordi- nation polyhedron of the rhodium atoms, which consists of one rhodium atom at the bonding distance of 285 7 pm and eight tin atoms at the almost equal distances of 273 7 (2 x), 274 4 (2 x) and 277 3 pm (4 x) The two different tin atoms of RhSn, have coordination numbers 15 and 12 The Sn(1)atoms of RhSn, are the only tin atoms in the structures of the present investigation which do not have two transition-metal neigh- bours They have four rhodium and eleven tin neighbours, while the Sn(2) atoms of this stannide are coordinated by two rhodium and ten tin atoms Again, the Sn-Sn distances cover a wide range extending from 286 1 to 368 6 pm For a further discussion of the P-CoSn,-type structure, refer to ref 3 The stannides RhSn, and IrSn, are isotypic with a structure (Fig 5) first described for IrGe, 25 The c/a ratios of these trigonal cells are slightly different, but the cell volumes are practically the same (Table l), and the interatomic distances in the two compounds are also very similar (Table 5) The rhodium atoms (the corresponding values for the iridium compound are listed in parentheses) are coordinated by eight tin atoms at an average distance of 275 8 (276 4) pm forming a distorted square antiprism (Fig 6), similar to the environ- ment of the Os(1) atoms in Os,Sn,, The three different tin atoms all have two Rh (Ir) and twelve Sn neighbours and, as was discussed above for Os,Sn,, and RhSn,, the Sn-Sn distances cover a wide range, almost continuously extending from 298 6 (2962) pm to 405 5 (407 6) pm There are no further Sn-Sn distances up to 450 pm The average Sn-Sn distances of 349 8 (350 4), 340 8 (340 5) and 352 1 (352 3) pm for the Sn(l), Sn(2) and Sn(3) atoms reflected the average Sn-Rh(1r) bond lengths of 271 7 (272 5), 286 2 (286 4) and 273 8 (273 9) pm, respectively, I e the longest Sn-Rh(1r) bonds are found for the Sn(2) atom, which has the shortest average Sn-Sn distances We conclude with a remark about the thermal parameters Usually these are strongly affected by absorption errors, which cannot fully be accounted for by the correction from psi-scan data The relative values within one compound, however, are much more reliable It can be seen (Table 2), that the thermal parameters of the transition elements are all smaller than those of the tin atoms This is also the case for the stannides MoSn,,' VSn2,2 NbSn2,2 CrSn22 and the two modifications of CoSn,,, even though the transition metals are sometimes lighter and sometimes heavier than the tin atoms We believe this reflects Fig.5 Projection of the hexagonal IrGe,-type structure of RhSn, and IrSn, RNlr (2) 4 Fig. 6 Coordination polyhedra and site symmetries in the stannides RhSn, and IrSn, the regular coordination of the transition-metal atoms on the one hand and the irregular coordinations of the tin atoms on the other, the latter with relatively large voids in their vicinity, as we have analysed in more detail for Os,Sn,,. We thank Dip1.-Ing. U. Rodewald, Dip1.-Chem. M. Gerdes and Mr. K. Wagner for the intensity data collection on the four-circle diffractometer, for the magnetic susceptibility measurements in the SQUID magnetometer and for the EDX investigation.Dr. R-D. Hoffmann contributed to this work in the early stages of the structure determinations and refinements. We are also indebted to Dr. G. Hofer (Heraeus Quarzschmelze, Hanau) for a generous gift of silica tubes. This work was supported by the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie. References 1 T. Wolpl and W. Jeitschko, 2. Anorg. Allg. Chem., 1994,620,467. 2 T. Wolpl and W. Jeitschko, J. Alloys Compd., 1994,210, 185. 3 A. Lang and W. Jeitschko, Z. Metallkde., 1996,in press. 4 K. Schubert, H. L. Lukas, H-G. Meissner and S. Bahn, Z. Metallkde., 1959,50, 534. 5 K. Schubert, Z. Naturforsch., Teil A, 1947,2,120. 6 P.Bussibe and K. Lazar, Hyperfine Interact., 1988,41,559. 7 P. Bussikre, M. Boge and K. Lazar, Hyperfine Interact., 1990, 54, 775. 8 C. Venturini, B. Malaman and B. Roques, cited in ref, 6 and 7. 9 V. I. Larchev and S. V. Popova, J. Less-Common Met., 1984,98, L1. 10 K. Yvon, W. Jeitschko and E. Parthe, J. Appl. Crystallogr., 1977, 10, 73. 11 G. M. Sheldrick, SHELXL86, a computer program for crystal structure determination, Universitat Gottingen, Germany, 1986. 12 B. A. Frenz & Associates, Inc. and Enraf-Nonius, Structure Determination Package V3.0, College Station, Texas, USA, and Delft, Holland, 1986. 13 L. M. Gelato and E. Parthe, J. Appl. Crystallogr., 1987,20, 139. 14 L. J. van der Pauw, Philips Res. Rep., 1958,13,1. 15 E. Gruneisen, Ergeb. Exakten Naturwiss., 1945,21, 50. 16 B. W. Roberts, J. Phys. Chem. Ref. Data, 1976,5, 581. 17 F. C. Frank and J. S. Kasper, Acta Crystallogr., 1958,11, 184. 18 F. C. Frank and J. S. Kasper, Acta Crystallogr., 1959,12,483. 19 K. Schubert, Kristallstrukturen zweikomponentiger Phasen, Springer, Berlin, 1964. 20 S. Samson, in Structural Developments in Alloy Phases, ed. B. C. Giessen, Plenum Press, New York, 1969. 21 C. B. Shoemaker and D. P. Shoemaker, Monatsh. Chem., 1971, 102,1643. 22 W. B. Pearson, The Crystal Chemistry and Physics of Metals and Alloys, Wiley, New York, 1972. 23 P. Villars and L. D. Calvert, Pearson’s Handbook of Crystallographic Data for Intermetallic Phases, ASM International, Materials Park, Ohio, 2nd edn., 1991. 24 J. Donohue, The Structures of the Elements, Wiley, New York, 1974. 25 P. K. Panday and K. Schubert, J. Less-Common Met., 1969,18,175. Paper 6/04697E; Received 4th July, 1996 J. Muter. Chem., 1996,6( 12), 1897-1903 1903

ISSN:0959-9428    

DOI:10.1039/JM9960601897

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Magnetic iron oxide-silica nanocomposites. Synthesis and characterization Corinne ChanCac, Elisabeth Tronc and Jean Pierre Jolivet Chimie de la Matih-e Condensie, URA-CNRS 1466, Universiti Pierre et Marie Curie, T54, E 5, 4 Place Jussieu, 75252 Paris Cedex 05, France Composite materials containing nanoparticles of maghaemite (y-Fe203) dispersed in a silica matrix have been made by polymerizing a silica precursor (triethoxysilane or silicic acid) inside an aqueous sol of maghaemite particles. After gelation, the examination of xerogels by electron microscopy does not reveal noticeable aggregation of particles. The structure and composition of the silica matrices were deduced from ,'Si MAS NMR spectroscopy. Thermal analysis, FTIR and NIR spectroscopic studies showed that the particles and silanol groups of the matrix remain solvated in the composite materials.No Si-0-Fe bonds are formed in the xerogels and the dispersion of particles in the matrix seems to result from the mutual solvation of particle surfaces and remaining silanol groups, as indicated by strongly associated hydrogen-bonded water molecules. Nanophase materials are interesting for various technological applications because of the specifically size-related properties (mechanical, electronic, optical, magnetic, etc.) of crystalline domains or particles.lY2 Nanoparticles of ferrimagnetic oxides are typically characterized by the superparamagnetic relaxation phenomenon, which is strongly dependent on the particle size and shape, on the magnetic interactions between particles and on various surface effect^.^ The synthesis of calibrated and well dispersed particles into a rigid matrix is consequently a pre- requisite for the study of such phen~mena.~ Composites formed by nanoparticles of maghaemite, y-Fe203, a spinel iron oxide, dispersed in silica have been tentatively synthesized by heating a mixture of iron nitrate and silicon alkoxide between 700 and 900°C.5*6 The size and the dispersion of the maghaemite particles formed during the thermal treatment seem to be determined by the porosity of the silica matrix.It is, however, difficult to obtain only the spinel phase in the form of calibrated and well dispersed particles. Other attempts consisted of making multiple coatings of silica on more or less oxidized magnetite particles peptized in alkaline medium, but the aggregation of particles always seems to OCCU~.~*~ We showed previously that composites with well dispersed y-Fe,03 particles can be obtained by in situ polymerization of a silica precursor in a sol of well dispersed y-Fez03 particles, and we focused on the behaviour of the composites at high temperatures.' Here, we report the synthesis of composites from a different silica precursor and we focus on the charac- terization of the interactions between particles of varying mean size and the silica matrices by IR spectroscopies and thermal analysis.Experimental Syntheses y-Fe,03 aqueous sols. y-Fe203 nanoparticles result from the oxidation of nanoparticles of stoichiometric magnetite, Fe,O,.They were prepared as described previously," by coprecipi- tation of Fe( NO3), and FeC1, (Fe"/Fe"' =0.5) in alkaline medium under vigorous stirring. The control of the precipi- tation conditions [pH, nature of the base (NaOH or NH,), ionic strength imposed by a salt, e.g. NaNO,] allowed us to prepare calibrated particles with a mean size in the range 4-10 nm.'l,12 The black precipitate immediately formed was first decanted by magnetic settling on a permanent magnet and then isolated and treated with a concentrated HClO, (3 mol 1-l) solution. The Fe" ions released in solution13 were eliminated after centrifugation. The precipitate was treated repeatedly with the perchloric acid solution until the Fe"/Fe"' ratio in the solid was ca.0.01. After the last separation by centrifugation, the particles were dispersed in pure water, giving a stable sol of y-Fe203 particles (iron concentration, 1 mol 1-l) at pHz2. Under such conditions, the aggregation of the particles, which are electrostatically positively charged (C104- counter ions), is at a minimum because of the high surface charge density (ca. 0.9 C mV2) and the low ionic strength.', Dispersion in a silica matrix. Two precursors of silica glasses, triethoxysilane and silicic acid, were used to prepare iron oxide-silica nanocomposites. In acidic aqueous medium, hydrolysis and condensation of the precursors yield a gel which leads, after drying, to a transparent monolithic gla~s.'~,~~ By mixing the precursor with the aqueous maghaemite sol, hydrolysis and condensation of the precursor take place in situ, the gelation being catalysed by the acidity of the sol.Such a procedure leads to homogeneous solid composites. Triethoxysilane precursor. Acidic water contained in the maghaemite solution was used for hydrolysis and condensation of triethoxysilane according to the rea~tion.','~ HSi(OCH,CH,), +3/2 H20-+HSiOl.,+3 CH,CH,OH The alkoxysilane was added to the sol in a proportion corre- sponding to a molar ratio H,O/Si>2. In every case, under vigorous stirring, an emulsion formed initially, then a homo- geneous solution and finally a gel after 30-45 mn at room temperature. The gelation time seems to be independent of the particle concentration. After drying at room temperature, a brown monolith was obtained.Samples with an Fe/Si ratio varying between 0.03 and 1.6 were thus prepared. Fe/Si=O.O3 corresponds to H,O/Si =2 using a sol with an Fe concentration of 1moll-l. Fe/Si=1.6 corresponds to the amount of precur- sor just needed to cover particles of diameter 10nm with a monolayer of silica. Silicic acid precursor. Na,Si03 (1 mol 1-l) was exchanged on a Dowex 50 W X2 resin in H-form and a solution of silicic acid (ca. 0.45 mol l-l, pH M 3) was collected. The pH was adjusted to 2 by adding perchloric acid. For preparing com- posites with Fe/Si 20.1, the silicic acid solution was introduced directly into the sol of y-Fe20, particles. For low particle concentrations (Fe/Si <O.l), the same procedure leads to a fast flocculation of the particles.In order to prevent flocculation, J. Muter. Chem., 1996, 6(12), 1905-1911 1905 it is necessary to control the pH during the mixing The silicic acid solution was thus introduced drop by drop into the sol and the pH of the mixture was kept at 2 by adding perchloric acid We prepared samples with Fe/Si varying between 003 and 0 7 In all cases, homogeneous gels were obtained after a few days at 50 "C, the gelation time decreasing with increasing particle concentration Drying at room temperature yielded homogeneous brown monoliths Aggregation phenomena observed for low Fe/Si ratios pre- sumably result from the acidic catalysis of the condensation of silicic acid l6 This reaction initially consumes protons from solution as it is observed when hydrochloric acid is mixed with silicic acid Thus, when a small quantity of sol (pH=2) is introduced into a large excess of silicic acid (pH32 5, Fe/Si <0 l), the reaction of silicic acid at the particle surface rapidly lowers the surface charge density and leads to the flocculation'' before the proton adsorption equilibna at pH 32 5 take place As aggregates are embedded into silicic acid polymers, the surface cannot be recharged On the other hand, if protons are supplied by perchloric acid addition, the flocculation is prevented If the concentration of particles is higher (Fe/Si >0 l),the amount of protons introduced is larger This allows the protonation of silicic acid with a lesser relative decrease in surface charge density Consequently, the stability of the colloids is not affected significantly and no aggregation occurs Techniques Transmission electron microscopy (TEM) experiments were performed using a JEOL 100 CX I1 apparatus Observations of the particles alone were achieved by evaporating a drop of the very dilute aqueous sols onto a carbon-coated grid The composites were examined in the form of thin sections, 80-100 nm thick X-Ray diffraction (XRD) patterns were recorded using a powder diffractometer (Philips PW 1830) operating in the reflection mode with Cu-Ka radiation The average particle size was deduced from the line broadening by applying the Scherrer formula assuming Gaussian profiles for experimental and instrumental broadenings Quasi-elastic light scattering measurements were performed on aqueous sols with an Amtec SM200 apparatus equipped with a Brookhaven BI2030 correlator These were used to control the lack of significant aggregation in the sols Thermogravimetry and differential thermal analyses (TG-DTA) were effected using a Netzsch STA 409 apparatus IR spectroscopic studies were performed in the range 1800-400 cm-' using a FTIR Nicolet 550 spectrometer Samples dehydrated at room temperature were pelleted with dned KBr (1% by mass) Near-IR spectroscopic studies in the wavelength range 2500-800 nm were performed in a closed cell using a Vanan Cary 5E spectrometer with a 240 nm min-l scanning rate 29S1 NMR spectra of silicic acid solutions were obtained using a Bruker AM 250 spectrometer (49 6 MHz) The probe signal was avoided by the use of a Hahn echo sequence with a delay of 30ms The silica xerogels were studied using a Bruker MSL 300 spectrometer (59 6 MHz) with standard single-phase or cross-polanzation techniques under magic angle spinning (MAS) conditions Results and Discussion Fig la shows the XRD patterns of y-Fe20, particles of different mean sizes The lattice parameter is 0 835 nm, independent of the particle size The line broadening is essentially due to the size effect The average sizes, deduced from the full width at half maximum, are consistent with the mean sizes deduced 1906 J Muter Chem , 1996, 6(12), 1905-1911 A 81 I I 1 I I I I bI C 10 20 30 40 50 60 70 80 2Bldeg rees Fig.1 X-Ray diffraction patterns of samples dried at room tempera- ture a, y-Fe,O, particles of different average diameter, D, composites made from b, triethoxysilane (D = 10 nm) and c, silicic acid (D=7 4nm) from TEM observations In the composites, the average particle size remains unaltered (Fig lb,c) Fig 2a and b show micrographs of composites made from triethoxysilane and silicic acid, respectively In each case, the particles appear to be well separated The condensation of silicic acid is much slower than that of the hydrolysis products of triethoxysilane In both cases, the polymerization of the precursors inside the sol allows particles to be trapped without significant aggregation, and allows the preparation of dispersed materials in a silica matrix over a large range of iron oxide concentration However, the two matrices are chemically very distinct which leads to different effects on the particles during thermal treatment of the composites (see above) The dispersability of the particles in the silica matrices may result from various types of interactions covalent, through Si-0-Fe bond formation, electrostatic, between negatively charged Sir0 terminal ligands and positively charged groups on the particle surface, or by hydrogen-bond interactions between hydration layers of silanol groups and the particle surface In order to characterize the interactions between the particles and the silica matrices, we studied separately the pure silica xerogels and the particles in various states in the initial aqueous sol, after drying at room temperature and in the composites Maghaemite particles For particles with an average diameter of 5 nm, as deduced from XRD and TEM, the average hydrodynamic diameter measured by light scattering on a dilute sol at pH=2 is ca 8 nm The large difference in the two sizes shows that a large amount of water is taken up by the particles in their thermal motion in the sol, which results from the solvation of the positively charged surface groups Of course, a rapid exchange between water molecules in the solvation layer and the bulk of the sol must occur Fig* TEM images Of Y-Fe203 in matrix Obtained: a,from triethoxysilane (10 nm y-Fe203 particles, molar ratio Fe/Si = 0.03); b, from silicic acid (7.4 nm y-Fe203 particles, molar ratio Fe/Si =0.07) Fig.3a shows typical DTA-TG curves of 5 nm y-Fe203 particles. They indicate a loss of water up to the transformation into a-Fe203, at ca. 400-500°C depending on the sample. The water hydration retained by the particles after drying depends strongly on the mode of dessication. As shown in Fig. 4 the thermal treatment of powders obtained by drying the sol at pH =2 for a few days at room temperature (RT)proceeds with three successive mass losses, which represent in all ca. 12% by mass for 10 nm particles (Fig. 4a) and 25% for 5 nm particles (Fig. 4b); they take place from ca.RT to 180 "C, from 200 to 240°C and beyond 250"C, respectively. The first mass loss corresponds to the elimination of hydration water which is relatively weakly bonded (physisorbed) to the particles. For both particle sizes, this hydration represents ca. three water layers around each particle, taking into account the mean volume of the water molecules (3 xlop2nm3) in the liquid phase. The second mass loss corresponds to the removal of the perchlorate counter ions which are always present because the particles remain electrostatically charged after drying. The Fig. 3 DTA-TG curves: a, of 5 nm y-Fe203 particles; b, of composites from silicic acid (molar ratio Fe/Si =0.7) 0. -2 ' -4 -8 h -6j8-101v h $-12C.,, , ,,.,.I.., ,,.., , v uJo8 -4'-8 -12 -16 -20 -24 0 100 200 300 400 T1"C Fig. 4 Thermogravimetry curves: a, 10 nm y-Fe203 particles, pure silica xerogel from triethoxysilane and composites [SiHo *O, 5(OH)0(O), b, 5 nm y-Fe203 Fe/Si=O.17 (m),Fe/Si=1.6 (O), y-Fe203 (O)]; particles, pure silica xerogel from silicic acid and composites [SOl 75(OH)o (O),Fe/Si =0.03 (W), Fe/Si =0.7 (O),y-Fe,03 (O)].The y-Fe,03 particles were isolated from aqueous sols (pH =2) and dried at temperature. presence of perchlorate was shown by its characteristic IR absorption bands (see below), which are still present for samples heated at 180 "C but are absent for samples heated at 240 "C. Such a temperature range for the removal of perchlorate as perchloric acid corresponds well to the boiling point (200 "C) of hydrated perchloric acid." The amount of perchlorate remaining after drying of the sample is, however, smaller than that expected from the surface electrostatic charge of the particles determined in s01ution.l~ This is an effect of the drying process.The drying of particles, involving their progress- ive concentration, induces a relaxation phenomenon of the double layer between the oxide surfaces." The increase of the electrostatic interactions between the particle surfaces leads to a partial desorption of protons from charged surface groups as perchloric acid, which evaporates during the drying. Above 250"C, the mass loss results from the removal of strongly adsorbed water and from the dehydration of surface hydroxy ligands.A sudden small mass loss is systematically observed (Fig. 3a) when the y-Fe20, +a-Fe203 transformation occurs. As the transformation begins by the sintering of aggregated J. Muter. Chem., 1996,6(12), 1905-1911 1907 particles,20 water trapped inside the aggregates is suddenly removed during the recrystallization of the aggregates When the drying conditions are harsher (at 50°C or at RT in the presence of P205),the total mass loss, of 5% for 10 nm particles and 7 5% for 5 nm particles, represents only one water layer around each particle Perchlorate is still present, but in a smaller amount, and the extent of hydration is comparable with that existing on particles flocculated at pH=7 (near the point of zero charge) and then dried FTIR spectra of the particles dried at RT (Fig 5a,b) show one band in the range 630-580 cm-' (denoted +) correspond-~-~ing to the stretching vibration v~ of tetrahedral iron atoms Perchlorate groups present as counter ions exhibit two bands at 1110 and 630 cm-' (denoted *) 21 Silica ma trices The matnx has a very different chemical composition depending on the precursor For tnethoxysilane-based silica, the relative amounts of various sites (Table 1) were deduced from 29S1 MAS NMR spectra and refined by 29S1 CP-MAS NMR using MAS data The NMR studies of this pure silica matrix show that, under the hydrolysis conditions employed, 80% of the Si-H bonds of the precursor are preserved The elimination of ethoxy groups is quasi-complete and the silicium atoms are in different environments The high proportion of trifunctional silicium atoms (HSiO, groups) suggests a rela- tively low degree of crosslinking of the matrix of mean composi- tion SiH, 5(OH), The condensation of pure silicic acid was studied by 29S1 NMR spectroscopy in solution One hour after the exchange, no monomer was observed The monomers condense rapidly to form oligomers which aggregate to form a in which 67% of the silicon atoms carry at least one Si-OH terminal bond After gelation and drying at room temperature, the relative proportions of the different groups in the xerogel (Table2) deduced from the 29S1 MAS NMR study, lead to a mean composition of SiO, 75(OH)0 In view of the chemical compositions, the matnx appears more crosslinked and more hydroxylated than that formed by triethoxysilane hydrolysis The xerogel formed from triethoxysilane is therefore more flexible than the xerogel obtained from silicic acid The pure xerogel obtained from triethoxysilane is thermally * I I I I 1*1 1 I I I I I 191 J 1800 1400 lo00 600 vlcm Fig.5 FTIR spectra of pure xerogels, y-Fe,O, particles and nanocom- posites a, from ethoxysilane, b, from silicic acid (symbols see Table 3) 1908 J Muter Chem, 1996, 6(12), 1905-1911 Table 1 Distribution of the different groups into a pure silica matrix formed by hydrolysis of tnethoxysilane (H,O/Si =2, pH =2) 6 YO OH I -0-Sl-O-I -76 6 13 5 H I 0 -0-Sl-O--85 1 66 2 IH OH I -0-SI-0 -I -101 5 0 I -0-SI-0 --111 1 13 5 I 0 Table 2 Distnbution of the different groups into a pure silica matrix formed by condensation of silicic acid (pH z 2, Ca=0 45 mol I I) 6 % OHI -0-SI-OHI -91 5 710I OHI -0-SI-0 --101 1 38 5 0 I I0 -O-SI-O--1100 54 4 I 0 I stable up to 350°C17 and exhibits only a slight mass loss (<2%) at ca 100 "C (Fig 4a) corresponding to the elimination of hydration water Above 350"C, the cleavage of the Si-H bond occurs, up to the crystallization of the silica network into cristobalite near 1400 "C Up to 350 "C the pure silica xerogel obtained from silicic acid dned at RT (Fig 4b) shows similar behaviour 22 However, the water loss at ca 150°C reaches 17% by mass showing greater hydration of the xerogel in agreement with the NMR analysis Beyond 450"C, a con- tinuous thermal condensation of the Si-OH groups, rep- resented by a slight mass loss (2-3%), occurs up to the transformation of the xerogel into glass The IR absorption bands of the silica networks obtained from tnethoxylsilane or from silicic acid in the range 1800-400 cm-' (Fig 5a,b) are listed in Table 3 on the basis of the absorption spectra of conventional vitreous silica l5 23 24 Silanol group vibrations appear at 3680-3650 cm-' if the SiO-H group is isolated and at 3400 cm-' if it is hydrogen bonded" However, in silica xerogel spectra, these bands are very broad and are not distinguishable from one another because of superimposition of stretching vibrations of hydro- gen-bonded water molecules adsorbed on the surface 25 The Table 3 Assignments of the absorption bands (in cm-') in pure SiH, 8015(OH)0 and SiO, 75(OH)0 xerogels SiHO 8(OH)0Zol 5 Si(OH)O 5Ol 75 v,, (Sl-0-S1) v 1154 (LO) 1205" 1065 (TO) 1082' V, (Si-0-S1) A 824 797 v (Si-OH) 0 936 959 -6 (Si-H) 0 880 6 (Si-0-Sl) 0 452 459 " Longitudinal mode.'Transverse mode. contributions of various silanol species and water molecules to the envelope of vibrations near 3400cm-' may be deter- mined using near-IR spectro~copy.'~ Near-IR studies have been performed in the wavelength range 2500-800nm where the main groups of absorption bands of Si-OH and H20 have been a~signed.~~,~~ In prin- ciple, it is possible to separate free or hydrogen-bonded silanol groups2, and isolated or associated H20 molecules.The spec- trum of the pure xerogel obtained from silicic acid (Fig. 6) essentially shows the vibrations of the silanol groups and water molecules.29 Three main domains can be distinguished. In the 2400-2250 nm region only the stretching vibrations of Si- OH groups and a contribution from the bulk of the matrix arise. In the 2050-1850nm region, there are the combinations of stretching and deformation vibrations of water, and in the range 1600-1300 nm are the overtones of the stretching fre- quencies of silanol groups and water molecules. For each vibration, the band is asymmetrical.At low energy, the main band corresponds to isolated groups, free silanol or isolated water hydrogen-bonded to a silanol group. At higher energy, the shoulder corresponds to hydrogen-bonded groups, silanol groups or hydrogen-bonded water molecules.30 Composites Fig. 3b shows typical TG-DTA curves for a silicic acid-based composite. The exothermic peak characteristic of the y+a-Fe203 transformation (Fig. 3a) is no longer observed. XRD patterns of composites after heating at various temperatures up to 1400°C are shown in Fig. 7. It is clear that the y-Fe203 particles are stabilized in the composite up to at least 1000"C. At 1200"C, is the major iron oxide phase and a-Fe203 appears only beyond 1200 "C. The matrix containing the particles acts as an antisintering agent, and stabilizes the spinel structure.In composites with low particle concentration, structural transformation of the iron oxide occurs only when the matrix crystallizes into cristobalite. Similar features were observed with the triethoxysilane-based composites in an oxid- izing atmosphere.' Below ca. 400"C, all composite materials exhibit the com- bined thermal behaviour of the maghaemite particles and the pure matrix (Fig. 4a,b). This suggests that the formation of Isolaud H20 ,- - +Bonded SiOH IsolatedH70 0.3 . Q,0c ([I4! 0.2 . 8n I ([I 0.1 1 I I I I 1 1200 1400 1600 1800 2000 2200 2400 Alnm Fig. 6 Reflectance NIR spectrum of SiO, 75(OH)* xerogel dned at 100°C for 2 h Cnstobi E-Fe,03 + a-Fe203I II,1 10 20 30 40 50 70 80 2Bldeg rees Fig.7 X-Ray diffraction patterns of a silicic acid-based composite (Fe/Si =0.07) after heating at different temperatures: a, 25 "C; b, 750°C; c, 1000°C; d, 1200°C; e, 1400°C nanocomposites involves neither the dehydration of the par- ticles nor that of the matrix. The FTIR spectra of RT-dried nanocomposites obtained with the highest and intermediate particle concentrations [Fe/Si= 1.6 and 0.17 for SiH, ,O, 5(OH), matrix and Fe/ Si=0.7 and 0.07 for SiOl 75(OH)o matrix] are shown Fig. 5. The absorption bands due to the matrix and those due to the iron oxide are not shifted. No additional band characteristic of Si-0-Fe bonds appears, especially around 900 or 680 cm-', as is observed for Fe-substituted silicalite in ZSM5 zeolites33 or ferrisilicates with the sodalite structure.34 The spectra of the composites can thus be described as superim- posed spectra of the iron oxide and of the silica matrix.In Fig. 8 are shown FTIR spectra of composites after heating at high temperature and characterization using the XRD patterns given in Fig. 7. The continued polycondensation of the silica matrix manifests itself (Fig. 8) by the decreasing intensity of the 970cm-' band assigned to Si-OH vibrations and the evolution of Si-0-Si vibration bands at 1100, 790 and 500 cm-I is in agreement with the structural transformation of the silica network. After heating at 1200 and 1400"C, well resolved bands appear at 580-620 cm-', due to Fe-0-Fe bonds in E-and a-Fe,03.35,36 In these spectra, there is no band which can reasonably be attributed to Si-0-Fe bonds.In order to prove the existence of weak interactions between the surface of the particles and the matrix, it is important to examine the spectral range 3800-3000 crn-', characteristic of the stretching vibration of SiO- H and water molecules. Near-IR investigations on silicic acid-based composites were carried out on materials dried for 4 days at various tempera- tures up to 450°C and corresponding to different degrees of dehydration. The spectra (Fig. 9) of all samples, pure Si(OH), 501 xerogel and composites, show bands located 75 between 2100 and 1250 nm characteristic of both isolated and l...I.V.l...l...l...I 1600 1400 1200 loo0 800 600 400 v1crn-l Fig.8 FTIR spectra of a silicic acid-based composite (Fe/Si = 0.07) heated at different temperatures: a, 25°C; b, 750°C; c, 1000°C; d, 1200°C; e, 1400°C. J. Muter. Chern., 1996, 6(12), 1905-1911 1909 2.5 I 1 0.8 0.6 0.4 0.2 n $j 0.8 -0a 0.8 A 0.6 0.6 0.4 0.4 0.2 0.2 ,......IL 7200 1400 1600 lsoo 2000 Xlnm Fig. 9 NIR spectra of silicic acid-based samples thermally treated for 4 days at different temperatures: A, 25°C; B, 100°C; C, 250°C; D, 450 "C. a, Pure xerogel; composites with 7.4 nm y-Fe,O, particles and molar ratios b, Fe/Si=O.O3 and c, Fe/Si=O.O7; d, 4.5 nm y-Fe203 particles and Fe/Si =0.07. associated water molecules. After dehydration above 100 "C, the band at 1900nm becomes more defined.The hydrogen- bonded water molecules are eliminated progressively as shown by the relative decrease of the shoulder near 1950nm and some isolated water molecules attached to Si-OH groups are still present. Simultaneously, free silanol groups appear as shown by the sharp peak at 1360 nm. After thermal treatment at 25O-45O0C, the spectra are practically the same for all samples. This is in good agreement with the quasi-complete dehydration of the samples as shown by the TG curves (Fig. 4b). The near-IR spectra (Fig. 9) do not allow us to differentiate clearly the hydration of the matrix from that of the particles. However, for the pure silica matrix, free silanols are the main groups remaining after prolonged drying at 100°C and no evolution is observed after heating to 250°C (Fig.9a). In the composites, (Fig. 9b,c,d), the band near 1450 nm, characteristic of associated water molecules, is still present after treatment at 250 "C and such hydration water probably forms multilayers between the particles and the matrix, as shown schematically in Fig. 10. This also suggests that the surface of the iron oxide particles is more strongly solvated than the matrix and the remaining water is in fact essentially a solvation layer of the particles. This could indicate that the dispersal of the particles into the silica matrix results from solvation of silanol groups Multilayers ,,-HIk..._ 6 Fig. 10 Schematic representation of the interactions between the surface of the y-Fe20, particles and the silica matrix 1910 J.Muter. Chem., 1996, 6(12), 1905-1911 of the matrix by the associated water layers around the colloids, without other chemical surface interactions. Very similar results are obtained for the composites made from ethoxysilane in the same thermal treatment range. An advantage given by the ethoxysilane precursor is the rapidity of the gelation. This matrix is much less hydrated than the matrix formed from silicic acid. It is also less crosslinked which is likely to favour particle dispersal. The behaviour of these matrices becomes very different at high temperatures in an inert atmosphere when the cleavage of the Si-H bond leads to the reduction of the iron oxide particles to U-F~.~ Preliminary Miissbauer spectroscopy investigations of the same y-Fe203 particles dispersed, at the same concentration, in silica matrices or in polyvinylic alcohol4 indicate no signifi- cant changes either in the low-temperature spectra or in their temperature dependence, Since the superparamagnetic relax- ation rate is, in principle, very sensitive to surface effect^,^' these observations support the results of the IR spectroscopic studies, namely the presence of only weak interactions between the particle surface and the matrix, and the absence of Fe-0-Si bonds in our materials.This is in contrast with the conclusions of Jung3* for silane-coated spinel iron oxide particles. Conclusion In this study, we showed that the polymerization of silica precursors in an aqueous dispersion of iron oxide nanoparticles allows the formation of nanocomposites free from aggregation of particles.This evidently requires control of the particles dispersion in the starting aqueous sol. Note that the inter- actions between the surface of particles and the matrix are very weak and probably involve only the solvation layers. References 1 H. Gleiter, Nanostruct. Muter., 1995,6, 3. 2 Nanophase Materials. Synthesis, Properties, Applications, ed. G. C. Hadjipanayis and R. W. Siegel, NATO ASI Ser. E, Applied Sciences, Kluwer Academic Publishers, Dordrecht, 1993, vol. 260. 3 J. L. Dormann, D. Fiorani and E. Tronc, Adu. Chem. Phys., 1996, 98, in press. 4 E.Tronc, P. Prene, J. P, Jolivet, F. D'Orazio, F. Lucari, D. Fiorani, M. Godinho, R. Cherkaoui, M. Nogubs and J. L. Dormann, Hype$. Interact., 1995, 95, 129; J. L. Dormann, F. DOrazio, F. Lucari, E. Tronc, P. Prene, J, P. Jolivet, D. Fiorani, R. Cherkaoui and M. Nogues., Phys. Rev. B, 1996,53,14291. 5 M. N. Asuha, J. Muter. Sci. Lett., 1993, 12, 1705. 6 D. Niznansky, J. L. Rehspringer and M. Drillon, IEEE Trans. Mag., 1994, 30,821. 7 A. P. Philipse, M. P. B. van Bruggen and C. Pathmamanoharan, Langmuir, 1994,10,92. 8 T. Gacoin, F. Chaput and J, P. Boilot, J. Sol-Gel Sci. Technol., 1994,2,679. 9 C. Chaneac, E. Tronc and J. P. Jolivet, Nanostruct. Muter., 1995, 6,715. 10 J. P. Jolivet, J. M. Fruchart and R. Massart, Nouv. J. Chim., 1983, 7,325.11 J. P. Jolivet, E. Tronc and L. Vayssieres, in Nanophase Materials. Synthesis, Properties, Applications, ed, G. C. Hadjipanayis and R. W. Siegel, NATO ASI Ser. E, Applied Sciences, Kluwer Academic Publishers, Dordrecht, 1993,vol. 260, p. 45. 12 L. Vayssibres, E. Tronc and J. P. Jolivet, J. Colloid Interface Sci., submitted. 13 J. P. Jolivet and E. Tronc, J. Colloid Interface Sci., 1988, 125, 688. 14 C. Chantac, L. Vayssitres, E. Tronc and J. P. Jolivet, J. Colloid Interface Sci., submitted. 15 C. J. Brinker and G. W. Scherer, Sol-Gel Science, Academic Press, New York, 1990. 16 R. K. Iler, The Chemistry of Silica, Wiley, New York. 1979. 17 V. Belot, R,J. P. Corriu, D. Leclercq, P. H. Mutin and A. Vioux, J. Muter. Sci. Lett., 1990, 9, 1052.18 J. P. Jolivet, De la solution ci I'oxyde, Inter Editions, Paris, 1994. 19 Handbook of Chemistry and Physics, 64th edn., CRC Press Inc., Boca Raton, FL, 1983. 20 E. Tronc, J. P. Jolivet and J. Livage, Hyperfine Interact., 1990, 54, 737. 31 32 S. Schrader and G. Buttner, Z. Anorg. Allg. Chem., 1963,320,220. L. Walter-Levy and E. Quemeneur, C. R. Acad. Sci., 1963,6,3410. 21 G. Socrates, Infrared Characteristic Group Frequencies, 2nd edn., Wiley, Chichester, 1994. 33 D. Scarano, A. Zecchina, S. Bordiga, F. Geobaldo, G. Spoto, G. Petrini, G.Leofanti, M. Padovan and G.Tozzola, J. Chem. Soc., 22 A. Bertoluzza, C. Fagnano, M. A. Morelli, V. Gottardi and Faraday Trans. 1993,89,4123. 23 24 25 M. Guglielmi, J. Non-Cryst. Solids, 1982,48, 117. J. Y. Ying and J. B. Benziger, J. Am. Ceram. Soc., 1993,76,2571. M. Prassas, J. Phalippou, L. L. Hench and J. Zarzycki, J. Non-Cryst. Solids, 1982,48, 79. M. L. Hair, J. Non-Cryst. Solids, 1975,19,299. 34 35 36 R. Szostak, V. Nair and T. L. Thomas, J. Chem. Soc., Faraday Trans. 1,1987,83,487. I. Dezsi and J. M. D. Coey, Phys. Status Solidi A, 1973,15, 681. F. Vratny, M. Dilling, F. Gugliotta and C. N. R. Rao, J. Sci. Znd. B, 1961,20, 559. 26 J. H. Anderson, Jr. and K. A. Wickersheim, Surf. Sci., 1964,2,252. 37 E. Tronc and J. P. Jolivet, Hyperfine Interact., 1986,28, 525. 27 28 29 C. Morterra and M. Low, Ann. N. Y. Acad. Sci., 1973,220, 133. F. Orgaz and H. Rawson, J. Non-Cryst. Solids, 1986,82, 57. C. C. Perry and X. Li, J. Chem. Soc., Faraday Trans., 1991, 87, 38 C. W. Jung and P. Jacobs, Magn. Reson. Imaging, 1995, 13, 661; C. W. Jung, Magn. Reson. Imaging, 1995,13,675. 30 761; 3857. D. L. Wood and E. M. Rabinovich, J. Non-Cryst. Solids, 1986, Paper 6/04363A; Received 24th June, 1996 82, 171. J. Muter. Chern., 1996, 6(12), 1905-1911 1911

ISSN:0959-9428    

DOI:10.1039/JM9960601905

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Synthesis, reactivity, structure and electronic properties of [N(CH,),] CGo1.5thf: fullerides with simple hexagonal packing Richard E. Douthwaite,? Mark A. Green, Malcolm L. H. Green and Matthew J. Rosseinsky" Inorganic Chemistry Laboratory, University of Oxford, South Parks Road, Oxford, UK OX1 3QR The preparation of [N(CH3),] C60 1.5thf and characterisation of its electronic properties, using susceptibility, EPR and 13C MAS NMR measurements, is reported. X-Ray powder diffraction shows a simple hexagonal stacking sequence of close-packed c60 layers, with cations and solvent molecules located in the trigonal-prismatic sites. The electronic properties are consistent with narrow-band metallic behaviour at high temperature. A transition in susceptibility and EPR measurements is seen between 200 and 230 K.A similar synthesis affords a c602-compound with the same simple hexagonal fulleride packing. c60 is an electronegative molecule. One of the most intensely studied aspects of its chemistry is the preparation of com-pounds containing C60"- fulleride anions.' If the counter ions are electropositive metals from Groups I or 11, they are small enough to occupy the interstitial sites in close-packed arrays of the fulleride. The resulting salts then have sufficient c60-c60 overlap between the frontier t,, orbitals to allow delocalisation of the electrons, yielding metallic conductivity, and supercon- ductivity in some compo~nds.~?~ The A3C,o fullerides have the highest superconducting transition temperatures (T,s) of any systems except the cuprates.Of particular relevance to the search for high- T, molecular superconductors is the monotonic increase in T, with volume per C603- anion in this series., The highest T, observed to date is that of 40K for cS3c60, a metastable material produced by low-temperature ~ynthesis.~ In conventional solid-state synthesis, at the temperatures required to prevent c60 decomposition (<650-800 "C), the volatility of most metallic elements is low and cation mobility in the c60 lattice can be insufficient for reaction. The use of solution-chemistry routes to new fullerides can circumvent these problems, allowing access to a wide range of structural types and interesting physical properties, and opening up the possibility of extending the number of fullerides with molecular counter ions. On the other hand, cation solvation may pre- vent fulleride close packing. The known superconductors K3C6, and Rb3C60 have been synthesised from s~lution,~-~ as have several novel fulleride salts with organic counter cat- ions such as [ (C6H5 )4 3ZC60 [I10.35 ,lo [ (C6H5 )4 1ZC60 cc1 1," {N[P(C6H5)312 1c6012 and {N[P(C6H5 )312 2c6013 {NLP-(C6H5)3]2= PPN} and the ternary fulleride B~,CSC,~.'~ The elemental closed-shell counter cations discussed in the preceding paragraph have spherical symmetry and a restricted size range. The electronic properties of the fulleride are con- trolled by the inter-C,, separation and the relative orientations of the c60 molecules.Molecular, as opposed to elemental, cations offer a wider distribution of sizes. In addition, the molecular shape can influence the fulleride packing and orien- tation. A first step in the use of non-elemental cation size and shape to control the electronic properties of fulleride salts was the formation of complex counter cations by ligand co-ordination to the intercalated metals. N~,CSC,,'~ affords ( NH,),N~,CSC,~ containing the tetrahedral (NH,),Na+ species on the octahedral site, body-centred tetragonal ( NH3)K3Cso is prepared from K3C6,16 while face-centred cubic (fcc) Na3C,0 yields body-centred cubic (bcc) ( NH3),Na3C6017 with linear (NH,),Na+ groups. Salts of large molecular counter and [P(c&f5)4]2C6010.3,~~and ( PPN)C,O'~ and ( PPN),C,O~~ all have crystal structures with large interfulleride separations.TDAECgo [where TDAE is tetrakis(dimethylaminoethy1ene)l has 10A c6o**.c60 contacts in one dimension'* and is a molecular ferromagnet with T,= 16 K.,' In this paper, we report the solution synthesis, reactivity, magnetic and electronic properties of [N(CH,),] C,, 1.5thf. Tetraalkylammonium c60 compounds have been studied at electrode surfaces20-22 but no bulk preparations or characteris- ation have been reported. The route adopted here is amenable to the preparation of molecular fullerides of higher charge, and the preparation and structure of the dianion salt [N(CH~)&C,O *xCH3CN is also reported. Experimental Materials Pure c60 powder was obtained by spark erosion,23 chromatog- raphy on alumina and sublimation under a dynamic vacuum of Torr at 500°C.NaC,, -5thf was prepared as described previously.24 N(CH,),F was prepared by a modified literature method.25 N(CH,),F -4H20 (Aldrich) was recrystallised three times from isopropyl alcohol [refluxed and distilled from CaH, (twice)] and the resulting alcoholate decomposed at 80 "C under dynamic vacuum (lo-' Torr). Purity was checked by 'H NMR spectroscopy and elemental analysis. Solvents were distilled under dinitrogen from sodium potas- sium alloy (thf) and CaH, (CH,CN). CD3CN was distilled from CaH, (twice), degassed using freeze-thaw cycles, and stored under argon. Synthesis All samples were manipulated in a helium-filled MBraun Labmaster drybox, where a 25 W light bulb with a 1 cm hole would typically burn for 5 days.Reactions were performed under argon using standard Schlenk techniques with dual argon and vacuum manifolds. [N(CH3),]C6o 1.5thf was prepared via salt metathesis of NaC,O Sthf with N(CH,),F in acetonitrile at -30 "C. NaC,, -Sthf is an oxygen- and water-sensitive compound in the solid state and in solution. It is appreciably soluble in thf, CH3CN and decomposes in EtOH and slowly in CH2C12. The tetramethylammonium halides (Cl, Br and I) are insoluble in thf and CH3CN at room temperature but are soluble in cations do not show fulleride close packing: direct c6O*'*c60 solvents such as alcohols and water. N(CH,),F is insoluble contact is lost as both cation and anion take part in the in thf, reacts with CH3CN at room temperature and is soluble packing to form the solid.The double salts [P(C~H~),],C~OC~'~ However, reaction with in alcohols without decomp~sition.~~ J. Muter. Chern., 1996, 6(12), 1913-1920 1913 CH,CN is slowed at lower temperatures and is negligible below -30 "C. This allows the metathesis reaction with NaC6,*5thf to be carried out in CH3CN at low temperature. At -30 "C a CH,CN solution (20 ml) of N(CH3),F (6.4 mg, 0.07 mmol) was added rapidly via cannula to a CH3CN solution (40 ml) of NaC,, -5thf (75.0 mg, 0.07 mmol). The volatiles were removed under reduced pressure at -30 "C and the black residue extracted with thf (2 x 20 ml) at room tem- perature giving an intense cherry red solution and leaving a pale grey residue.The red thf solution was concentrated to ca. 2 ml and cooled to -80 "C depositing black needle-like crystals in the supernatant. Filtration at -78 "C and washing with pentane (2 x 5 ml) gave black fern-like microcrystals of [N(CH,),] C6, 1.5thf in 80-90% yield. Filtration at -78 "C and drying under reduced pressure or under a stream of Ar produced similar results. The black microcrystalline solid is considerably less oxygen- and moisture-sensitive than NaC,, -5thf and can be handled in air for a few seconds without significant decomposition. It is soluble in CH3CN and is extremely soluble in thf (>25 mg ml-'), in contrast to NaC6,.5thf which has a con- siderably reduced solubility in thf and CH,CN (<5 mg ml-').Elemental analysis corresponded to the composition [N(CH,),]C,,. 1.5thf: found (calc.): C, 91.4 (93.1); H, 2.6 (2.7); N, 1.5 (1.6); Na, <0.1% (0). Integration of the 'H NMR (CD,CN) resonances of thf, N(CH3),+ and an internal calibr- ant of ferrocene of several samples from separate preparations also gave a stoichiometry of [N(CH,),]C,,-( 1.5 +O.l)thf. NIR spectroscopy in CH,CN showed a band at 1080nm which is in accordance with the c60 species in solution being a radical monoanion.26 The alkali-metal fullerides Na,C,,, K3C60 and K4C60 are all soluble in CH,CN and can also be reacted with N(CH,),F at -30 "C and then extracted at room temperature with CH,CN (this is necessary as none of these compounds are appre- ciably soluble in thf) to form [N(CH3),],C6, .xCH,CN, [N(CH,),],C,, .xCH,CN and [N(CH,),]&, -xCH3CN, respectively.Powder X-ray diffraction Samples were contained in 0.5 mm diameter capillaries sealed under helium. Data was collected in transmission geometry on a Siemens D5000 instrument with a 6 degree linear position- sensitive detector and monochromatic Cu-Kol, radiation from a Ge( 11 1)incident beam monochromator. Solid-state NMR spectroscopy Samples for I3C magic angle spinning (MAS) NMR spec- troscopy (ca. 50 mg) were contained in a KEL-F insert within a 7 mm zirconia rotor and spun typically at ca. 3 kHz. Samples for static 13C NMR spectroscopy (ca. 50 mg) were flame-sealed in 5 mm internal diameter Pyrex tubes.I3C MAS NMR and static spectra were recorded at 50.32 MHz on a Bruker MSL200 spectrometer. Temperature control was achieved using a Bruker VT-1000 temperature control unit. Spectra were referenced externally to adamantane, relative to tetra- methylsilane. Solution NMR spectroscopy Spectra were recorded on a Bruker AM-300 spectrometer at 300.13 MHz ('H) referenced internally to the residual protio- solvent resonance of CD3CN relative to tetramethylsilane. Magnetic measurements Dc static susceptibility measurements were performed using a Cryogenic Consultants S600C SQUID magnetometer equipped with a Lakeshore temperature controller. Samples were contained in gelatin capsules with a sample mass of 1914 J.Muter. Chem., 1996, 6(12), 1913-1920 typically 50-80 mg. Correction for the diamagnetism of the container was achieved by measuring the moment of the empty gelatin capsule over a series of temperatures and external fields prior to filling the capsule with sample. Diamagnetic contri- butions from the core electrons of thf and N(CH3),+ were corrected for using Pascal's constants and those of c60 using a correction of -258 x lod6emu mol-1.27 EPR spectroscopy Spectra were recorded on a Varian E-LINE Century series X-band spectrometer. Temperature control was achieved using an Oxford Instruments ITC 4 controller. Samples were con- tained in quartz tubes (Spectrosil) sealed with a Teflon stop- cock. Spectra were referenced externally to a powdered sample of a,a'-diphenyl-p-picrylhydrazyl (dpph) radicals and the line- width taken as peak to peak of the first derivative of the resonance.Conductivity The conductivity as a function of temperature from 120 to 300K was measured using a two-probe dc technique on a section of a pellet (2 x 5 x 0.5 mm3) pressed at room tempera- ture (15 tons) and the four contacts made with silver paint. The sample was mounted on an eight-pin chip and onto a probe on a copper block. The probe was covered with a copper cover, sealed with indium wire and removed from the drybox and placed directly into the cryostat. Temperature control was achieved using an ITC 4 Oxford Instruments intelligent tem- perature-control unit. Two-probe resistances were measured before and after each experiment in the drybox and confirmed that the sample was not decomposing during the measurement.Measurements were performed on cooling and warming. No significant hysteresis was observed. Results Powder X-ray diffraction The powder X-ray diffraction pattern of [N(CH,),] c60-1.5thf is shown in Fig. 1. The cell was determined to be hexagonal and refinement of observed reflFctions gave the c$l dimensions as a= 10.131(4), c= 10.23(1)A and 7/=909.8 A3. These cell parameters correspond to the anions being in van der Waals contact in all three dimensions; the structural motif suggested by the cell size is an AAA stacking of close-packed layers of c60 molecules, as previously reported for (P4)2C,,28,29 and (12)2C60.30In this simple hexagonal structure, there are only two suitable positions [shown in Fig.2(u)] for species of the size of thf or N(CH,),+ in the unit cell: the two trigonal prismatic sites at (+ 1/3, & 2/3, 1/2) [these sites are fully Fig. 1 Powder X-ray diffraction pattern of [N(CH,),]C,, 1.5thf 860 Fig. 2 (a)A simple hexagonal array of c60 molecules, with the trigonal-prismatic and square-planar sites shown. (b)Calculated total energy for various structural arrangements of “(CH,),] and thf. (c)Minimised structure of “(CH,),] C,, ethf. (d)Minimised structure of [N(CH,)4]2C6,. occupied in (P4)&0], or the square sites in the ac or bc faces at (1/2, 0, 1/2), (0, 1/2, 1/2) or (1/2,1/2,1/2). Molecular mech- anics computations, using the consistent-valence forcefield (CVFF) within the BIOSYM were employed to calculate and minimise the possible locations of the N(CH,),+ and thf molecules.All c60 molecules were assumed to be orientationally ordered in the P3 space group. Introduction of the tetramethylammonium cation into the simple hexagonal array clearly favours the trigonal-prismatic site over the square one. Diffraction patterns were calculated in which these sites are 50% occupied by the tetramethylammonium cation in a positionally disordered manner. However, this model clearly fails to reproduce the observed diffraction pattern [Fig. 3(u)] with the first two reflections calculated to be far too intense, suggesting the importance of modelling the scattering from the included solvent molecules which are required by the analytical and NMR data.Docking and energy-minimisation calculations in a simple hexagonal c60 array with half of the trigonal- prismatic sites occupied by the tetramethylammonium cations shows the thf molecules also prefer the trigonal-prismatic site over the square one. Two orientations of the thf, one with the molecular plane parallel and the other with this plane perpen- dicular to the layer stacking direction, are found to have very close total energies. The composition [N(CH,),] c60 thf [Fig. 2(c)] has a diffraction pattern [Fig. 3(b)] which depends only slightly on the orientation of the thf (shown as parallel). It is, therefore, impossible to determine this orientation with the present data.Calculations were then performed to explore the inclusion of a second thf molecule into the structure. It was found that two molecules can occupy the trigonal site, but with a considerable total energy increase over the single occupied arrangement, as shown in Fig. 2(b). Minimisation shows that these two molecules are both aligned parallel to the stacking direction, with a substantial energy increase if they are ordered perpendicular. Occupation of the square site is much higher in energy and thus ignored in the subsequent discussion. The possibility of solvent molecules occluded on the surface of the crystals makes the 1.5 thf molecules per formula unit an upper limit. Therefore, it is proposed that the observed composition results from predominantly single occu- pation of the trigonal sites by thf, with the possibility of a random distribution of doubly occupied sites.The inclusion of 50% extra thf produces a calculated diffraction pattern which is still in reasonable agreement with the observed pattern, shown in Fig. 3(c). [N(CH3),]& *xCH3CN is also hexagonal and refinement of eight deflections gives Fell parameters a =10.226( 2), c = 9.978(7) A and I/= 903.6 A3. The diffraction pattern can be approximately modelled by full occupancy of the trigonal prismatic sites by the tetramethylammonium cation [Fig. 2(d)]. The observed and calculated powder X-ray diffraction patterns of [N(CH3)4]& are shown in Fig. 4.Rietveld refinement using the program PROFIL33 with spherical shell scattering from the c60 and the N(CH3), cations gives R,= 18%, suggest- ing the model requires further modification for satisfactory agreement with the data. 13Csolid-stateNMR spectroscopy 13C NMR spectroscopy has been used extensively to probe the dynamics and relaxation mechanisms of alkali-metal fuller- ides and is an integral technique in the determination of the phase purity of a material containing c60.The room-tempera- ture 13C(1H) MAS NMR spectrum of [N(CH3),]C6, 1.5thf at 50.32 MHz shows only a single peak at 6 185 (v~,~<100 Hz) (Fig.5) which is assigned to the c60 monoanion. No 13C J. Muter. Chem., 1996, 6(12), 1913-1920 1915 I 10 15 20 25 30 35 40 45 I 10 15 20 25 30 35 40 45 ,hII I I II I IIII 1111 IIIIIIIMI I 5 10 15 20 25 30 35 40 45 2Bldegrees Fig.3 Powder X-ray diffraction patterns generated using BIOSYM software from the structures in Fig. 2 for: (a) [N(CH~)~]C~O;(b)[N(CH3),]C6,.thf; (c) [N(CH3)4]C60*1.5thf resonances could be observed from the thf or N(CH3),+ moieties (relaxation delays from 0.5 to 60 s were used). 13C{ 'H} CP MAS NMR with contact times ranging from 0.5 to 50 ms did not show any resonances except the single peak at 6 185 ppm. This could be due to positional disorder of thf and N(CH3), plus orientational disorder of the fulleride resulting + in the occupation of a large number of slightly differing sites and a very broad resonance, or broadening of the resonance by local moments on adjacent fulleride anions.Exposure of the solid to air resulted in decomposition over a period of minutes which was monitored by 13C{lH) MAS NMR and showed new peaks at 6 143 and 55 assigned to c60 powder and the N( CH3)4 cation, respectively. Peaks corresponding + to thf were not observed. This is probably due to evaporation of thf from the decomposing sample. Decomposition is accompanied by a colour change from black to tan brown. The room-temperature 13C{ 'H} static NMR spectrum of [N(CH,),]C,,. 1.5thf at 50.32 MHz showed a single peak at 6 185 (V~,~=~OO Hz) (Fig. 6). This indicates that the c60 monoanions are undergoing isotropic reorientation which is fast (> lo4 s-l) on the NMR timescale and thus averages out the chemical shift anisotropy.Cooling the sample below 200 K (Fig. 6) results in broadening and a large change in the chemical shift to lower field. The broadening is either due to restriction of the reorientation of the c60'-anions or to 1916 J. Muter. Chern., 1996, 6(12), 1913-1920 4000-h2?-53000-$ v 2000-.-c v)t al+-C--1000-I I I I II I I IIII Ill1 IIIII 111111 I 5 10 15 20 25 30 35 40 45 2 8 /d eg rees Fig.4 (a) Observed powder X-ray diffraction pattern of [N(CH,),],C,,; (b) X-ray powder diffraction pattern calculated according to the model of Fig. 2(d) 1,1.1.I.I.I 240 220 200 180 160 I40 6 Fig.5 Room-temperature 13C MAS NMR spectrum of N(CH,),C,, 1.5thf enhanced relaxation produced by local magnetic moments.The isotropic chemical shift (defined as the centre of gravity of the resonance) changes abruptly below 200 K, in agreement with the magnetisation and EPR measurements, but its tem- perature dependence cannot be fitted to a Curie-Weiss law at any temperature. The temperature independence of the shift at 230K and above is consistent with metallic behaviour in this temperature range. The I3C MAS NMR spectra of the [N(CH3),],C60 fullerides produced by reaction of N(CH,),F with the C60n- anions are shown in Fig. 7 (u)-(c). They show a common feature at 6 180, which we assign to [N(CH3),],C60. Magnetic measurements Dc magnetic susceptibility measurements of [N(CH3),] Cm 1.5thf as a function of temperature at 1T show non-Curie- Weiss behaviour over the entire temperature range and a large drop in the susceptibility at cu.200 K (Fig. 8). The suscepti- bility as a function of temperature after zero-field cooling 2S0K 22SK 2 350 300 250 200 150 100 50 6 Fig. 6 I3C wideline (static) NMR spectra of [N(CH,),]C,. 1.5thf at the temperatures indicated. Spectra were filled to 4096 points and 50 Hz Lorentzian line broadening applied (100Hz at 146 K). and field (1 T) cooling did not show any hysteresis below or above the transition at 200 K. The magnetisation as a func- tion of applied field at 6 K from 0 to 3 T was linear and did not show hysteresis or saturation. Attempts to fit the magnetisation data over the whole temperature range were unsuccessful.In itinerant electron systems, sharp drops in magnetisation are observed in the Kondo insulators, which are systems of considerable current interest.34 FeSi, now considered an early prototype of this class of material, has a qualitatively similar temperature dependent susceptibility to [N(CH,)4]C,O lSthf, but the narrow band 1 I I 1 1 1 200 I50 I00 50 6 Fig.7 13C MAS NMR spectra at room temperature of (a) [N(CH,),]&,; (b) the solid product of the reaction of K3C60 and N(CH,),F in CH,CN; (c) K4C60 and N(CH,),F in CH,CN 'i 0.0054 11.4 t-0 m 1.2 3rIE,O.O04 = E< 0.003: =t"0*..* equations used to fit its susceptibility fail quantitatively here.35 As I3C NMR measurements indicate a significant change in the molecular dynamics in the region of the magnetic transition, and the 13C NMR total shift becomes temperature dependent below this transition, we then analysed the data above and below the transition separately.A convenient starting point for understanding the data is to consider the effective magnetic moment per c60 [p,ff/pg =2.82(~,~~T)~'~],which is shown in Fig. 8. A localised array of non-interacting S = 1/2 c60-anions would have a temperature-independent value of 1.73 pB. Although the high-temperature limiting value is indeed 1.7 pB, there is no temperature region over which the Curie law is obeyed. At low temperature, below the 200 K transition, fits to x =C/T+ A +f(T),where f(T)is a slowly varying function of temperature (discussed in more detail with respect to the high-temperature data) and C/T and A represent respectively Curie and temperature-independent contributions to the sus-ceptibility, are successful [Fig.9 (a)].In all the fits, regardless of the precise form of f(T),C=O.O35-0.04 emu mol-' K, Fig. 8 Dc static susceptibility (open diamonds) measured in a 1 T field and effective magnetic moment (filled circles) as a function of tempera-ture for "(CH,),] c60* 1.5thf corresponding to a magnetic moment of 0.6 pugper c60, and A=(7-8) x emu mol-l. The data between 230 and 300 K cannot be fitted successfully with the function x =A + C/T;this functional form has incor- rect curvature. The Curie-Weiss law yields a significantly poorer fit than found at low temperature, with unphysically large values of 8 (-450 K) and peff (2.5 pB per c60).The second-order polynomial, x =u +bT+ cT2, provides a much better fit, with the size of the linear and second-order contri- butions being comparable to the constant term over this temperature range. This behaviour is appropriate for a metal with a small Fermi energy. Fits to the functional form appro- priate to the paramagnon-enhanced correlated x = xPaull[1-(T/ATF)2] also provide a considerable improvement over the Curie law fits. Including the low-T Curie tail as a J. Muter. Chem., 1996, 6(12), 1913-1920 1917 00 0 20 40 60 80 100 120 140 160 Fig. 9 (a) Fit to susceptibility below 160 K as described in the text (b) Fit to susceptibility of [N(CH,),]C,,* 1 Sthf above 260 K as described in the text fixed term makes little difference to the fit The fit shown in Fig 9(b) 1s to x=~o[l-(T/B)~]+O04/T with x0=16x lop4 emu mol-' and B=661 K Electron spin susceptibillty measurements were performed by double integration of the EPR signal between 130 and 290 K and showed a transition similar to that observed in the dc susceptibility data [Fig lO(u)] The linewidth [Fig 10(b)] T(a) 3300 3200 T t! 3 3100C $ 3000 v $' 2900 2C 2800 r 2700 me 2600 1 I I I I I 100 150 200 250 300 : (b) 30-r Q 5 20-1 15--0 lo -=I I I Fig.10 (a) Spin susceptibility of "(CH,),] C6, 1 Sthf derived from integration of EPR spectra as a function of temperature (b) EPR peak-to-peak linewidth as a function of temperature for "(CH,),] C,, 1 Sthf 1918 J Muter Chem, 1996, 6(12), 1913-1920 above 230 K shows an increase with increasing temperature, as has been noted previously in the literature The g value remained constant within expenmental error at 2 000 over the temperature range studied A benzonitrile glass of [N(CH,),] and the linewidth increased with temperature with a constant g value of 1999, consistent with other reports of frozen glass solutions of c60-22 24 26 Conductivity The room-temperature conductivity of a pressed pellet is 0 01 R-'cm-l, and the temperature dependence appears character-istic of a semiconductor with an activation energy of approxi-mately 0 36 eV, a distinct transition is observable below 200 K, which may be associated with the transition seen in the magnetism Significant grain-boundary effects are to be expected in such an unsintered pellet, but the key point for the purposes of the present paper is that the conductivity is significantly higher than in other c60-salts with isolated fulleride anions Some reactivity Exposure of solid [N(CH,),] C,o -1 Sthf to air causes oxidation to c60, observed by I3C MAS NMR spectroscopy On exposure to air, a red thf solution of [N(CH,),]C6,.1 Sthf deposits a tan-brown solid (c60 powder) within 30s However, a red CH,CN solution becomes a very deep orange-brown on exposure to air Removal of the volatiles under reduced pressure gives a dark brown-black reflective solid, 1, which is insoluble in all common organic solvents including CH,CN 13C MAS NMR spectroscopy of 1 showed a sharp peak at 6 55, assigned to N(CH3),+, and a broad peak centred at 6 143, assigned to a C60-baSed material (c60itself gives a very sharp resonance at this chemical shift value when pure) Solution 'H NMR spectroscopy in CD,CN of the orange-brown solution formed by exposure of [N(CH3),]C,O*l 5thf to air showed one broad peak at 6 23 in addition to the residual protio solvent (Fig 11) Powder X-ray diffraction showed that 1 was amorphous The fact that 1 is formed in CH,CN and not thf strongly suggests that CH,CN is involved in a reaction with c60 -in the presence of air Compound 1 is then possibly the result of polymerisation/oligomerisation of (or initiated by) C,,-based radicals on removal of the volatiles from the orange-brown CH,CN solution Species 1does not have an observable EPR spectrum at room temperature and it is essentially diamagnetic as determined by SQUID magnetometry Reaction of 1 equiv of [N(CH,),]C,,.l Sthf with [Fe(q-C5H5)(C0)2C1] in thf gave [Fe(q-C,H,)(CO),], c60 and 5.0 4.5 4.0 3.5 3 0 2.5 2.0 1.5 1.0 0.5 0.0 6 Fig.11 Solution 'H NMR spectrum of [N(CH3),]c6,.1 Sthf in CD,CN after exposure to air N(CH,),Cl. This is due to reduction of [Fe(q-C,H,)(CO),Cl] by c60'-. Suspension overnight of [N(CH,),] C6, -1.5thf with stirring in a solution of 1.5 mol dm-, n-butyllithium produced intercal- ation of lithium cations and presumably reduction of the [N(CH,),]C,, 1.5thf host giving a black solid, 2.After fil- tration and washing with hexane, 7Li MAS NMR spectroscopy showed a broad peak at 6 3 (vl,, =3 kHz) and 13C MAS NMR studies revealed two peaks at 6 183 and 181 (v~,~~2200Hz for both). Compound 2 is currently under investigation. Several attempts at intercalation of rubidium metal at temperatures from 100 to 300°C caused decomposition of [N(CH,),]C60*1.5thf as observed by 13C MAS NMR spec- troscopy. No superconductivity was observed in these solids, which had broad MAS NMR resonances at 6 145 and 186, tentatively consistent with disproportionation of the monoanion salt into c60 and a more reduced fulleride on reaction with rubidium.Solution 'H NMR spectroscopy shows that thf can be removed from [N(CH,),] C60 1.5thf under dynamic vacuum (lo-, Torr) at 80 "C, but 13C MAS NMR spectroscopy of the desolvated material always showed a broad peak at 6 143 in addition to a peak at 6 183, indicating that removal of thf is associated with decomposition. Powder X-ray diffraction of the desolvated material showed broad Bragg peaks that pre- cluded confident indexing. [N(CH,),]C,, -1.5thf is prone to loss of crystallinity on prolonged standing in the drybox, which we associate with thf loss. Discussion We have isolated simple hexagonal arrays containing close- packed layers of c60 anions, with tetramethylammonium as the counter cation. Energy minimisation and powder X-ray diffraction suggest location of the TMA' (tetramethylam-monium) cations at the trigonal-prismatic sites, although the details of the structures proved inaccessible from the powder data.The lattice parameters show that close packing is still maintained in two dimensions; it is simply the stacking sequence of the layers wbich changes from ABC to AAA, reducing the number of 10 A interfulleride contacts from twelve to eight. This is an interesting crystal chemical observation. We may roughly estima!e the van der Waals radius of the N(CH3),+ Fation as 2.2 A. The (NH,),Na+ cation is approxi- mately 3.2 A in size and still occupies the octahedral site in fcc (NH,),N~,CSC~~,and therefore an fcc TMAC6, with TMA+ on the octahedral site is not precluded on size grounds.It appears that the absence of small cations suitable to occupy the tetrahedral sites in the fcc structure and the presence of a second large group (the thf solvent or a second tetramethylam- monium cation) under the present low-temperature (-30 "C) synthesis conditions favours this simple hexagonal structure. The difference from the denser, close packed structures is that there are two large sites per c60 (the trigonal-prismatic ones), rather than the single octahedral site available in fcc structures, and so the presence of two large groups clearly favours simple hexagonal packing. The second moiety [thf or N(CH,),+] is apparently required to stabilise the hexagonal structure, to avoid vacancies on the large trigonal prismatic sites; this is demonstrated by the instability of the x= 1 material when desolva ted.The C60-c60 distances in (PPN)xC60 (x= 1, 2) and [P(S&)4]2C60 (X =Br, I) are considerably greater than 10 A, accounting for their strongly insulating character. [N(CH,),]C,, * 1.5thf is more conducting than these salts and has a transition in its magnetic susceptibility. The absence of any signature of the transition in the frozen-glass EPR spectra indicates that the transition is a solid-state effect, arising from the close interfulleride contacts. No hysteresis is observed and I3C wideline NMR studies do not show any signs of large line broadenings or of extra resonances. It therefore seems unlikely that the transition in [N(CH,),] C6, -1.5thf is due to cycload- dition producing poly- or oligo-merisation/dimerisation, which is found for the AIC60 systems (A=K, Rb, Cs).Transitions of the sort observed in [N(CH,),]C,, 1.5thf have been found in the compounds NaxC60.ythf (where x= 0.36-0.42 and y =1.5-2.9),,' NaC,, 3thf 38 and NaC,, 5thf.24 In the case of Na,C,,-ythf (where x=O.36-0.42 and y= 1.5-2.9) a metal-metal transition occurs at 180 K, with an increase in single-crystal conductivity from a virtually tempera- ture-independent value of 50 S cm-' to a maximum value of ca. 1000 S cm-' at 100 K. The observation of broadening of the 13C NMR spectra without hysteresis suggests that the transition in [N(CH,),]c60 -1.5thf is closely associated with a slowing down of the reorientation dynamics of the c60-species.Given the close interfulleride contacts in the simple hexagonal struc- ture, any structural transition will strongly influence the physi- cal properties by changing the nearest-neighbour transfer integral. The localised electron analysis of x above 230K produces unphysical parameters. An itinerant electron analysis is consist- ent with the high powder conductivity compared with (PPN),C,, and [P(C&,)&C&1. The temperature-indepen- dent terms in the fit above 230 K are close to xPaullfound for many fulleride metals [x(K3C6,)= 10 x lo-, emu mol-l]. The tight-binding bandwidth, W=2zt, may be estimated as 0.4 eV, using z= 8 and t=0.025 eV (from K3C6,).39 This is compared with estimates of the Hubbard U of 1-1.5eV,40 suggesting that it is appropriate to use the paramagnon model for spin fluctuations produced by interelectron repulsion in narrow-band metals to analyse the data.Here, the density of states at the Fermi energy, E,, evaluated from the magnetic response, N,, (x=2pB2N,, where N, is per spin, i.e. orbital states) is enhanced over that predicted from the bare density of states, N,, as N,=N,/(l --INB)where I is the Stoner parameter. In addition, excitation of paramagnons due to the incipient local moment character (spin fluctuations) gives the susceptibility a temperature dependence in narrow band systems. A quantitat- ive expression for this is x =xPaull[1-3.2.n2/24K2(T/TF)2], where K= l-IN, [see the fit in Fig.9(b)]. Here x= 16 x lop4 emu mol-I yields N,=20 states eV-' spin-' c60-I. We then use a free-electron estimate of N, of 8 states eV-' C60-1 spin-' to find a Stoner enhancement of the susceptibility of the order of 2.5 (similar to that found for K3C6,,') and an unenhanced Fermi temperature of 1865 K. The free-electron Fermi tempera- ture with one electron per c60 in a triply degenerate band is 1800K and so the magnitude and temperature dependence of the susceptibility are consistent with the correlated itinerant electron model. Within this interpretation of the high temperature data, the transition at 230 K would be due to significant changes in the transfer integrals being associated with the change in molecular dynamics indicated by the variable-temperature NMR measurements. The transition would then be similar to the metal-metal transition seen in Na,C6, ythf (x=0.36-0.42, y = 1.5-2.9).However, in order to account for the substantial Curie tail at low temperature, either 10% S =1/2 impurities (extraneous or trapped at c60 sites in the otherwise metallic solid) must be present in the sample, or the electronic structure of the solid is such that each c60 carries a local moment of 0.6 pugin the metallic state. An alternative interpretation of the transition is therefore as a metal-insulator transition where the delocalised electrons become 0.6 p, local moments on the c60 anions and the constant term A, derived from the analysis of the low-temperature data, is ascribed to van Vleck tempera- ture-independent paramagnetism.This description of the low- temperature state is consistent with either localised or delocal-ised models for the high-temperature behaviour and would agree with the temperature dependence of the 13C shift below J. Mater. Chem., 1996, 6(12), 1913-1920 1919 the transition The small local moment can then be ascribed to use of the simple Curie model in a situation where assign- ment of electronic character as purely localised is oversimplified The small variation of the EPR linewidth with temperature at high temperature is also suggestive of metallic behaviour above the transition Quantitative interpretation of the hne- width may be made following Janossy et a1 42 and results in a conduction electron scattering time of z,=4 x s, four times smaller than that found for Rb&0 The physically sensible values derived from analysis of the high-temperature phase EPR and x data with the assumption of a narrow-band metal increase our confidence in this assignment Below 200 K, the temperature dependence of the linewidth is more marked, suggestive of either a longer, more temperature-dependent zR in the more highly conducting metallic phase, or a transition to a localised c60-system [N(cH,)~]C60*1 5thf is an example of a c60monoanion salt in which the single t,, electron does not promote the formation of oligomeric or polymeric units, despite the close interfulleride contacts The magnetic data show that strong interactions between the t,, electrons make the magnetic response considerably mare complex than would be expected from a localised array of c60-anions a coherent interpretation of the EPR, susceptibility and conductivity data is possible within an itinerant electron model above the transition, but a decisive distinction between strongly coupled local spins and a narrow-band metal is difficult without further measurements, preferably on single crystals, which will help resolve the precise nature of the transition at 230K The structural chemistry provides a contrast with that of the larger alkali-metal cations and the fullerides of complex ammoniated cations in that the close-packed fulleride layers are stacked in a simple hexagonal AAA manner, this seems to require the presence of two large guest species (cation or solvent molecule), which occupy the two trigonal-prismatic sites in this array We thank Dr Mohammed Kurmoo of the Royal Institution for assistance with the conductivity measurement, St John’s College, Oxford for a Junior Research Fellowship to R E D and the EPSRC for a grant towards the purchase of the powder diffractometer References 1 D W Murphy, M J Rosseinsky, R M Fleming, R Tycko, A P Ramirez, R C Haddon, T Siegrist, G Dabbagh, J C Tully and R E Walstedt, J Phys Chem Solids, 1992,53, 1321 2 A F Hebard, M J Rosseinsky, R C Haddon, D W Murphy, S H Glarum, T T M Palstra, A P Ramirez and A R Kortan, Nature (London), 1991,350,600 3 R C Haddon, A F Hebard, M J Rosseinsky, D W Murphy, S J Duclos, K B Lyons, B Miller, J M Rosamilia, R M Fleming, A R Kortan, S H Glarum, A V Makhija, A J Muller, R H Eick, S M Zahurak, R Tycko, G Dabbagh and F A Thiel, Nature (London), 1991,350,320 4 R M Fleming, M J Rosseinsky, A P Ramirez, D W Murphy, J C Tully, R C Haddon, T Siegnst, R Tycko, S H Glarum, P Marsh, G Dabbagh, S M Zahurak, A V Makhija and C Hampton, Nature (London), 1991,352,701 5 T T M Palstra, 0 Zhou, Y Iwasa, P E Sulewski, R M Fleming and B R Zegarski, Solid State Commun , 1994,92,71 6 R P Ziebarth,D R Buffinger,V A Stenger,C RecchiaandC H Pennington, J Am Chem SOC ,1993,115,9267 7 H H Wang, A M Kim, B M Savall, K D Carlson, J M Williams, M W Lathrop, K R Lykke, D H Parker, P Wurz, M J Pellin, D M Gruen, U Welp, W-K Kwok, S Fleshier, G W Crabtree, J E Schirber and D L Overmeyer, Inorg Chem , 1991,30,2962 8 H H Wang, A M Kini, B M Savall, K D Carlson, J M Williams, K R Lykke, D H Parker, P Wurz, M J Pelhn, D M 1920 J Mater Chem, 1996, 6(12), 1913-1920 Gruen, U Welp, W -K Kwok, S Fleshier and G W Crabtree, Inorg Chem, 1991,30,2838 9 X Liu, W C Wan, S M Owens and W E Broderick, J Am Chem SOC, 1994,116,5489 10 A Penicaud, A Perez-Benitez, R Gleason, E Munoz and R Escudero, J Am Chem SOC , 1993,115,10392 11 U Bilow and M Jansen, J Chem SOC Chem Commun , 1994,403 12 H Kobayashi, H Moriyama, A Kobayashi and T Watanabe, J Am Chem SOC, 1993,115,1185 13 P Paul, Z Xie, R Bau, P P W Boyd and C A Reed, J Am Chem SOC, 1994,116,4145 14 A C Duggan, J M Fox, S J Heyes, P F Henry, D Laurie and M J Rosseinsky, Chem Commun , 1996,1191 15 0 Zhou, R M Fleming, D W Murphy, M J Rosseinsky, A P Ramirez, R B van Dover and R C Haddon, Nature (London), 1993,362,433 16 M J Rosseinsky, D W Murphy, R M Fleming and 0 Zhou, Nature (London), 1993,362,433 17 P F Henry, M J Rosseinskyand C J Watt, J Chem SOC Chem Commun, 1995,2131 18 P W Stephens, D Cox, J W Lauher, L Mihaly, J B Wiley, P-M Allemand, A Hirsch, K Holczer, Q Li, J D Thompson and F Wudl, Nature (London), 1992,355,331 19 P-M Allemand, K C Khemani, A Koch, F Wudl, K Holczer, S Donovan, G Gruner and J D Thompson, Science, 1991, 253, 301 20 R G Compton, R A Spackman, R G Wellington, M L H Green and J Turner, J Electroanal Chem , 1992,327,337 21 R G Compton, R A Spackman, D J Riley, R G Wellington, J C Eklund, A C Fischer, M L H Green, R E Douthwaite, A H H Stephens and J F C Turner, J Electroanal Chem, 1993, 344,235 22 K M Kadish, D Dubois and M T Jones, J Am Chem SOC, 1992,114,6446 23 W Kratschmer, L D Lamb, K Fostiropoulos and D R Huffman, Nature (London), 1990,347,354 24 R E Douthwaite, A R Brough and M L H Green, J Chem SOC Chem Commun , 1994,267 25 K 0 Christie, W W Wilson, R D Wilson, R Bau and J-A Feng, J Am Chem SOC, 1990,112,7619 26 C A Reed, J Stinchcombe, A Penicaud, P Bhyrappa and P D W Boyd, J Am Chem SOC, 1993,115,5212 27 R C Haddon, L F Schneemeyer, J V Waszczak, S H Glarum, R Tycko, G Dabbagh, A R Kortan, A J Muller, A M Mujsce, M J Rosseinsky, S M Zahurak, A V Makhija, F A Thiel, K Raghavachari, E Cockayne and V Elser, Nature (London), 1991,350,46 28 R E Douthwaite, M L H Green, S J Heyes, M J Rosseinsky and J F C Turner, J Chem Soc Chem Commun, 1994,1367 29 I W Locke, A D Darwish, H W Kroto, K Prassides, R Taylor and D R M Walton, Chem Phys Lett, 1994,225,186 30 Q Zhu, D E Cox, J E Fischer, K Kniaz, A R McGhie and 0 Zhou, Nature (London), 1992,355,712 31 P Dauber-Osguthorpe, V A Roberts, D J Osguthorpe, J Wolff, M Genest and A T Hagler, Proteins Structure Function and Genetics, 1988,4, 31 32 InsightII User Guide, Biosym/MSI, 1995 33 J K Cockcroft, PROFIL Rietveld refinement program, 1991 34 G Aeppli and Z Fisk, Comments Condens Mater Phys, 1992, 16,155 35 V Jaccarino, G K Wertheim, J H Wernick, L R Walker and S Arajs, Phys Rev, 1967,160,476 36 N F Mott, Metal-Insulator Transitions, Taylor and Francis, 1990 37 H Kobayashi, H Tomita, H Moriyama, A Kobayashi and T Wanatabe, J Am Chem SOC, 1994,116,3153 38 R E Douthwaite, M A Green, M L H Green and M J Rosseinsky, unpublished results 39 S Satpathy, V P Antropov, 0 K Anderson, 0 Jepsen, 0 Gunnarsson and A I Leichtenstein, Phys Rev B, 1992, 46, 1773 40 R W Lof, M A van Veenendaal, B Koopmans, H T Jonkman and G A Sawatzky, Phys Rev Lett, 1992,68,3924 41 A P Ramirez, M J Rosseinsky, D W Murphy and R C Haddon, Phys Rev Lett, 1992,69, 1687 42 A Janossy, 0 Chauvet, S Pekker, J R Cooper and L Forro, Phys Rev Lett, 1993,71,1091 Paper 61059525, Received 28th August, 1996

ISSN:0959-9428    

DOI:10.1039/JM9960601913

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Preparation of BaTi,09 by a sol-gel method and its photocatalytic activity for water decomposition Mitsuru Kohno, Shuji Ogura and Yasunobu Inoue* Analysis Center & Department of Chemistry, Nagaoka University of Technology, Nagaoka, Niigata 940-21, Japan In the development of photocatalyst materials, barium tetratitanate, BaTi,O,, with a pentagonal-prism tunnel structure was prepared by a sol-gel method and calcined in air at various temperatures from 873 to 1273 K. The changes in the structures were investigated using TG-DTA, X-ray diffraction, Raman and FTIR spectroscopies, and the photoinduced properties were examined by EPR spectroscopy. X-Ray diffraction patterns showed that crystallization occurs by calcination above 973 K. A characteristic Raman peak at 860 cm-', assigned to the stretching vibration of short Ti-0 bonds, increased in intensity with increasing calcination temperatures.EPR signals with g= 2.018 and g=2.004 were produced for BaTi,O, calcined above 1173 K with UV irradiation at 77 K in the presence of oxygen. These signals are associated with an 0'-radical. Calcined BaTi,09 was combined with RuO,, and its photocatalytic activity for water decomposition increased with increasing calcination temperatures of BaTi,O,. From the findings that the formation of the 0'-species and the photocatalytic activity are the results of high efficiency for the separation of photoexcited charges and are closely associated with crystallization of BaTi,O,, it is proposed that the crystallization develops the pentagonal-prism tunnel structure of BaTi,O, and enhances the role of the polarization fields present in the TiO, octahedra which facilitates the separation of photoexcited charges.In view of the current interest in photocatalysis by metal oxides, the design and preparation of efficient photocatalyst materials have been among the most important subjects. Recently, we have found that barium tetratitanate, BaTi,09, having a unique pentagonal-prism tunnel structure, acts as an excellent photocatalyst for the decomposition of water' when the titanate is combined with RuO,. In the development of barium titanate-based photocatalysts applicable to various gaseous and liquid reactions, further research into the prep- aration of BaTi409 and the investigation of its photocatalytic properties are necessary.In the present work, we have employed a sol-gel method, since the method has the advan- tage of producing complex metal oxides with different morpho- logies such as thin films and ultrafine powders as a material. The sol-gel method has been applied for the preparation of Ti0,,2-5 BaTiO, ,6*7 Ba2Ti9020 and BaTi5Ol1 ,' but there have been few studies on BaTi,O,. In the present study, BaTi,O, powder was prepared from barium acetate and titanium isopropoxide. Since the photocat- alytic activities of metal oxides are closely associated with their structures, the prepared BaTi,O, was subjected to heat treat- ment at various temperatures and characterized by TG-DTA, X-ray diffraction, FTIR and Raman spectroscopies.For the evaluation of photocatalytic features, the efficient production of photoexcited charges is an important factor. Thus, the ability to produce surface radicals in the presence of oxygen with UV irradiation was measured by an EPR method. Furthermore, the prepared BaTi,O, was combined with RuO, and the photocatalytic activity of the resulting mixture for water decomposition was examined. On the basis of the obtained results, the relation between the structure of BaTi,09 and the photocatalytic activity is discussed. Experimental In the sol-gel method for the preparation of barium tetratita- nate, BaTi,O, (referred to as BTO), acetic acid anhydride solutions of barium acetate, Ba(CH,COO), (Extra pure grade, Nakarai Tesque Inc.), and titanium isopropoxide, Ti(OC3H7), (Extra pure grade, Nakarai Tesque Inc.), were mixed in a 1: 1 molar ratio in a nitrogen-purged glove box, and water was added dropwise at room temperature.The resulting viscous sol solution was dried at 473 K for 5 h and then calcined in air at various temperatures from 873 to 1273 K for 20h. Powder samples were obtained which are here referred to as BTO (calcination temp./K): for example, a powder obtained by calcination at 873 K is denoted as BTO( 873). TG-DTA curves were recorded on a Rigaku TG 8110 thermal analyser. The temperature was increased from 473 to 1473 K at a heating rate of 10 K min-'. Al,O, powder was used as a reference. The structures of the prepared samples were determined using a Rigaku RADII1 X-ray diffractometer.The degree of crystallinity was evaluated from the intensity of the diffraction peaks due to the (121) and (031) planes of BTO As a reference for the evaluation, a well sintered BTO sample, which was prepared from a mixture of BaCO, and TiO, by calcination at 1273 K, was employed. The surface area of BTO was measured via the BET method using nitrogen at 77 K. Raman spectra of BTO were measured at room temperature with a JASCO NR-1100 spectrometer in the wavelength region from 300 to 1000 cm-'. For measurements of FTIR spectra, BTO was mixed with a fine KBr crystal and then pressed into a disc. The FTIR spectra were obtained with a JASCO FTIR-500 spectrometer. For measurements of EPR signals, ca.300mg of sample was placed in an X-band quartz cell. After degassing at 573 K and then introduction of 30 Torr of oxygen at room temperature, the sample was cooled to 77 IS. The EPR signals were recorded on a JEOL JES-RE2X instru- ment with irradiation by a 500 W low-pressure mercury lamp. The g values were calibrated using Mn2+ in MgO. The photocatalysts were prepared by impregnation of BTO with an RuC13 aqueous solution at 353 K, which was then dried in air and then reduced in a hydrogen atmosphere at 673 K for 2 h, and oxidized in an oxygen atmosphere at 773 K for 2 h. The ruthenium loading was 1mass% (metal content). The powder Ru0,-BTO photocatalyst was placed in a quartz cell and filled with pure water.The decomposition of water was carried out in an Ar atmosphere of 40Torr using a gas circulation system under irradiation by an Xe lamp operated at 400 W. Hydrogen and oxygen evolved in the gas phase were analysed by a gas chromatograph. Details of the appar- atus and procedure for the photocatalytic reaction have been described elsewhere.' J. Muter. Chem., 1996, 6(12), 1921-1924 1921 I -1 92 90 473 673 873 1073 1273 1473 TIK Fig. 1 TG and DTA curves of BTO dried at 473 K Results Fig 1 shows the TG and DTA curves of a dried sol at 473 K In the TG curve, a ca 5% mass loss occurred in the temperature region 573-723 K, followed by a gradual mass loss at higher temperatures The total mass loss reached ca 7% at 1473 K In the DTA curve, a strong exothermic peak was observed at ca 630 K, which corresponded to the significant mass loss Two small exothermic peaks appeared at ca 980 and 1095 K without noticeable mass loss Fig 2 shows X-ray diffraction patterns of BTO treated at calcination temperatures between 873 and 1273 K Table 1 lists the degrees of crystallinity and BET surface areas of the BTO samples prepared BTO(873) was amorphous, as seen in a largely broadened X-ray diffraction peak With increasing calcination temperatures, crystallization proceeded and reached 83% at 1273 K The X-ray diffraction pattern of BTO( 1273) was exactly the same as that of BaTi,O, reported previously The specific surface area decreased with increasing calcination temperature it was 37 m2 g-' for BT0(873), 20 m2 g-' for BT0(973), 1 m2 g-' for both BTO(1073) and BTO( 1173), and c1 m2 g-' for BTO( 1273) Fig 3 shows the Raman spectra of the prepared samples In the spectrum of BT0(873), broad peaks appeared in the I frC a cc 20 30 40 50 60 70 80 90 28/degrees (Cu-Ka) Fig.2 X-Ray diffraction patterns of (a) BT0(873), (b) BT0(973), (c) BTO( 1073), (d) BTO( 1173) and (e) BTO( 1273) Table 1 Crystallinity and surface areas of the prepared BTO samples BTO calcination temperature/K crystallinity (%) surface area /m2 g-' BTO( 873) BTO(973) BTO( 1073) BTO( 1173) BTO( 1273) 873 973 1073 1173 1273 0 14 35 70 83 37 20 1 1 <1 1922 J Muter Chem, 1996, 6(12), 1921-1924 x10 1 I 1 10 800 600 400 wavenumbedcm-' Fig.3 Raman spectra of (a) BT0(873), (b) BT0(973), (c) BTO( 1073), (d) BTO(1173) and (e) BTO(1273) following four regions 430-450, 520, 590-650 and 860 cm-' Upon calcination at 973 K, the peak at 520 cm-' disappeared completely, whereas considerable increases in the intensities of the remaining peaks were observed Further calcination at 1173 K resulted in a remarkable growth of these three peaks The spectral patterns of BTO( 1173) and BTO( 1273) are nearly the same Fig 4 shows the FTIR spectra of BTO in the range 2700-3100 cm-' The spectrum of BTO(873) shows two peaks at 2850 and 2930 cm-' with a shoulder peak at 2955 cm-' which are assigned to symmetric and asymmetric C-H stretch-ing vibrational modes, respectively The two peaks became considerably smaller when the BTO was heated at 973 K, and completely disappeared at temperatures > 1173 K Fig 5 shows the EPR spectra of BTO measured at 77 K in the presence of gaseous oxygen Two signals appearing at g= 2034 and g= 1981 were due to Mn2+ which was used to calibrate the g values No characteristic EPR signal was observed upon UV irradiation of BTO(873) For BTO calcined at 1173 K, two very strong peaks appeared at g=2018 and g=2004 These signals were not observed prior to UV 1 1001 3100 3000 2900 2800 2700 wavenumberkm-' Fig.4 FTIR spectra of (a) BT0(873), (b) BT0(973), (c) BT0(1073), (d) BTO( 1173) and (e) BTO( 1273) ~~2.018-3 I I L 3200 3250 3300 magnetic field/G Fig.5 EPR spectra of (a) BT0(873), (b) BT0(973), (c) BT0(1073), (d) BTO( 1173) and (e) BTO( 1273).All the EPR spectra were obtained at 77 K with UV irradiation after evacuation at 573 K for 1 h followed by the introduction of O2 at a pressure of 30 Torr. irradiation, and they remained stable until the UV irradiation was turned off. The peaks increased further for BTO( 1273). BTO samples treated at various calcination temperatures were combined with RuO, and employed as photocatalysts for the decomposition of water. Fig. 6 shows the photocatalytic activity for the production of hydrogen and oxygen. The activity of Ru0,-BTO( 873) is quite low: neither hydrogen nor oxygen was produced to a noticeable extent. The activity became higher with increasing calcination temperature of BTO: both hydrogen and oxygen were produced for Ru0,-BTO( 1073).The activity increased significantly for Ru02-BT0(1173), but it decreased by 20% for Ru0,-BTO(1273). The ratio of reaction rate for hydrogen prod- uction to that for oxygen production was different from the expected stoichiometric ratio of 2: it was 4 for a catalyst using BT0(973), 2.9 for BTO( 1073), 2.7 for BTO( 1173) and 2.6 for BTO(1273). This shows that the production of oxygen is lowered for BTO calcined at lower temperatures, but it is improved with increasing calcination temperatures. 10 t Fig, 6 Water decomposition on Ru0,-combined BTO. (a) BTO( 873), (b) BT0(973), (c) BTO( 1073), (d) BTO( 1173) and (e) BTO( 1273). U, H2; 0,02.Discussion A large mass loss appearing at temperatures between 573 and 673 K in the TG curve accompanies a large exothermic peak. In the reaction of barium acetate with titanium isopropoxide in the~presence of water, hydrolysis and polymerization pro- duce acetic acid and propanol as products. Thus, the changes correspond to the evaporation of these products. Two small peaks were observed at 980 and at 1095 K in the DTA curve, which were accompanied by no significant mass loss, and it appears that these are associated with phase changes. BTO( 873) showed no characteristic X-ray diffraction peaks and provided broad and weak Raman peaks in the four regions observed. Since the strong peaks due to the C-H vibration were present in the FTIR spectra, these results indicate that the structure of the titanate is amorphous and contains hydro- carbon residues.The X-ray diffraction pattern and Raman spectrum of BTO(973) show the development of definite peaks, followed by a decrease in the intensity of the C-H peaks. These findings mean that crystallization takes place at ca. 973 K, removing the hydrocarbon residues. Drastic increases in the intensities of both the diffraction and Raman peaks for BTO( 1173) show that crystallization proceeded intensively above 1173 K. At these temperatures, the hydrocarbon residues were eliminated so that they were not detected in the FTIR spectra. The Raman spectrum of BTO( 1273) exhibited major peaks in three regions: 430-450, 590-650 and 860cm-'.The peak appearing at the highest wavenumber is interesting, since it represents a feature of Ti-0 bonding. Gratzel and Rotzinger" showed that TiO(C10J2 gives rise to a peak at 970cmP1, which is assigned to the stretching vibration of short Ti-0 bonds. Dehnickell also pointed out that the peak of short Ti-0 bonds appeared at 836 cm-'. Therefore, the 860 cm-' peak of BTO( 1273) is associated with the stretching vibration of Ti-0 with a small separation. This assignment is supported by the results obtained from the X-ray diffraction of a BTO crystal: the TiO, octahedra of BTO contain short Ti-0 bonds, with distances of 0.17-0.18 nm. Because of the presence of the short Ti-0 bonds, the TiO, octahedra are distorted so significantly that the position of the Ti ion deviates from the centre of gravity of the surrounding six oxygens so producing a dipole moment.There was no characteristic EPR peak for BTO(1073), whereas for BT0(1173), EPR signals with g=2.018 and g= 2.004 were obtained in the presence of oxygen at 77 K upon UV irradiation [Fig. 5(d)]. The signal did not appear in the dark in an oxygen atmosphere, or in the absence of oxygen under UV irradiation: note that both oxygen and UV light are essential for the generation of the signal. Lunsford et ~1.'~ showed that the adsorption of N20 on ZnO gives rise to EPR signals with g =2.021 and 2.0026, which are assigned to surface 0-species. Shvets and Kazansky13 also assigned the EPR signals with g =2.020 and g= 2.006 for N,O adsorption on Mo/Si02 to 0-.The g values and the shape of the signal observed here are in good agreement with those values associ- ated with an oxygen radical, 0'-.The formation of the EPR- active species indicates that the photoexcitation and separation of the excited charges occurs efficiently for BTO calcined at higher temperatures. The calcination temperature at which the 0'-peak appeared is in agreement with that for the crystalliz- ation of BTO. These findings imply that crystallization of BTO leads to an increased ability for the separation of photoex- cited charges. The photocatalytic activity for water decomposition increased with increasing calcination temperature of BTO. Fig. 7 shows the relation between the crystallinity of BTO and the photocatalytic activity. This indicates that the well defined crystal structure is preferred to the amorphous structures.As shown in the Raman spectra, BTO provides distorted TiO, J. Mater. Chem., 1996, 6( 12), 1921-1924 1923 crystallinity (%) Fig. 7 Dependence of photocatalytic activity on crystallinity of BTO octahedra such that a dipole moment is generated in the distorted Ti0, octahedra In previous studies, we proposed that the dipole moment is useful for the charge separation in photoexcitation l4 The distorted structure played a much more important role in the rigid crystal structure than in the amorphous structure, since the distorted structure is more stably preserved for the crystal phases This explains the relation between the high photocatalytic activity and the crystal structure of BTO The decrease in the photocatalytic activity for Ru02-BT0(1273) is possibly due to the growth of RuO, particles deposited on BTO, since the surface area of BTO is significantly smaller This makes it difficult to deposit a small particle of Ru02 on the BTO surface, which lowers the activity It is proposed that the formation of a well defined pentagonal-prismatic tunnel structure due to the crystallization of BTO is important for the effective separation of photoex- cited charges The H2/02 ratios were different from the expected stoichio- metric ratio of 2 it was 4 for a catalyst using BT0(973), 29 for BTO( 1073), 2 7 for BTO( 1173) and 2 6 for BTO( 1273) The evolution of oxygen increased with increasing crystallinity of BTO One of the reasons for lower oxygen production with less crystallized BTO may be the consumption of oxygen by the reaction with hydrocarbon residues within the BTO The generation of 0 -species upon UV irradiation occurs at a calcination temperature of 1173 K, whereas the photocatalyst became active at lower calcination temperature, 1073 K The difference is attributed to the surface-bulk properties for EPR measurements, bulk phenomena are involved, while the photo- catalytic activity is based on the surface region Since the removal of hydrocarbon residues occurs preferentially at the surface, it follows that the photocatalytic activity is generated at calcination temperatures lower than the generation of 0 -species In conclusion, BaTi409 synthesized by a sol-gel process exhibits photocatalytic activity for water decomposition The preparation of BTO films coated on porous supports is feasible by this method, which is able to broaden the application of this material to practical photocatalytic reactions such as the purification of water and air This work was supported by a Grant-in-Aid for priority areas from the Ministry of Education, Culture, Sports and Science References 1 Y Inoue, Y Asai and K Sato, J Chem Soc Faraday Trans, 1994, 90,797 2 C Anderson and A J Bard, J Phys Chem ,1995,99,9882 3 I Moriguchi, H Maeda and S Kagawa, J Am Chem Soc 1995, 117,1139 4 T Yoko, A Yuasa and K Kamiya, J Electrochem Soc, 1991, 138,2279 5 N Negishi, T Iyoda, K Hashimoto and A Fujishima, Chem Lett, 1995,841 6 T Yoko, K Kamiya and K Tanaka, J Mater Sci ,1990,25,3922 7 Y Nosaka, M Jimbo, M Aizawa, N Fujii and R Igarashi, J Mater Scz Lett, 1991,10,406 8 H Lu, L E Burkhart and G L Schrader, J Am Ceram SOC, 1991,74,968 9 K Lukaszewicz, Rocz Chem, 1957,31,1111 10 M Gratzel and F P Rotzinger, Znorg Chem ,1985,24,2320 11 K Dehnicke, 2 Anorg Allg Chem ,1961,309,266 12 N B Wong, Y B Taant and J H Lunsford, J Chem Phys, 1974, 60,2148 13 V A Shvets and V B Kazansky, J Catal, 1972,25,123 14 M Kohno, S Ogura and Y Inoue, manuscript in preparation Paper 6/043561, Received 24th June, 1996 1924 J Mater Chem, 1996, 6(12), 1921-1924

ISSN:0959-9428    

DOI:10.1039/JM9960601921

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

_________~~~~~~~~~~ Preparation of anatase, brookite and rutile at low temperature by non-hydrolytic sol-gel methods Pascal Arnal, Robert J.P. Corriu, Dominique Leclercq, P. Hubert Mutin and AndrC Vioux* Laboratoire des pricurseurs organomitalliques de matiriaux, case 007, Universitt? Montpellier II, 34095 Montpellier Cidex 5, France Titania samples prepared by different non-hydrolytic sol-gel methods, mainly based on the etherolysis and alcoholysis of titanium tetrachloride, have been found to differ in both structure and texture. Thus, the reaction of diethyl ether with TiCl, at 110 'C affords anatase, which begins to convert into rutile only around 1000"C. The reaction of TiC1, with ethanol leads to rutile as early as 110 "C, whereas the reaction of tert-butyl alcohol at 110 "C leads to the singular formation of brookite.The development of low-temperature routes to transition-metal oxide processable materials underlies the recent works in the field of sol-gel synthesis.' In these methods based on the hydrolysis of molecular precursors, such as metal alkoxides, the major problem is control of the reaction rates which are generally too fast. An attractive solution is to use organic additives which act as chelating ligands (carboxylic acids, p-diketones, etc.) and modify the reactivity of the precursors.2 Here we propose various non-hydrolytic sol-gel routes tested in the case of titanium, chosen as a representative transition metal. As reported previously, a novel sol-gel route is provided by the thermal condensation of metal halides with metal alkox- ides:3 MX, +M(OR), -t2MO,,, +nRX (1) Alternatively, it is possible to generate the alkoxyl groups in situ by the action of alcohol4 or dialkyl ether5 on metal halides: =M-X +ROH-M-OR +HX or EM-X +ROR-tE M-OR +RX (2) Another possible variation is the direct reaction of anhydride with alkoxide precursors,6 which leads to acylation, then condensation, via the formation of an ester: M(OR),+n/2 (R'C=O),O+MO,,,+nR'COOR (3) In these non-hydrolytic reactions, the metal alkoxide, dialkyl ether, alcohol and anhydride act as oxygen donors, instead of water.The main features of these methods are: (i) low synthesis temperatures (cu. 100-150°C); (ii) a simple reaction of TiC1, [or Ti (OR),] with readily available compounds; (iii) an easily removed by-product (hydrogen halide, alkyl halide or acetate); (iv) no cosolvent required (otherwise needed to dissolve water).This paper falls into two parts. First, four typical non-hydrolytic routes to titania were compared [systems TiC1, plus Ti(OPr'),, TiCl, plus diisopropyl ether, TiC1, plus isopropyl alcohol, Ti(OPr'), plus acetic anhydride]; the structures and textures of the different samples were studied by means of X- ray diffraction (XRD) and BET measurements. Secondly, the influence of the nature of the oxygen donor on the structure and crystallisation behaviour of the TiO, precur-sors was investigated in the etherolysis and alcoholysis methods. Background Hydrothermal and sol-gel methods have attracted much inter- est in the preparation of titania powders and colloids because of the many uses of this material, as a pigment, filler and, more recently, as a membrane, anti-reflection coating, catalyst and pho toca taly st.Titania sol-gel synthesis has been developed from inorganic precursors and from metal-organic Ti(OR), precursors.' Thus, TiO, gels can be obtained by adding a weak base [e.g.Na,C03, (NH,),C03] to a solution of sodium titanate in concentrated hydrochloric acid, whereas sols can be obtained by the addition of TiC1, or TiO(NO,), to acidic aqueous solution.'*7 Moreover, stable clear sols, which tend to turn into monolithic gels, can be obtained from Ti(OR), precursors by using high water/ alkoxide ratios and inorganic acids as peptization Ti0,-based gels or colloids can also be obtained after a chemical modification of titanium alkoxides by chelating agents, such as acetic acid and acetylacetone." Titanium dioxide is known to exist in three crystalline modifications, namely rutile (tetragonal), anatase (tetragonal) and brookite (orthorhombic).Anatase and rutile are the common polymorphs of synthetic titania. The titanium atoms are octahedral in both structures. In fact, crystallisation is highly influenced by the hydrolysis conditions." During the condensation process, the formation of the kinked chains of edge-sharing octahedra corresponding to anatase appears more probable than the formation of the straight chains typical of rutile.Therefore, anatase is obtained in processes under kinetic control, whereas processes involving Ostwald ripening lead to the equilibrium phase, i.e. rutile.'* On the other hand, the brookite structure, in which each octahedron shares one edge, has not been X-ray characterised to date, to our knowledge, in sol-gel products. Usually, amorphous TiO, crystallises into anatase below 4OO0C, which is further converted to rutile from 600 to 1100"C.I3The rates of transformation are markedly influenced by particle size or the presence of imp~rities.'~ The anatase- rutile transformation may be detrimental in many applications. It involves a collapse from the relatively open anatase structure with a cell volume of 0.068 nm3 to rutile with a cell volume of 0.0624nm3, a volume change of ca.8%. So the structural transformation may be damaging for supported membranes owing to the extra stresses introduced by this volume change during sintering.15 Moreover, the anatase-rutile transformation is considered to represent the major factor in catalyst deacti- vation in selective oxidation catalysis.16 In the same way, rutile has been found to be unable to sensitise the photomineralis- ation of phenol derivatives, unlike anatase." This behaviour J. Muter. Chem., 1996, 6(12), 1925-1932 1925 appears to be associated with the concomitant decrease in specific surface area Experimental General Titanium chloride and titanium tetraisopropoxide [Ti(OPr'),] were purchased from Aldrich Titanium tetra-tert-butyloxide [Ti(OBut),] was prepared from Ti(OPf), and tert-butylacet- ate according to the literature l8 Ethers and alcohols were distilled from the appropriate drying agents prior to use All manipulations were carried out under argon, in oven-dried glassware, to preclude any hydrolysis side reaction Reactions of TiCl, with titanium isopropoxide and ethers The oxygen donor was added to titanium chloride (typically 2-5 ml) at 0 "C in a tube under argon (Table 1) An exothermic reaction occurred and in most cases a precipitate was formed, which melted on heating The tube was frozen in liquid nitrogen then sealed under vacuum The sealed tube was held at 110 "C for 7 days, then it was opened in a glove bag under argon The solid and the liquid phases were separated by filtering The solid was washed with successive portions of dichloro- methane, then dried under vacuum Reactions of titanium isopropoxide with acetic anhydride Acetic anhydride (6 4 ml, 67 mmol) was added to titanium isopropoxide (10 ml, 33 5 mmol) in a Schlenk tube under argon, an exothermic reaction occurred, which led to a yellow solution After the addition of a small amount of TiC1, (004 ml, O36mmol), the mixture was transferred to a tube, frozen in liquid nitrogen, then sealed under vacuum The tube was kept at 140°C for 7 days A white solid was obtained which was isolated by the above procedure Reactions of TiC14 with alcohols Titanium chloride was added dropwise at room temperature with stirring to an excess of alcohol (TiC1,-ROH, 1 6) under argon Viscous yellow solutions were obtained, except in the case of tert-butyl alcohol where a yellow precipitate formed initially, which then dissolved on slow heating The solutions were stirred at room temperature for 2 h to allow complete HC1 evolution Thereafter they were allowed to stand for 7 days at 110 "C, under autogenous pressure in sealed tubes C haracterisa tion methods The syneresis liquids were analysed by 'H NMR spectroscopy and by gas chromatography (GC) The elemental analyses of solids were performed by the Service Central &Analyses of CNRS (Vernaison, France) C content was determined by IR spectroscopy after high-temperature combustion, C1 content was determined by potentiometric titration, Ti content was determined by induction coupling plasma (ICP) from an aqueous solution Thermal analyses were performed at a heating rate of 10Krni11-~ in a 20 80 0,-N, mixture on a Netzsch STA 409 thermobalance In some cases the thermobal- ance was coupled with a Balzers QMG 421 mass spectrometer, thus allowing the continuous analysis of the gases evolved Speclfic surface areas were determined by the Brunauer-Emmett-Teller (BET) method, using nitrogen adsorption- desorption isotherms recorded on Micromentics ASAP 2400 analyser (estimated error d 5%) X-Ray powder diffraction (XRD) patterns were recorded with Cu-Ka radiation using a SEIFERT MZ IV diffractometer (4 scans, with 1000 digitised points and 500 ms acquisition time, 8 angle ranging from 5 to 40") The mass percentage of rutile in the mixtures was calculated using %R= 1/[ 1+0 8 (IA/IR)],where I, and IRare the intensities of the (101) reflection of anatase and the (1 10) reflection of rutile, respectively l9 Results Comparison of different non-hydrolytic methods Four preparation modes, involving parent byproducts (isopro- pyl chloride or isopropyl acetate), were compared (1) the reaction of TiCl, with Ti(OPr'), (equimolar), (2) the reaction of TiCl, with diisopropyl ether, Pf20[2 (A) and 3 (B) equiv 3, (3)the reaction of Ti(OPr'), with acetic anhydride, (CH3C0)20 (2 equiv), (4) the reaction of TiCl, with an excess of isopro- pyl alcohol Experimental details are given in Table 1 In methods 1 and 2, the reactions were carried out at 110 "C under autogenous pressure in sealed tubes, without any cosolvent In fact, when mixed at room temperature TiCl, and Ti(OPf), give chloroal- koxide precipitates which melt readily below 110 "C 2o 21 TiCl, is known to form Lewis adducts with ethers, however, with diisopropyl ether the adduct was not isolated and the precipi- tate obtained at room temperature was ascribed to the forma- tion of the substitution compound TiCI,(OPr') 22 No cosolvent was needed in method 3 either, since the reaction of Ti(OPr'), with 2 equiv of acetic anhydride is known to yield liquid Ti(OAc),(OPr'), 23 The strong evolution of HC1 (exothermic reaction) which occurred when isopropyl alcohol was added to TiCl, clearly indicated the formation of alkoxide groups In fact, whereas the reaction of alcohol with silicon tetrachloride usually leads to tetraalkoxides Si(OR), , compounds of formula TiC12(0R), .ROH (which are soluble in alcohol and are poten- tial candidates for non-hydrolytic condensation) are generally obtained from TiCl, and alcohol 24 Polycondensation took place within a few hours at 110 "C, except in the reaction of Ti(OPr'), with acetic anhydride where a temperature of 140°C was needed Note that in the latter case no solid formed in the absence of some TiCl, catalyst The total heating time was always 7 days The solids were obtained as white agglomerates (grains gathered in a mass), except in the case of alcoholysis in which a precipitate (separate grains) was obtained In all cases the GC analysis of the expelled liquid (syneresis liquid) established the expected formation of isopropyl chloride Table 1 Comparison of different non-hydrolytic sol-gel routes to TiOz preparation of the samples (7 days reaction time) ~~~~~ method reagents 1 TiCl,, Ti(OPr'),, no solvent 2A TiCl,, 2Pf20, no solvent 2B TiCl,, 3Pr'20, no solvent 3 Ti(OPr'),, 2(CH,CO),O, 1% TiCI,, no solvent 4 TiCl,, excess Pr'OH, c= 1 8 mol I-' "Time for which the formation of a solid phase was observed 1926 J Muter Chem, 1996, 6(12), 1925-1932 gel time"/h, temperaturePC ca 5, 110 ca 2, 110 ca 3, 110 ca 19, 140 ca 6, 110 appearance white agglomerate white agglomerate white agglomerate white agglomerate white precipitate TiOZ yieldb byproducts (%I Pr'C1+ hydrocarbons 90 Pr'Cl +hydrocarbons 95 Pr'Cl +hydrocarbons 91 AcOPr' 75 Pr'Cl +hydrocarbons 86 From TG, after calcination in air at 1000 "C Table 2 Comparison of different non-hydrolytic sol-gel routes to TiO,: characterisation of the different titanias dryingmethod' temperature/"C 1 110 180 2A 110 180 2B 110 3 140 4 110 ~~ ~ S/m2 g-' XRD (after XRD (after composition from elemental analysis Am/m (Yo)(from formula) Am/m (YO) (after calcination (TG) at 500 "C for 5 h) calcination at 500"C for 5 h) calcination at 950 "C) TiC1O.llCO.85~ Tic1,., 1(0pf)0.2801 .6 160 anatase anatase 100% 80% condensation Tic10.04c0.67 TiC10.04(0Pf )0.2Z01 .87 -13 -14 94% condensation TiC10.28C0.87 120 anatase anatase 86% Tic10.28(OPf )O.2g0l72% condensation .43 +rutile 14% Tic1,.14c0.41 14(0Pf)0.1401.86 -12 -16 93% condensation TiC10.09Co.10 63' anatase rutile 100% Tic10.09 (OPf )0.03O1.94 -8 -8 97% condensation Tic10.08c2. 18 47 anatase anatase 25% Tic10.08 (0Pr')0.44(0Ac)0.~401.56b -36 -39 +rutile 75% 76% condensation TiC10.07C0.03 40 anatase anatase 63 YO Tic10.07~0pf~0.0101.95 -3 -7 +rutile 37% 98% condensation Systems 1 TiCl4/Ti(0Pf),; 2A TiCl,, 2Pr1,0; 2B TiCl,, 3Pt20; 3 Ti(OPr'),, 2(CH,CO),O; 4 TiCl,, excess PfOH. Assuming an equal number of OPf and OAc groups. 174 m2 g-' after calcination in air at 200 "C for 3 h.or acetate. In methods 1,2 and 4, the presence of hydrocarbons was detected as well as the alkyl chloride; they may be ascribed to the polymerisation of the alkene arising from the partial dehydrochlorination of the alkyl chloride (both elimination of HC1 and polymerisation of alkenes are known to be catalysed by titanium chloroalkoxides).25 Table 2 sets out the compositions (based on elemental analyses) of the samples after vacuum drying for 3 h at the given temperature. The number of residual OR groups (and OAc and C1 groups) per Ti atom are calculated from carbon contents (and chloro contents), and the number of 0x0 bridges are deduced by difference (to meet valence requirement, assuming no presence of hydroxy groups). Therefore the calcu- lated formulae TiCl,(OR),O, give an estimation of the degree of condensation [condensation (%)=z/2 x 1001.The mass losses, Am/m, based on thermogravimetry (TG) from 200 to 1000°C are in rather good agreement (except for sample 4) with those calculated from the formulae (Table 2); this confirms the postulated small amount of hydroxy groups. X-Ray diffraction (XRD) indicated the incipient crystallis- ation of anatase in all the as-prepared samples. In Fig. 1, the XRD patterns of sample 1 after annealing at different tempera- tures show only the reflections of anatase. However, the progress of the anatase-rutile transformation at 950 "C (with- out holding time) was found to be quite uneven depending on the preparative method (Table 2; e.g.75% rutile in method 3 us. 100% anatase in method 1). No explicit exothermic peak associated with crystallisation or phase transformation (and not associated with mass loss) was recognized from the differential thermal analysis (DTA) curves. An additional etherolysis experiment, involving an excess of ether (3 equiv.; Table 1, method 2B), was performed for a comparison with method 4 in which an excess of alcohol was used. Anatase was produced again; however, the anatase-rutile transformation was found to be promoted (accompanying a higher degree of condensation) compared to the stoichiometric etherolysis (Table 2, methods 2A and 2B). The specific surface areas of the samples were measured by the BET method after 5 h heat treatment at 500 "C (Table 2).The low values found in methods 3 and 4 (i.e. acetate and isopropyl alcohol methods) are related to the progress of crystallisation. On the other hand, the specific surfaces found for methods 1 and 2A are 160 and 120 m2 g-', respectively (174 m2 g-' in method 2B after calcination at 200 "C for 3 h). The nitrogen adsorption-desorption isotherms obtained for methods 1 and 2A are of type IV according to the BDDT classification26 (mesoporous solids; Fig. 2). .-1801 5 10 15 20 25 30 35 40 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 e/degrees relative pressure,P/Po Fig. 1 X-Ray diffraction patterns of sample 1, arising from the reaction of TiCl, with Ti(OPr'),, after heating at (a) 180, (b) 500, (c) 950°C Fig.2 Nitrogen adsorption-desorption isotherms of sample 2A,arising from the reaction of TiCl, with Pr',O, after calcination at 500 "C for 5 h J. Mater. Chem., 1996, 6(12), 1925-1932 1927 Influence of the nature of the organic oxygen donor (ether or alcohol ) Etherolysis. Titanium chloride was reacted with diisopropyl ether, di-n-propyl ether, tetrahydrofuran (THF), diethyl ether and dimethoxyethane (DME) Table 3 shows that gel time (the time for which a solid phase was observed) depends on the dialkyl ether used and it increases in the order Pr',O <Et,O <THF <PP,O <DME The analysis of the syn- eresis liquid confirms the formation of alkyl chloride as bypro- duct (with the partial isomerisation of n-propyl chloride into isopropyl chloride in the case of di-n-propyl ether) In the case of DME no liquid could be isolated The mass losses, Am/m, measured by thermogravimetry (Table 4) are related to the amount of residual groups (organics and chloride) present In the dried samples (after vacuum drying at 180 "C), this was found to be rather low (16-22%), except for with DME In this case the high mass loss (Am/m= 47%) must be related to the absence of 1,2-dichloroethane as the syneresis liquid Apparently DME reacts mainly through the methoxy end groups, giving an inorganic-organic hybrid compound (volatile methyl chloride was not isolated) -TiCl+ CH3-0CH,CH,O-CH3 + ~Ti-OCH2CH20-Ti~ +2 CH3C1 (4) This was supported by the continuous analyses of the gases formed during the heat treatment, by coupling the thermobal- ance with a mass spectrometer, which allowed the detection of (Fig 3) (a) between 150 and 400°C 172-dichloroethane(m/z 62) and acetylene (m/z 25), as well as HCl (m/z 36) and H20 (m/z 18), (b) between 400 and 650°C CO, (m/z 44) and H20 (m/z 18) associated with the combustion In fact, the gases produced between 150 and 400°C may be accounted for by the following reactions ~T1-OCH2CH20-Ti--' +2 =TlCl+ 2 Ti-0-Ti= +C1CH2CH2C1 (5) ClCH,CH,Cl+HC=CH +2 HC1 (6) -Ti- OCH2CH20-Ti= +2=Ti- OH +HC-CH (7) 1 TG -1004 -.-. -. -. . 0 200 400 600 800 lo00 ,II, ml544 II /I I-........0 200 400 600 800 loo0 TI% Fig. 3 Thermogravimetric and differential thermal analysis in air of the sample arising from the reaction of TiCl, with DME, and continuous mass spectrometry of the gases evolved during the heat treatment Note that the precursor derived from DME was quite amorphous after vacuum drying at 180°C for 3 h, whereas incipient crystallisation of anatase was discernible in other precursors (Fig 4) However, XRD indicated that all the samples gave anatase after calcination at 500°C One can see from Table4 that the surface areas, which varied from 45 to 120m2 g-' are related to the crystallinity of the samples [Fig 5(u)] Upon heating to 950°C (without holding time), the transformation to rutile was complete only for the sample Table 3 Preparation of the samples by etherolysis of TiCl, (7 days reaction time) Ti0, yield' ether gel timea/h appearance by products (%) Pt,O ca 2 white agglomerate Pf20 ca 70 white agglomerate PtC1+ hydrocarbons Pr"C1 +Pr'C1-t hydrocarbons 95 81 THF ca 60 white agglomerate CI(CH,),Cl 89 Et,O ca 17 white agglomerate EtCl 74 DME ca 85 black powder - 80 a At 110 "C 'From TG, after calcination in air at 1000"C Table 4 Characterisation of the different titanias prepared by etherolysis of TiCl, composition ether (condensation degree)" pr'20 14(OPf)0 14O1 86 (93%)..Pr"20 Tic1, 34(OPf)0 0701 80 (90%) THF TIC10 13(ORC1)~ 1501 86' (93%) Et20 41(OEt)0 16O1 67 (83%) DME 09(ORo)0 55OO 91' (46 Yo) XRD (after vacuum XRD (after XRD (after s/mzg (after Am/m (%) (from formula) Am/m (YO) (TG) drying at 180°C) calcination at 500°C for 5 h) calcination at 950 "C) calcination at 500°C for 5 h) -12 -16 -15 -16 -16 -18 amorph + anatase amorph + anatase amorph + anatase amorph + anatase amorph + anatase amorph + anatase anatase 86% anatase 89% anatase 57% +rutile 14% +rutile 11YO +rutile 43% 120 58 73 -17 -40 -22 -47 amorph + anatase amorph amorph + anatase amorph + anatase 42% rutile 100% +rutile 58% 45 52 anatase a After vacuum drying at 180 "C 'R =(CH2)4 R =CH,CH, 1928 J Muter Chem , 1996, 6(12), 1925-1932 Pr',O I DME I...-.----...--.--...--....-...-.-.I S 10 IS 20 25 30 35 40 Bldegrees Fig.4 X-Ray diffraction patterns of the samples arising from the reactions of TiC1, with DME and with Pr',O (sample 2A),after vacuum drying at 180°C for 3 h derived from DME, whereas the transformation had just begun for the sample derived from Pf,O [Fig.5(b)]. Alcoholysis. Titania precursors were prepared from primary alcohols (ethanol, propan-1-01, butan-1-01, ethylene glycol), secondary alcohols (isopropyl alcohol, butan-2-01) and tertiary alcohols (tert-butyl alcohol, tert-amyl alcohol). Experimental details for gelation are given in Table 5. Gelation occurred within the range 0.5 h (tertiary alcohol) to 18 h (butan-1-01). In all the cases the expected alkyl chloride was identified by GC in the syneresis liquid. However, note the partial isomeris- ation of Bu'Cl into ButCl in the case of butan-2-01. The mass losses were measured by TG (Table 6).Values as low as 44% were indicative of high condensation degrees.The unique large mass loss (Am/m =32.6%) observed with ethylene glycol as the oxygen donor may be accounted for as in the above-mentioned case of DME. Table 6 indicates the components of the samples after vacuum drying at room temperature or at 11O"C, and after calcination in air at 500 and 950 "C, based on X-ray diffraction analyses (Fig. 6). Dramatic differences are observed, depending on the nature of the alcohol, as early as room temperature. Thus, anatase was visible in the samples derived from propan- 1-01, propan-2-01, butan-1-01 (and this phase was retained to 950 "C), whereas rutile was present in the samples arising from ethanol and butan-2-01.The reaction of tert-butyl alcohol with TiC1, leads to the formation of a precipitate, which is yellow and which does not melt below 110°C. However, it dissolves in an excess of alcohol. A translucent gel was obtained within 1h at 110 "C, afterwards the gel expelled all the solvent on heating, and turned into an agglomerated powder. The X-ray diffraction pattern exhibits the typical reflections of brookite and rutile (Fig. 7). The conversion of brookite to rutile was almost complete after 5 h heat treatment at 500 "C. Traces of brookite were also detected in the as-prepared xerogel derived from tert-amyl alcohol, but they disappeared after vacuum drying at 110 "C. The unique formation of brookite from tertiary alcohols led us to conduct further investigations.The role of the alcohol Prn20 THF Eta0 DME I t I I I I 1 1 II 5 10 15 20 25 30 35 40 5 10 15 20 25 30 35 40 eldegrees Fig. 5 X-Ray diffraction patterns of the different titanias prepared by etherolysis of TiC1, after calcination (a) at 500°C for 5 h; (b)at 950°C without holding seems to be decisive in the formation of brookite. Indeed, brookite was not formed in the absence of ButOH: neither by reacting TiC1, and Ti(OBut),, nor by reacting TiC1, and Ti(OPr'), in the presence of 4equiv. of ButCl (only anatase Table 5 Preparation of the samples by alcoholysis of TiCl, (7 days reaction time) gel time/h TiO, yield" alcohol (110 "C) appearance byproducts (%) EtOH ca. 6 white agglomerate EtCl 91 Pr"0H ca.7 white agglomerate Pr"C1 85 Bu"OH ca. 18 white agglomerate Bu"C1 83 Pr'OH Bu'OH HOCH,CH,OH Bu'OH AmtOH ca. 6 ca. 3 ca. 50 ca. 0.5 ca. 0.5 white agglomerate white agglomerate white gel white gel white gel Pr'C1 +hydrocarbons 86 Bu'Cl +Bu'Cl +hydrocarbons 88 60 CH,CH,C( CH,),Cl +hydrocarbons 87 BuTl +hydrocarbons 95 a From TG; after calcination in air at 1000 "C. J. Muter. Chem., 1996, 6(12), 1925-1932 1929 Table 6 Charactensationof the different titanias prepared by alcoholysis of TiCl, and from the reaction of TiC1, with Ti(OBu'), oxygen donor EtOH HO(CH2)20H PPOH Bu"0H Pr'OH Bu'OH Bu'OH Am'OH Ti(OBu'), composition (from elemental anal after vacuum drying at 110"C) O4(OEt)003O1 96 (97% condensation) 15(0R0)0 78O1 14 (57%) TIC10 07(OPr')o 0201 95 07(0Bu")0 04O1 94 07(0Pr')0 Olol 95 04(0Bus)0 OOSO1 97 04(0Bu')0 OOSO1 98 TIC10 os(OAm')o 0101 97 27(OBU')0 19O176 (88 5%) Amlm (YO) AmJm (%I XRD (vacuum XRD (vacuum XRD (calcined XRD (calcined S/m2 g-' (calcined (formula) (TG) dried RT) dried 1lOOC) 500"C, 5 h) 9503C) 500"C, 5 h) -3 -4 amorph +rutile amorph +rutile rutile rutile 10 -32 -33 amorph +anatase amorph +anatase amorph +anatase anatase 2% +rutile 98% 35 -4 -7 amorph +anatase amorph +anatase anatase anatase 45% 30 frutile 55% -5 -8 amorph +anatase amorph +anatase amorph +anatase anatase 64% 43 +rutile 36% -3 -7 amorph +anatase amorph +anatase anatase anatase 63% 40 +rutile 37% -2 -5 amorph +anatase +rutile amorph +anatase +rutile anatase +rutile anatase 9% +rutile 91% 16 -2 -4 amorph +rutile +brookite amorph +rutile +brookite amorph +rutile +brookite rutile 23 -2 -4 amorph +rutile +brookite amorph +rutile amorph +rutile rutile 28 -20 -19 - amorph +anatase amorph +anatase anatase 100% 64 was obtained in both cases at 110 "C).Nevertheless, upon heat treatment in ButOH at 110°C for 7 days of an amorphous sample prepared from Ti(OPr'), and TiC1, (7 h heating at 110 "C in a sealed tube), only anatase was obtained. Moreover, the composition of the precipitate formed initially in tert-butyl alcohol before dissolution did not correspond to the expected formula TiCl,(OR), -ROH.The elemental analy- sis found (C, 4.68; C1, 18.66; Ti, 36.65%) better corresponds to the hydroxide; TiC1, 68(0B~t)o 127(0H)319, or to oxotitanate, TiCl,.,8(OBut)o 12701The mass losses, Arn/rn, calculated59. from these formulae are 41 and 25%, respectively. The exper- imental value (38%) is rather close to the former; furthermore, the latter would correspond to a condensation degree as high as 80%, which is inconsistent with the dissolution of the precipitate. However, the hydroxide precipitate did not crystal- lise to brookite on heating at 180 "C, after Bu'OH was removed by filtering and drying. Discussion Elemental analyses and mass losses measured by TG permit the evaluation of the amount of residual groups and, therefore, the comparison of the condensation degrees achieved in the different non-hydrolytic methods.Note that the reaction of acetic anhydride with Ti(OPr'), hardly occurs at 140 "C, and only then in the presence of TiC1, as a Lewis-acid catalyst. The catalytic efficiency of the Lewis acid FeC1, has been reported in the case of the polycondensation of R,Si(OAc), with PhSi(OR')3,27 but this type of catalysis has never been used in the sol-gel area to our knowledge, although it may be of some interest.28 However, this condensation reaction is a mediocre method in the case of titanium, compared to the reaction of Ti(OPr'), with TiCl,, or the etherolysis and alcoholysis of TiC1,. In the latter cases the non-hydrolytic condensation degree is improved by using an excess of the oxygen donor (in Table 2, compare samples 2A and 2B) or an additional heat treatment (in Table 2, compare samples 1 and 2A after drying at 110 and 180 "C).Thus, by reacting an excess of diisopropyl ether with TiC1, at 110 "C, an amount of about one residual group (OPr' plus C1) per every eight Ti atoms was found on average in the solid (Table 2). It is interesting to compare this result to that reported for the conventional aqueous route: alkoxy groups (Pr'O) remained at more than about one group for every six Ti atoms in samples produced by the hydrolysis of Ti(OPr'), (after heat treatment at 110°C for 7 h), even though a large amount of HCl solution was used." It is known that residual alkoxy groups determine the crystallisation behaviour of TiO, precursors.As a matter of fact, steric hindrance by residual unhydrolysed isopropoxy groups, preventing crystallisation to anatase, was reported." However, it appears that huge amounts of alkyls left over, such as OCH,CH,O bridges (Tables 4 and 6), bring forward the anatase-rutile transformation (whereas the presence of OCH,CH,O bridges delayed anatase formation at lower tem- perature, as shown in Fig. 4). This behaviour might be related to the exothermic oxidation of organics (at ca. 350"C), which brings about a local increase in temperature sufficient to promote the crystallisation of rutile when the organic content is high. Moreover, one has to keep in mind that the crystallis- ation of the anatase phase and the transformation from anatase to rutile are known to also take place on ageing of gel precursors." Our experimental conditions, especially in the case of alcoholysis, are reminiscent of solvothermal conditions (a prolonged heat treatment in an autoclave in the presence of a protic ~olvent).~' The effect of the alcoholic medium on crystal- lisation is most plausible.The ability of alcohols to cleave Ti-0-Ti linkages (thus slightly reversing the condensation reaction) may be invoked. Indeed, it has been reported recently that ethanol rinses, at room temperature, are able to convert a nanocrystalline anatase precipitate to an amorphous phase.30 Moreover, note the recently reported synthesis of materials with particular structures and morphologies in alcoholic media, more particularly via glycothermal treatment.31 However, no template effect (in the sense of structure director) was observed with ethylene glycol in our case.Moreover, note that no effect was observed when ButCl was used as the solvent in the reaction of TiC1, with Ti(OPr'),, although one would have expected that it would be able to reverse condensation reactions. The formation of brookite from tertiary alcohols, as well as the early crystallisation of rutile in the precursors arising from the alcoholysis of TiC1, with ethanol and butan-2-01, are difficult to rationalise. In addition to the effect of the alcoholic medium, the crystallisation behaviour is most probably influ- enced by the initial ultrastr~cture~~ of the amorphous precur- sor.Thus, the unique formation of hydroxy groups in the reaction of tert-butyl alcohol with TiCl, might be related to the crystallisation of brookite. This formation may be accounted for by the occurrence of reaction (8) (liberation of ButCl) instead of reaction (9) (liberation of HCl), owing to the increased cationic character on the tertiary carbon group 1930 J. Muter. Chew., 1996, 6( 12), 1925-1932 Am'OH I 14- A. L. A. BuSOH A A,- Pr'OH A A A M c BunOH Am'OH I A.A A. BunOH I I, & I PPOH II L.. .* L. ..Ah-Fig. 6 X-Ray diffraction patterns of the different titanias prepared by alcoholysis of TiCl, after calcination (a)at 500 "C for 5 h, (b)at 950 "C without holding which favours the nucleophilic attack of chloride ETiCl+ Bu'OH+ rTiOH +ButCl (8) =TIC1 +ButOH+ -TiOBu'+ HCl (9) Moreover, note that the probable presence of remaining OH groups in the dried precursor makes it difficult to evaluate the condensation degree from the elemental analyses (as is con- firmed by the discrepancy between the calculated and exper- imental values of mass loss in Table 6) R ~...(....,..,.,....I....I..... 5 10 15 20 25 30 35 40 eldeg rees Fig.7 X-Ray diffraction patterns of the sample arising from the reaction of TiCl, with Bu'OH, after drying and heating at (a) 20, (b) 110, (c) 500, (d) 950 "C Conclusion This study highlights the capability of non-hydrolytic sol-gel methods in preparing Ti02 precursors with various ultra- structures which in turn induce various crystallisation behav- lours These methods enable us to produce a particular crystalline structure depending on the nature of the oxygen donor (mainly ether or alcohol), as do conventional sol-gel methods by controlling the hydrolysis conditions 30 Thus it is possible to delay the anatase-rutile transformation up to ca 950°C, or to obtain the rutile phase directly as well Note that brookite, which was obtained at low temperature by action of tert-butyl alcohol on titanium chloride, is very uncommon in synthetic products 33 Further work is needed to rationalise the action of organic 0x0 compounds A solution-chemistry study of a non-hydro- lytic sol-gel route to titania will be published in the future for the reactions of titanium tetrachlonde with Ti(OPr'), and Pr',O 34 References 1 J Livage, M Henry and C Sanchez, Prog Solid State Chem ,1988, 18,259 2 C Sanchez, J Llvage, M Henry and F Babonneau, J Non-Cryst Solids, 1988,100,65 3 R J P Corm, D Leclercq, P Lefevre, P H Mutin and A Vloux, J Mater Chem ,1992,2,673 4 R J P Corm, D Leclercq, P Lefevre, P H Mutin and A Vioux, J Non-Cryst Solids, 1992,146,301 5 P Arnal, R J P Corriu, D Leclercq, H Mutin and A Vioux, Better Ceramics through Chemistry VI, Muter Res SOC Symp Proc ,1994,271,339 6 R C Mehrotra and R Bohra, Metal Carboxylates, Academic Press, London, 1983 7 J L Woodhead, J Phys Colloq ,1986,47, C1-3 8 B E Yoldas, J Muter SCI , 1986,21, 1087 9 A Larbot, J P Fabre, C Guizard and L Cot, J Am Ceram SOC, 1989,72,257 10 (a) S Doeuff, M Henry, C Sanchez and J Livage, J Non-Cryst Solids, 1987, 89, 206, (b)A Leaustic, F Babonneau and J Livage, Chem Muter ,1989,1,248 11 K Terabe, K Kato, H Miyazaki, S Yamaguchi, A Imai and Y Iguchi, J Muter Sci ,1994,29,1617 12 J P Jolivet, in De la solution a 1 oxyde, InterEditions, Paris, 1994, P 98 13 E F Heald and C W Weiss, Am Mineral, 1972,57, 10 J Muter Chew , 1996,6( 12), 1925-1932 1931 14 M Ocaiia, J V Garcia-Ramos and C J Serna, J Am Ceram SOC, 26 S Lowel and J E Shields, Powder surface area and porosity, 15 1992,75,201 V T Zaspahs, W Van Praag, K Keizer and J R H Ross, J Muter Sci ,1992,27,1023 27 Chapman and Hall, London, 1991 K A Andnanov, N N Sokolov and E N Khrustaleva, Zh Obshch Khim , 1956,26,1102 16 G Ohven, G Ramis, G Busca and V S Escribano, J Muter 28 M Jansen and E Guenther, Chem Muter, 1995,7,2110 17 18 19 20 21 22 23 24 Chem , 1993,3,1239 A Mills, R H Davies and D Worsley, Chem SOC Rev, 1993,417 R C Mehrotra, J Am Chem SOC, 1954,76,2266 R A Spurr and H Myers, Anal Chem ,1957,29,760 D C Bradley, D C Hancock and W Wardlaw, J Chem SOC, 1952,2773 C Dijkgraaf and J P G Rousseau, Spectrochim Acta Part A, 1968,24,1213 P M Hamilton,R McBeth, W BekebredeandH H Sisler,J Am Chem SOC ,1953,75,2881 K C Pande and R C Mehrotra, 2 Anorg Allg Chem, 1957, 290,95 J S Jennings, W Wardlaw and W J R Way, J Chem SOC, 1936,637 29 30 31 32 33 34 A M Chippindale, A R Cowley and R I Walton, J Muter Chem , 1996,6,611 D C Hague and M J Mayo, J Am Ceram Soc, 1994,77,1957 M Inoue, H Otsu, H Kominami and T Inui, Ind Eng Chem Res ,1996,35,295, and references therein The ultrastructure level is considered to be the range of somewhe5e between molecular and submicron dimensions, I e 10-1000 A J D Mackenzie and D R Ulrich, Ultrastructure Processing of Advanced Ceramics, Wiley, New York 1988 I Keesmann, Z Anorg Allg Chem , 1966,346, 30, S Komarneni, R Roy and E Breval, J Am Ceram Soc , 1985,68, C41 P Arnal, R J P Corriu, D Leclercq, P H Mutin and A Vioux, Chem Muter, in the press 25 G A RasuwaJew, L M Bobinowa and V S Etlis, Tetrahedron, 1959,6,154 Paper 6/04130B, Received 12th June, 1996 1932 J Matev Chem, 1996,6(12), 1925-1932

ISSN:0959-9428    

DOI:10.1039/JM9960601925

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Sodalite-type Na8Mg,Si90m (OH) and Na8Mg3Si9Om (OH,CI), novel framework magnesiosilica tes John G. Thompson,* Julieanne Dougherty Alexandra Melnitchenko Charlene Lob0 and Ray L. Withers Research School of Chemistry Australian National University Canberra AC T0200 Australia Na8Mg3Si,024(0H)2 and Na,Mg3Si902,(OH,Cl)2 have been prepared by solid state reaction at ca. 700 "C in air. These new materials have the sodalite structure type with I-centred cubic unit cells of dimensions a =9.059(2) and 8.989(1)A,respectively. The magnesium and silicon atoms in the sodalite framework are disordered with the interstitial sodium ions being partially exchangeable at 85 "C. These materials represent only the second type of magnesiosilicate framework structure reported. Materials with the sodalite structure while not generally considered as belonging to the family of zeolites are nonethe- less very closely related.The building block of the sodalite structure is a cuboctahedron of corner-connected tetrahedra or P-cage (see Fig. l),which is also a component of zeolites A X and Y. For this reason and because many of the sodalites display base exchange and/or interstitial water they are treated together with zeolites.' Sodalite itself has the chemical composition Na8Al6Si6024C12 which Can be rewritten as NagC12(A16Si6O2,) to represent the composition of the cuboctahedral interstices and framework respectively. The sodalite structure is chemically highly adaptable in terms of substitution into both the framework and the cuboc- tahedral interstices.Sodalite-like structures have been reported related with compositions at or close to Na2MgSi0,23-25 or K2MgSi04,26 or tridymite-related with composition K,MgSi308 .26 The proposed magnesium analogue of the framework aluminosilicate analcime namely doranite has recently been di~credited.~~ Zeolite and molecular sieve research has only produced a handful of magnesium-substituted materials where a significant level of magnesium is demonstrably located in the tetrahedral framework. All of these involve substitution of magnesium for aluminium into aluminophosphate frameworks. They have been variously labelled as and MAPOS.~' Notably approximately one third seems to be the maximum level of substitution of magnesium for aluminium.31P MAS NMR spectroscopy has been used to confirm that magnesium occurs in the framework and not in the interstices. A similar with framework compositions (A112024),2 (Alo.03Gao~97Sil~28),3 level of substitution was (A16P6024 (A12Be2Si8024 ),5 ( 12024 [Be6(AS,P)60241,7 (Be6Ge6024),8 (Be$i6024),9'10 (Ga6Si6024)," (si12024),12313and [Zn,( P,AS~)O~~].'~ A wide variety of species can occupy the cuboctahedral interstices including halides chalcogenides and oxyanions,' which together with cations such as Li+ Na+ K+ Rb+ Ag+ Ca2+ Zn2+ Mn2+ Co2+ and Cd2+ reported for the large-pore mag- nesium-containing aluminophosphate DAF- l.31 Substitution of magnesium for aluminium in alumino-phosphate framework structures compared with aluminosilic- ate structures is considered more likely for two reasons.First a magnesiosilicate tetrahedral framework would possess a much higher negative charge than magnesiophosphate requir- provide charge balance for the negatively charged framew~rk.~ ing a very high positive charge density in the interstice for In the pure silica sodalite frameworks the interstices may be charge balance. However this should not be an absolute filled by charge neutral templates such as ethylene glyc01,~~*'~ impediment as pure aluminate sodalites e~ist.~,~~ Secondly the 1,3,5-trio~ane'~.'~and dioxolane,16 or in highly siliceous alumi- nosilicate sodalite frameworks by low charge density templates such as TMA+.17-22 Among the many reported sodalite analogues to date there have been no magnesium silicates.This is not surprising as magnesium has a strong preference for octahedral coordination in oxide structures. The only magnesium silicate structures where magnesium is tetrahedrally coordinated are cristobalite-aluminium 0 silicon Fig. 1 Schematic representation of the cuboctahedron of corner-connected tetrahedra or P-cage for sodalite Na,Al,Si,O,,Cl showing the exact Si:Al ordering. The oxygen atoms linking the aluminium and silicon atoms are omitted for clarity. metal-oxygen bond length in an idealised MgO tetrahedron is closer to that in AlO than that in SiO,; the idealised metal- oxygen bond 1:ngths are derive! from bond valFnces giving &-o= 1.949 A dAl-o= 1.757 A &-o= 1.624 A.33 There are other reasons why framework magnesiosilicates might not be prepared.We have already mentioned the crystal chemical preference of magnesium for higher coordination. Another reason is the relative insolubility of magnesium under alkaline conditions which precludes conventional preparative routes for zeolites. Therefore it is not surprising that the new materials Na8Mg,Si902,(OH) and Na,Mg3Si90,,(0H,C1) which are the subject of this paper could only be prepared by solid-state reaction. The syntheses of these new members of the sodalite family their characterisation and cation exchange properties are described below. Experimental Synthesis Magnesiosilicate analogues of both sodalite and hydroxysoda- lite have been prepared.Both solid-state and hydrothermal reaction conditions were investigated though only the solid- state processes were successful in yielding the desired products. AR grade reagents [NaNO NaOH MgCl Mg(NO,),] were reacted with various sources of silica under a range of con- ditions The sources of silica used were sodium silicate solution (Aldrich ca 14 mass% NaOH ca 27 mass% SiO,) colloidal silica (Ludox AM du Pont ca 30 mass% SiO,) talc and a commercial magnesium silicate gel (Florisil U S Silica 15 5 mass% MgO 840 mass% SiO,) Detailed descnptions of successful syntheses are given below as well as a brief summary of unsuccessful strategies Magnesiosilicate analogue of hydroxysodalite Na8Mg,Si,0zlr(OH)2To a solution of 2 82868 NaNO and 42662g Mg(NO3),*6H,O in 10ml H,O log of colloidal silica (Ludox AM) was added dropwise with stirring resulting in the formation of a gel A solution of 0448 g of NaOH in 4ml of H20 was added to the stirred gel causing the gel to thicken Stirring was continued for ca 20 min the gel was then dehydrated at 130 "C overnight The powdery product was ground divided into three portions of 2 7 g and pelleted using 10 tonnes uniaxial pressure on a 12mm die Apart from a small proportion of crystalline NaNO the material prior to pressing was XRD-amorphous.The pellets were placed in a furnace at 250"C then ramped to 650°C over a period of 2 h and held at this temperature for 1 2 and 3 days respectively XRD patterns of pulverised pellets showed that the sodalite-type material was the predomi- nant phase with crystallinity increasing with heating time A small quantity of amorphous material was also observed in the XRD patterns but the proportion decreased with longer heating A small amount of a P-cristobalite-type phase was also produced and was minimised in the sample heated for 2 days The XRD profile of this matenal which is discussed in more detail later is given as Fig 2(a).The ,'Si MAS NMR 2Bl degrees (Cu-Ka) Fig.2 X-Ray powder diffraction patterns of (a) Na,Mg,Si902,(0H)2 and (b) Na,Mg,Si,O,,(OH,Cl) whose syntheses are described in detail m the Experimental section 1934 J Muter Chern 1996 6(12) 1933-1937 spectrum and ion exchange data for this sample are also described later.Further heating of repelleted material at higher temperatures (665 "C) for short times resulted in a slight improvement in crystallinity and reduction in P-cristobalite-type phase content but at higher temperatures (2675 "C) the magnesiosilicate sodalite material began to decompose in favour of the amorph- ous material. Attempts to prepare the magnesiosilicate sodalite by solid state reaction using other reactive forms of silicate (eg gel prepared from sodium silicate solution) or magnesiosilicate (Florisil delaminated talc) were all unsuccessful though the reason for this is not understood The main crystalline products in these cases were P-cristobalite-type phases34 and sodium silicate Na,SiO Attempts to synthesise the magnesiosilicate sodalite material from colloidal silica without pelleting the reaction mixture before firing were also unsuccessful resulting in similar unwanted products Magnesiosilicate analogue of sodalite 'Na8Mg&,0&1,' To a stirred solution of 3 772 g NaNO 1 128 g MgC1 *6H,O and 2 844 g Mg(NO,) *6H,O in 8 5 ml H,O 10 g of colloidal silica solution (Ludox AM) was added slowly forming a gel A further 15 ml of H20 was used to rinse the beaker Stirring of the gel was continued for ca 20 min the gel was then dehydrated at 110 "C overnight The reaction mixture was pelleted as above and fired at 650°C for 4 days The product at this stage contained some cristobalite-related impurities which were removed by heating the specimen at 795°C for 1 h The principal product observed by XRD [Fig 2(b)] was sodium magnesiosilicate sodalite with a small impurity of NaCl and some XRD-amorphous material contributing to the background.Attempts at hydrothermal synthesis. Attempts to synthesise sodium magnesiosilicate sodalite using hydrothermal con-ditions were made using reaction conditions similar to those reported for the hydrothermal synthesis of hydroxysodalite 3s Reactions were carried out in a Teflon-lined stainless-steel autoclave fitted with an internal Teflon-coated thermocouple Various reaction mixtures were heated at 100-200 "C under ambient pressure for periods of 1-12 days (up to 14 days) In all hydrothermal preparations a five-fold excess of Na' over that required by the desired stoichiometry was used as for many standard sodalite preparations The tetrapropylam-monium cation (TMA') commonly used as a template for the formation of the sodalite structure was included in some of the hydrothermal preparations In those preparations using TMA' where temperatures above 100 "C were used glycerol was added TMA+ had been used successfully in the synthesis of MAPO-20 a partly magnesium substituted aluminophosph- ate with the sodalite-type structure.Reactions using colloidal silica and sodium silicate solution produced an amorphous product whereas those using ball milled talc and Florisil gave a poorly ordered vermiculite-like reaction product In none of these experiments were there any signs of framework magnesiosilicate formation.Characterisation Specimens were examined initially by X-ray powder diffraction using a Siemens D5000 diffractometer with Cu-Ka radiation (A= 15418 A) To obtain accurate unit-cell dimensions selected specimens were further examined using a Guinier-Hagg camera with monochromated Cu-Ka radiation (A= 1 5406 A) with Si (NBS No 640) as an internal standard. Selected area electron diffraction patterns (SAEDPs) from sodalite-type containing specimens were obtained in a JEOL 1OOCX transmission electron microscope (TEM). Finely ground specimens were dispersed onto a holey carbon grid for examination in the TEM. Close to single-phase specimens of the sodalite-type material were also studied using solid-state ,'Si NMR spectroscopy.29Si MAS NMR spectra were obtained using a Bruker MSL400 spectrometer operating at 79.488 MHz. Samples were spun at the magic angle at a frequency of 4.2 kHz in Bruker double- air bearing rotors and the spectra collected using the single- pulse excitation technique with various recycle times. The compositions of the reaction products were analysed quantitatively using energy dispersive X-ray spectroscopy (EDS) in a JEOL 6400 scanning electron microscope (SEM). Analyses were made at 15 kV and 1 nA using a Link ATW detector (138 eV resolution) and data processed using the Link ISIS system. ZAF corrections were made using the SEMQUANT software package. Ion exchange Small quantities of Na8Mg3Si902,(OH) were treated with solutions containing a 100-fold excess of Li+ ,K+,Rb' Cs+ Ag+ and T1+ at 85°C to assess the exchangeability of the interstitial Na+ ions.For K+ ,exchange was also attempted at 40 "C. Following exchange the specimens were rinsed thor- oughly with distilled water dried at ca. 100°C then analysed using EDS in the SEM. Results Synthesis Sodium magnesiosilicate sodalite (Na-MgSOD) was first observed by solid-state synthesis using colloidal silica as reagent but all attempts to synthesise the material using other sources of silica were unsuccessful. The sodalite-type material could only be prepared over a narrow range of conditions; 650-700°C with heating times of 1-4 days or 700-800°C with heating times of several hours. Higher temperatures or longer heating times led to the decomposition of the Na- MgSOD.We were not able to prepare the sodalite-type material or any other framework magnesiosilicate for that matter using hydrothermal conditions. XRD The XRD profiles (Fig. 2)t of the Na-MgSOD specimens whose synthesis is described above in detail both showed great similarity to the XRD data reported for sodalite and hydroxy- sodalite. In each case the observed peaks could be fitted to a body-cenped cubic (bcc) unit cell with a=9.059(2) and 8.989( 1) A for Fig. 2(a) and (b) respectively. The cubic dimen- sions were-slightly larger than those reported for sedalite [a = 8.8784(4) A]36 and hydroxysodalite [a =8.750(1) A].36 There was no evidence from the XRD profiles of violation of the body-centring condition h +k +1=2n.The Na-MgSOD speci- men shown in Fig. 2(a) contains two small peaks corresponding to a small impurityoof a P-cristobalite-like phase of cubic cell dimension a =7.32 A. The specimen in Fig. 2( b) was unrinsed and contained a small amount of NaCl byproduct. Inspection of the XRD profiles in Fig. 2 also reve5ls a broad background "hump centred on 28~28 (ca. 3.2 A) indicative of some remnant glassy material being present in each case though relatively less in Fig. 2(a). 7 XRD data for Na,Mg,Si,O,,(OI.f) [Fig. 2(a)] and Na,Mg,Si,O,,(OH,C1) [Fig. 2( b)] are available as supplementary data (SUP. No. 57176) from the British Library. Details are available from the Editorial Office. TEM As the sodium magnesiosilicate sodalite materials decompose rapidly in the electron beam and the crystal domain size is relatively small (< 1pm) it is not possible to obtain high quality electron diffraction patterns.The recording of conver- gent beam electron diffraction patterns to circumvent the small domain size was not possible owing to the rapid beam damage. Fig. 3 presents SAEDPs for the three principal zone axes of Na8Mg3Sig0,,(OH),~ .All diffraction patterns could be indexed to a bcc unit cell of dimension a z9.05 A. No further extinction conditions were observed consistent with both 143m and lm3m which are possible for sodalite-type structure^,^^ in agreement with the XRD data. ,'Si NMR spectroscopy Fig.4 shows the 29Si MAS NMR spectra for the Na,Mg,Si,O,,(OH) specimen shown in Fig. 2(a).The spec- trum consists of a broad composite signal centred on 6 ca. -88 (relative to Me,Si) which can be simulated by a series of Gaussian peaks as shown. The positions half-widths and relative intensities are listed in Table 1. As there are no published 29Si NMR data for framework magnesiosilicates due mainly to there being almost no reported examples it is only possible to interpret the spectra by analogy with framework aluminosilicates which display a stepwise change in chemical shift with change in c~ordination,~~.~' and which to a first approximation is independent of the type of framework. The deconvolution presented in Fig. 4 and Table 1 is pro- posed on the basis that the spectra are dominated by the various Si(nMg) (n=0-4) environments with a 6 ca.7 increase in chemical shift with each substitution compared to 6 ca. 6 for Si(nA1) in framework aluminosilicates.39~40 The assignment is made on the basis that there is a silica-rich XRD amorphous component at low frequency. Further argument to support this assignment is presented in the Discussion section. SEM/EDS Characterisation of the products of successful syntheses in the SEM revealed a glass-like mass with no obvious crystal formation. Quantitative microanalysis of the various specimens Fig. 3 Selected area electron diffraction pstterns (SAEDPs) for Na8Mg,Si,0,,(OH) along (a) (Ool) (b) (110) and (c) (111) type zone axes. Owing to the small single-crystal domain size and rapid damage in the electron beam the SAEDPs are unavoidably contami- nated with diffraction spots from neighbouring crystal domains.J. Muter. Chem. 1996,6( 12) 1933-1937 1935 -88 0h9 calculated residual ... ... ... ... .. -60 -iO -40 -1io 3 s(29sl) Fig.4 29S1 MAS NMR spectrum for Na,Mg,Si,O,,(OH) as shown in Fig 2(a) collected on a Bruker MSL400 at 79 468 MHz A sample spinning speed of 4 2 kHz was used with a recycle time of 20 s A simulated spectrum comprising five Gaussian peaks is Juxtaposed together with the residual spectrum when the simulated spectrum is subtracted from the observed Details of the simulated spectrum are presented in Table 1 Table 1 Details of simulated 29S1 MAS NMR spectrum for Na,Mg,Si,O,,(OH) in Fig 4 6 FWHM observed' (%) calculatedb (YO) Sl(4Mg) -00 04 Si(3Mg) -740 6 5 26 47 Si(2Mg) -81 5 65 21 9 21 1 Si(1Mg) -882 65 43 8 42 2 Si(0Mg) -954 65 31 7 31 6 -'Silica' -101 0 100 27 7 'Only Si(nMg) (n=0-4) signals included in summation Calculated using binomial theorem for random distnbution of Si and Mg confirmed within error the proposed magnesium silicon ratio ze 1 3 for both the sodalite and hydroxysodalite analogue specimens However the analytical volume of spot analyses (several pm3) was almost certainly greater than the crystallite size and therefore could not alone provide conclusive evidence for the composition of these new materials gven the presence of some XRD amorphous material Microanalysis of the sodalite analogue whose XRD profile is shown in Fig 2(b) after rinsing to remove the NaCl byproduct confirmed the presence of chlorine in the structure but only about half of that required for charge balance suggesting an actual composi- tion closer to Na,Mg3Si9O2,C1(OH) Owing to variability in the C1 content we represent the formula for this material as N%Mg3S19024 (C1,OH )Z Ion exchange At 40 "C Na,Mg,Si,O,,(OH) showed negligible Na+-exchange capacity in the 100-fold excess K+ solution However at 85"C partial exchange (up to cu 50%) was observed for Li' K+ Rb' Cs' Ag+ and T1+ With none of the exchange solutions was complete exchange observed In hydroxysodalite complete exchange was observed at 85°C for Li' Na+ and Ag' ,and partial exchange for K+,T1+,Rb' and Cs' Our results indicate that the interstitial sodium is less exchangeable in Na8Mg3Si902,(OH) than in hydroxysodalite Discussion The combined XRD and electron diffraction data for these new materials provide strong evidence for the formation of magnesiosilicate analogues of hydroxysodalite and sodalite given the composition of the reaction mixture The only other possibility in the absence of aluminate would be the formation of a pure silica or highly siliceous sodalite While this would be highly implausible in the absence of an organic template the unit-cebl dimensions further preclude tkis as the observed 9 05-9 00 A are much larger than the 8 83 A reported for pure silica sodalite 42 The increased unit-cell dimensions of the Na- MgSOD materials over hydroxysodalite and sodalite are quite consistent with the presence of 25% MgO tetrahedra in the sodalite framework Depmeier3 has presented the structural relationships between the various symmetries observed for members of the sodalite family The highest possible symmetry is lmh but this is rarely observed as the ideal tetrahedral framework usually distorts by means of concerted rotation of the tetra- hedra about their 4 axes The resultant space-group symmetry is I43m Ordering of the tetrahedral framework atoms for the I43m structure preserves the cubic symmetry but destroys the body centring The resultant space-group symmetry is then P43n Such lowering of symmetry would occur in Na-MgSOD if the composition of the tetrahedra alternated strictly between (Mg 5Sio 5)0 and SiO making the two sites inequivalent However our XRD and electron diffraction data demonstrate that our Na-MgSOD materials conform to bcc symmetry which requires that there is no Mg Si ordering in the frame- work Givtn that the cubic cell dimension is significantly less than 9 50 A which is the theoretical dimension for an uncol- lapsed framework of composition Mg,2sSio75 it is most probable that the correct symmetry for Na-MgSOD is 143m This is further supported by the close similarity in unit-cell dimensions and relative intensities of XRD peaks between the two Na-MgSOD materials and their aluminosilicate ana-logue~,~~37 indicative of similar degrees of collapse Having established the probable space group for Na-MgSOD we can consider further the 29S1 NMR data If we assume a random distribution of Mg and Si within the framework we can calculate the expected distribution of Si(nMg) (n=0-4) environments using the binomial theorem for MgxSil -where Si( OMg) =(1-x) Si( 1 Mg) =4x( 1-x)~ Si(2Mg)=6~~(1-~)~,Si(3Mg)=4x3(1-x) Si(4Mg)=x4 A framework composition of Mg 25Sio 75 would give a distri-bution of intensities as follows Si(OMg)=31 6% Si( 1Mg)= 42 2% Si(2Mg)=21 1% Si(3Mg)=4 7% Si(4Mg)=O4% Deconvolution of the spectrum of Na,Mg,Si,O,,(OH) shown in Fig 4 allowing for the presence of some silica-rich XRD- amorphous matenal gave relative intensities of 31 7% 43 8% 21 9% 26% and 0% if our proposed peak assignment is correct While there is always an element of subjectivity in spectrum deconvolution it would be fair to propose that the 29S1 NMR data were consistent with a random distribution of Mg and Si on the tetrahedral framework cation sites in Na8Mg3S19024(0H 12 A number of reasons were proposed in the introduction as to why framework magnesiosilicates might not be prepared Our inability to prepare Na-MgSOD or any other framework magnesiosilicates by hydrothermal methods up to 200 "C agreed with these propositions This does not preclude the possibility that Na-MgSOD might be prepared hydrothermally under more rigorous conditions Hydroxysodalite can be synthesised hydrothermally at 450-750 "C and 1000 kg cmP2 pressure 43 Also Shannon24 succeeded in growing hydrother- mally crystals of the cristobalite-related framework magnesios- ilicate Na2MgSi04 at 700 "C and 3000 atm pressure That Na-MgSOD forms at all by our solid-state synthesis is almost certainly due to the pelleting of the reaction mixture This assists in the retention of the water required in the structures of both NasMg3Sig024(OH)2 and Na8Mg3Si9024 (OH,Cl) Without pelleting the phases do not form While other sodalites can be synthesised by solid-state none of these require water in the structure We attnbute the 10 11 12 13 14 15 16 J J Glass R H Jahns and R E Stevens Am Mineral 1944 29,163 L B McCusker W M Meier and K Suzuki Zeolites 1986,6,388 D M Bibby N I Baxter D Grant-Taylor and L M Parker in Zeolite synthesis ed M L Occelli and H E Robson ACS Symp Ser 398 ACS Washington DC 1989 ch 15 p 329 D M Bibby and M P Dale Nature (London) 1985,317,157 T M Nenoff W T A Harrison T E Gier and G D Stucky J Am Chem SOC 1991,113,378 J Keijsper C J J den Ouden and M F M Post in Zeolites Facts Figures Future ed P A Jacobs and R A van Santen Elsevier The Netherlands 1989 p 237 K Futterer W Depmeier F Altorfer P Behrens and J Felsche ready decomposition of Na-MgSOD above 750-800 "C to water loss Structural water loss has been reported to occur at 860 "C in hydroxysodalite 43 17 18 Z Kristallogr 1994,209 517 R H Jarman J Chem SOC Chem Commun 1983,512 B J Schoeman J Sterte and J-E Otterstedt Zeolites 1994 14 208 19 C J J den Ouden K P Datema F Visser M Mackay and Conclusion 20 M F M Post Zeolites 1991,11,418 P D Hopkins in Molecular Sieues Vol 1 ed M L Occelli and Our lack of success in reproducing the synthesis of Na-MgSOD using synthesis strategies other than by solid-state reaction from colloidal silica reagent descnbed in the present work 21 22 H Robson Van Nostrand Reinhold New York 1992 ch 11 p129 R M Barrer and P J Denny J Chem SOC 1961,971 C Baerlocher and W M Meier Helv Chim Acta 1969,52 1853 suggests that the preparation of other magnesiosilicate zeolite analogues will be difficult The temperatures necessary to promote solid-state reaction preclude the use of organic tem- plates.This limitation requires that the framework is built around thermally stable inorganic species such as those which 23 24 25 26 C M Foris F C Zumsteg and R D Shannon J Appl Crystallogr ,1979 12,405 R D Shannon Phys Chem Miner 1979,4,139 W H Baur T Ohta and R D Shannon Acta Crystallogr Sect B 1981,37,1483 E W Roeder Am J Sci ,1951,249,224 have been used in other solid-state syntheses of sodalite-type materials inevitably restricting the possible structures to small cage frameworks with non-exchangeable positively charged moieties occupying the interstices Also it is clear that for hydrothermal synthesis to succeed it will require much higher temperatures and pressures than were available to us again 27 28 29 30 D K Teertstra and A Dyer Zeolites 1994,14,411 P J Barrie and J Klinowski J Phys Chem ,1989,93,5972 F Deng Y Yue T Xiao Y Du C Ye L An and H Wang J Phys Chem ,1995,99,6029 S T Wilson and E M Flanigen in Zeolite synthesis ed M L Occelli and H E Robson ACS Symp Ser 398 ACS Washington DC 1989 p 329 preventing the use of conventional organic templates Nevertheless our synthesis of a zeolitic magnesiosilicate frame- work demonstrates that such structures are possible and further work in this area may well reveal further zeolite analogues 31 32 33 P A Wright R H Jones S Natarajan R G Bell J Chen M B Hursthouse and J M Thomas J Chem SOC Chem Commun 1993,633 W Depmeier Phys Chem Miner 1988,15,419 N E Brese and M O'Keeffe Acta Crystallogr Sect B 1991 47,192 References 34 R L Withers C Lobo J G Thompson and S Schmid Acta Crystallogr Sect B in press R M Barrer Hydrothermal chemistry of zeolites Academic Press 35 36 R M Barrer and E A White J Chem SOC,1952,1561 JCPDS-ICDD file no 37-476 London 1982 p 46 37 JCPDS-ICDD file no 40-100.4 V I Ponomarev D M Kheiker and N V Belov Sou Phys Crystallogr ,1971,15,799 D E W Vaughan M T Melchior and A J Jacobson in Intrazeolite Chemistry ed G D Stucky and F G Dwyer ACS Symp Ser 218 ACS Washington DC 1983,ch 14 p 231 S T Wilson B M Lok C A Messina T R Cannon and E M 38 39 40 41 W Depmeier Z Kristallogr 1992 199 75 E Lippmaa M Magi A Samoson M Tarmak and G Engelhardt J Am Chem SOC ,1981,103,4992 J M Newsam J Phys Chem 1985,89,2002 R M Barrer and J D Falconer Proc R SOC London 1956 236,227 5 Flanigen J Am Chem SOC ,1982,104,1186 M Dano Acta Crystallogr 1966,20 812 42 S D Loades S W Carr D H Gay and A L Rohl J Chem SOC Chem Commun 1994,1369 6 7 C Fouassier A Levasseur J C Joubert and P Hagenmuller Z Anorg Allg Chem ,1970,375,202 T E Gier W T A Harrison and G D Stucky Angew Chem Int 43 44 I P Ivanov V F Gusynin Yu E Gorbatyi M B Epel'baum and M A Glagoleva Ocherki Fiz Khim Petrol 1970,2 50 R Kondo Yogyo Kyokai shi 1965,71 1 8 Ed Engl 1991,30,1169 0 K Mel'nikov B N Litvin and S P Fedosova in Hydrothermal 45 46 W Depmeier Krist Tech 1972,7,229 M E Brenchley and M T Weller Chem Mater 1993,5,970 Synthesis of Crystals ed A N Lobacher Consultants Bureau 47 J S Prener and R Ward J Am Chem SOC,1950,72,2780. New York 1971 9 S E Dann and M T Weller Inorg Chem ,1996,35,555 Paper 6/04946J Received 15th July 1996 J Mater Chern 1996 6(12) 1933-1937 1937

ISSN:0959-9428    

DOI:10.1039/JM9960601933

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

The crystal structure of Li,Sb04 Janet M. S. Skakle," Maria A. Castellanos R.,b Sonia Trujillo Tovar,b Susan M. Fray' and Anthony R. West" "Departmentof Chemistry, University of Aberdeen, Meston Walk, Aberdeen, UK AB24 3UE bUniversidad Nacional Autonoma de Mexico, Facultad de Quimica, Mexico DF04510,Mexico 'School of Chemistry, St. Andrews University, St. Andrews, Fife, UK K Y16 9ST The crystal structure of Li3Sb04 has been determined by analogy with Na,BiO,, and refined by the RiFtveld method using X-ray powder diffraction data. It is monoclinic, space group P2/c, a= 5.1579(2),b= 6.0922(3), c=5.1397(2) A, B= 108.839(4)". The structure is an ordered rocksalt type, with distorted M06 octahedra. The crystal structure of Li,SbO, was first examined by Blasse using X-ray powder diffraction (XRD) d?ta.' The unit cell was given as tetragonal, a=6.12, c =8.50 A; the structure was suggested to be an ordered rocksalt structure, where a=$4 and c =2aR (aR =unit cell of rocksalt).The Sb atoms were esti- mated to be located in positions ($ 3 0), (+ $), (a + 3)and (31 $) within a cubic close-packed (ccp) oxide array, giving a cation order related to that of spinel. Based on the Sb-0 framework, the agreement between observed and calculated intensities was reasonably good, with an R factor of 12%. The powder pattern of Li,SbO, given in the powder diffrac- tion file (card no. 17-824)2 was found to match that of Li2Te04 (card no. 26-8$1),, which has a similar unit cell with a=6.045 and c =8.290 A. The structure of Li2Te04 was determined in the space group P4,22, and was found to be a distorted inverse spinel with Te in octahedral sites and Li in both tetrahedral and octahedral sites., In spite of the similarity between the XRD patterns of Li,SbO, and Li2Te04, it seemed unlikely, from its stoichiometry, that Li3Sb0, could have a spinel structure.Since, however, the XRD patterns of both Li3Sb04 and Li2Te0, would be dominated by the scattering powers of Sb and Te, respectively, the similarity between the two patterns may confirm Blase's observation of a spinel-like cation arrangement. Other Li,MO, structures with similar-sized M atoms to Sb are Li,NbO, and Li,TaO,; these belong to the rocksalt family, with ordered Na,Ta04 and Na,Nb04 also have rocksalt-related structures and can exhibit both order and disorder of the cations, in different polymorphs.lO,ll The XRD pattern of Li,SbO, was examined as part of an investigation into rocksalt-related phases in the system Li,SbO,-CuO. Although the XRD data were collected on a high-resolution instrument, the pattern could not be indexed adequately on the previously described tetragonal unit cell.The structure of Li3Sb04 has therefore been reinvestigated, and we report the results here. Experimental Li3Sb0, was prepared by reaction of stoichiometric quantities of Li2C0, and Sb205 in an Au foil boat, initially at 700 "C for a few hours to expel C02 and then at 1000 "C for 24 h. XRD data for Rietveld refinement were collected with a Stoe Stadi/P diffractometer in transmission mode using a small linear position-sensitive detector (PSD) and a germanium mono-chromator providing Cu-Ka, radiation (A= 1.540 56 A).A scan range of 10<28 <110 " in steps of 1 was used in the refine- ment; the detector resolution was 0.02 ",Refinement was carried out using a squared Lorentzian function to model peak shape. This has been found to give the best fit for the peak shape profile on the Stoe diffractometer. Analysis was carried out using the Stoe software packages: initial indexing of the room- temperature powder pattern was attempted using the auto- indexing program INDEX, unit-cell refinement was performed using LATREF, theoretical powder patterns were generated using THE0 and Rietveld refinement used the pattern fitting structure refinement (PFSR) program.Differential thermal analysis (DTA) was carried out on a Stanton Redcroft STA675 instrument, using A1203 as a refer- ence material. Results The structure postulated by Blasse gave moderately good agreement with the intensities of the observed powder pattern. Hence, the structure was essentially correct, in so far as the Sb ions, which give most of the scattering, are located in a spinel- like arrangement within a ccp oxide structure. However, there are serious deficiencies in the fit of the proposed tetragonal cell to the observed d spacings for Li,SbO,. Table 1 gives the the first ten lines of the observed data together with the calculated d spacings obtained by refinement of the proposed tetragonal unit cell.It can be seen that the fit is extremely poor, and a number of lines could not be indexed in the complete pattern. Indexing of the powder pattern of Li,SbO, was therefore attempted using the auto-indexing procedure. No solutions gave a reasonable figure of merit; an orthorhombic tell of approximate dimensions a=5.985, b =6.084, c =8.362 A was found to index many of the lines, but the fit was still poor. It was felt, therefore, that lower symmetry was required to fully index the powder pattern. A survey of other compounds of the type A&O, revealed the monoclinic phase Na,BiO,, which has an ordered rocksalt structure.12 Chemically, it seems reasonable that there would be a similarity between the two structures, since Na and Bi lie one period below Li and Sb, respectively, in the Periodic Table.Na3Bi04 is moneclinic, space group P2/c, with a=5.871, b=6.696, c=5.659 A, /?= 109.8'. A theoretical powder pattern was generated for Li3SbO4 using the atomic coordinates for Na3Bi04 and was found to be similar to the observed powder pattern. The unit cell for Li3Sb04 was therefore indexed by comparison, and the first ten lines of the fully indexed pattern are given in Table 1. Comparison of the two sets of refinements in Table 1 shows an order of magnitude improvement in the fit of the monoclinic unit cell; the original tetragonal unit cell is therefore clearly incorrect. To refine the structure of Li,SbO,, 14 profile parameters were refined initially, including four cell parameters, instrumen- tal zero-point and scale factor.The variation of the full width J. Muter. Chem., 1996, 6(12), 1939-1942 1939 Table 1 The observed d spacings for Li,SbO, together with calculated d spacings for the tetragonal‘ and monoclinicb unit cells (first ten lines) tetragonal cell monoclinic cell d(obs)/A 1001/1, d(calc)/A h k I d(obs)-d(calc)/A d(calc)/A h k I d(obs)-d(ca1c) 6 0924 63 4 6 0389 1 0 0 0 0535 6 0923 0 1 0 00001 4 8817 1000 4 9024 1 0 1 -0 0207 4 8814 1 0 0 0 0003 3 8045 98 6 3 8061 1 1 1 -0 0016 3 8094 1 1 0 -0 0049* 3 8013 0 1 1 0 0032 3 4509 54 8 3 4468 1 0 2 0 0041 3 4509 -1 1 1 OoooO 2 6882 21 2 2 7007 2 1 0 -0 0125 2 6883 1 1 1 -0 0001 2 5818 31 1 2 5709 2 1 1 0 0109 2 5843 1 2 0 -00025 2 5817 0 2 1 0 0001 2 5267 16 6 2 5391 1 0 3 -0 0124 2 5269 -1 0 2 -0 0002 2 4634 11 3 2 4512 2 0 2 0 0122 2 4633 -1 2 1 0 0001 2 4405 75 -0 0107 2 4407 2 0 0 -0 0002 2 3390 16 7 2 3406 1 1 3 -0 0016 2 3397 -2 1 1 -0 0007 *broadened peak, resolve3 8092 42 2 3 8061 d into two separ 1 ate refl 1 ections i 1 n Rietveld refinement -0 0016 3 8095 1 1 0 -0 0003 3 8016 63 6 3 8013 0 1 1 0 0003 ‘Tetragonal cell a=b=6039(2), c=8 395(6)A, figure of merit=5O Monoclinic cell a=5 1578(1), b=60923(2), c=5 1397(1)A, ,!?= 108 841 (3)”, figure of ment =58 6 Figure of ment gven by l/(av A20) x (no observed reflections)/(no possible reflections), where the number of possible reflections is generated by the lattice type and crystal symmetry Note space group absences are not included and so the number of possible reflections may be an overestimate Table 2 Structural parameters for Li,SbO, after Rietveld refinement atom position X Y Z u,,, /A2 Sb 2e 0 0 1421(6) 0 25 0 0057(4) L1( 1) 2e 0 0 604( 9) 0 25 0 027(7) L1(2) 2f 05 0 87(2) 0 25 0 027( 7) W3) 2f 05 0418(8) 0 25 0 027( 7) O(1) 4g 0 221(2) 0091(2) OOOl(2) 0 010( 3) O(2) 4g 0 240( 2) 0 367( 2) 0 472(2) 0 014(3) Space group P2/c, 2=2, u=5 1578(2), b=60923(3), c=5 1397(2) A,/?=lo8 841(4)”, R,=403%, R,,=566%, R,=3 53% Table 3 Selected bond lengths for Li,SbO, bond bond length/A 234(4) x2 200(4) x2 218(1) x2 208(8) x2 223(2) x2 2 16(8) x2 255(4) x2 205(1) x2 208(3) x2 199(1) x2 201(1) x2 195(1) x2 I I 1 I I 111111 111111111 11M111111II IIIIIINIIIIIIIIIIIIRI1II I1 25 50 75 100 with atomic coordinates and a selection of bond lengths and 2Bldegrees angles in Tables 2 and 3 t Na3Bi04 has been shown to have a high-temperature cubic Fig.1 Expenmental XRD pattern with difference profile after Rietveld rocksalt form, with Na and Bi disordered over the cation refinement positions This form was stabilised by annealing at low tem- perature in flowing oxygen l3 In an attempt to transform the at half maximum was described by three Tchebychev poly- monoclinic Li,Sb04 structure to such a disordered form, nomials The background was also modelled using a series of samples were annealed at temperatures between 300 and 400 “C shifted Tchebychev polynomials up to the 5th degree Starting in oxygen, and also quenched from high temperatures (between parameters for the structural refinement were taken from the 1000 and llOO°C) into mercury No changes in the XRD ~~~~~~~~~Na3Bi04 structure l2 For the structural refinement, all Li 7 Atomic coordinates, thermal parameters, and bond lengths and atoms were constrained to have the same thermal parameter angles have been deposited at the Cambridge Crystallographic Data In the final refinement, ten positional and four thermal param- Centre (CCDC) See Information for Authors, J Muter Chem , 1996,eters were refined giving final R factors of R, =4 03%, R,, = Issue 1 Any request to the CCDC for this matenal should quote the 5 66% and RI=3 53% The final profile fit is shown in Fig 1 full literature citation and the reference number 1145/16 1940 J Muter Chem, 1996, 6(12), 1939-1942 Fig. 2 Relationship of the monoclinic unit cell of Li,SbO, to the cubic rocksalt subcell pattern were observed.Differential thermal analysis also showed no evidence for a transition up to 1200°C. Discussion Description of the Li,SbO, structure Li,Sb04 has a crystal structure derived from that of rocksalt by cation ordering. The cation ordering gives rise to a mono- clinic supercell whose volume is twice that of the pseudo-cubic subcell. The unit cells are related, as shown in Fig.2, by the transformations: umono=+I:1 2 11cutic=(J6/2)~; bmono=Ci 0 1Icubic =J~u;c,,~, =+ 1 2 1lcubic=(J6/2)~;p =109.5 O. As well as doubling the unit cell volume, the cation ordering causes a distortion from the pseudo-cubic symmetry of the subcell (e.g. umono#cmono, Table 1). The SbO, octahedra show small distortions in bond lengths (Table 3) whereas the LiO, octahedra exhibit considerable variations. Bond angles for the SbO, octahedra range from 162.0( 5) to 174.2( 5) O for the three 'ideal' 180 O angles and from 78.5(5) to 96.9( 5) O for the twelve ideal 90 angles. The Li06 octahedra show variations in bond angles from 163(2) to 174(4)O and from 75(1) to 102(1)". Projections of the crystal structure down b and a are shown in Fig.3 and 4. In Fig. 3, ccp oxide layers (open spheres) run horizontally and are interleaved by the cations in octahedral sites, with site ordering and overall, full site occupancy. The SbO, oxtahedra are not isolated from each other but link at opposite edges to form zigzag chains running parallel to csinp (Fig. 4). The structure shows considerable departures from Fig. 3 Projection of the Li,SbO, structure onto the ac plane. Large white circles represent oxygen, small white circles Sb and black circles Li. Layers of LiO, octahedra alternate with layers of mixed LiO,, SbO, octahedra. "tL c sin p Fig.4 Projection of the Li,SbO, structure down a, showing SbO, octahedral chains parallel to csinp local electroneutrality, using the simple criteria enunciated in Pauling's rule.The two types of oxygen have, as cation nearest neighbours, 4Li + 2Sb and 5Li + lSb, respectively, giving electrostatic bond strengths (ebs) for each of: 0(1):ebs=(4x-+ 2x-=2.333 ( 3 O(2): ebs= (5x-+ lx-=1.673 ( 3 Several worker~~~-l~ have developed the concept of bond valences from the electroneutrality principle of Pauling; throughout the cation-anion network, the valence of an atom is assumed to be distributed between the bonds it forms. Each bond can be assigned a bond valence, S, such that the sum of the bond valences at each atom is equal to the atomic valence. The bond valence can be calculated from the bond lengths using the simple relationship. S= exp["", "1=($)-N~ where R,, N and B are tabulated parameters." Using these equations and the bond lengths from Table 3, the bond valences and hence atomic valences were calculated; the results are given in Table4.Had there been any serious errors in the structure solution, the bond valence sum around each atom would differ greatly from its atomic valence; however, the results show the sums to be in good agreement with the oxidation state of each atom. Table 4 Bond valence sums (BVS) for the Li,SbO, structure bond bond length/A bond valence BVS 2.34(4) x2 0.113 2.00(4) x2 0.229 1.oo 2.18(1) x2 0.157 2.08(8) x2 0.194 2.23(2) x 2 0.142 1.oo 2.16(8) x2 0.164 2.55(4) x2 0.073 2.05(1) x2 0.206 0.95 2.08(3) x2 0.194 1.99(1) x2 0.784 2.01(1) x2 0.739 4.82 1.95(1) x2 0.886 J.Muter. Chem., 1996, 6(12), 1939-1942 1941 Comparison with other Li3M04 structures The phases L13M04 fall into two general structural families, depending on the size of M For smaller M (P, As, V, Cr and Mn), tetrahedral coordination of both Li and M is preferred and there are two main structure types, the so-called p and y structures l9 The p structures are ordered wurtzite superstruc- tures, with cation ordenng over one set of tetrahedral sites within a hexagonal close packed (hcp) oxide array The y structures have the cations ordered over both sets of tetrahedral sites within an oxide ion array that is somewhat distorted from hcp and may alternatively be descnbed in terms of tetragonal packing 2o 21 Most phases are polymorphic with /3 as the low- temperature form and y as the high-temperature form With increasing slze of M, the structures change completely and both Li and M occupy octahedral sites within a ccp oxide array to give a family of structures based on rocksalt Several structural variants occur and full structural details for all of them are not yet available In spite of the fact that Nb, Ta and Sb often form crystallochemically smilar phases, their phases L13M04 are significantly different from each other, whilst st;ll belonging to the rocksalt family Li3Nb04 is cubic (a= 8 405 A, 2=8)5 6, but can also be preppred as a cation-disordered, metastable form with a=4 212 A, Z= 1 by low temperature synthesis ' Li3Ta04 forms three polymorphs,22 the low- and high-temperature foFs are both monoclinic C2/c, a = 8 500, b=8 500, c=9 34,4 A,8 p= 117 05" and P2/n, a=6 018, b= 5 995, c= 12 865 A, p= 103 53 O respectively The intermediate temperature form, occurring between 900 and ca 1400OC2, has not been characterised fully8 since it has proved difficult to obtain the phase at room temperature by quenching 2o 23 It has been suggested, ho?ever,* that it is the disordered cubic We would like to thank Dr R A Howie for his assistance with bond length calculations, and we acknowledge the use of the EPSRC funded Chemical Database Service at Daresbury M A C and S T T thank UNAM for support, project number IN 101893 PAPIIT References 1 G Blasse, Z Anorg Allg Chem , 1963,32644 2 Powder Diffraction File, PDF card no 17-824, International Centre for Diffraction Data, Newtown Square, Pennsylvania, USA 3 PDF card no 26-861, in ref 2 4 F Daniel, J Moret, E Philipott and M Maurin, J Solid State Chem ,1977,22,113 5 J C Grenier and G Bassi, Bull SOC Fr Miner Crist , 1965,88,345 6 K Ukei, H Suzuki, T Shishido and T Fukuda, Acta Crystllogr Sect C, 1994,50,655 7 J C Grenier, C Martin and A Dunf, Bull SOC Fr Miner Crist, 1964,87,3 16 8 M Zocchi, M Gatti, A Santoro and R S Roth, J Solid State Chem, 1983,48,420 9 M Zocchi, M Gatti, A Santoro and R S Roth, J Solid State Chem , 1984,55,277 10 J Danet and J Galy, Bull SOC Fr Miner Crist , 1974,97, 3 11 M G Barker and D J Wood, J Chem SOC Dalton Trans, 1972,19 12 B Schwedes and R Hoppe, Z Anorg Allg Chem , 1972,393,136 13 S M Fray, PhD Thesis, University of Aberdeen 1996 14 I D Brown, in Structure and Bonding in Crystals, vol 11, ed M O'Keeffe and A Navrotsky, Academic Press, London, 1981 15 G Donnay and J D H Donnay, Acta Crystallogr Sect B, 1973, 29,1417 16 I D Brown and R D Shannon, Acta Crystallogr Sect A 1973,rocksalt form (a = 4 203 A), reported by Lapicky and Siman~v~~ and by Pfeiffer 25 The structure of Li3Sb04 reported here differs from these various Li3Nb04 and Li3Ta04 polymorphs Other L13M04 structures have also been reported Both L13uo4 and L13BlO4 havf tetragonal, ordered rocksalt struc- tures% with a z4 5, c = 8 5 A for Li3U0423 26 27 and a = 8 75, c = 4 22 A for Ll3B104 23 Li3Re04 has two polymorphs,28 the low- temperature form is monoclinic (C2/c) and is isostructural with Li,Sn03, the Sn sites being occupied by statistically disordered Re and Li to give Li, [ Re, 75Lio 25]0323 The high-temperature form is a disordered cubic rocksalt structure, isostructural with Li3OsO4 27 and Li3Ta04 24 25 Recently, the structure of Li3Ru04 was determined by energy minimisation procedures 29 Li3Ru04is, essentially, isostructu- ral with Li3Sb04 and Na3Bi04, however, one of the oxygen positions in Li3Ru04, 0(2), is shifted across a mirror plane with respect to the corresponding oxygen [0(l)] in Li3Sb04 This difference has been confirmed by Rietveld refinements of the Li3Ru04 structure using powder X-ray diffraction data, based on both the proposed Li3Ru04 structure and the Li3Sb04 structure 30 29,266 17 R Allman, Monatsh Chem ,1975,106,779 18 I D Brown, Acta Crystollagr Sect B, 1977,33, 1305 19 A R West, Z Kristullogr , 1975,111,422 20 A R West and P G Bruce, Acta Crystollagr Sect B, 1982, 38, 1891 21 W H Baur, Muter Res Bull, 1981,16,339 22 L C Martel and R S Roth, Bull Am Ceram SOC, 1981,60,376 23 G Blasse, Z Anorg Allg Chem , 1964,331,44 24 A V Lapicky and J P Simanov, cited in Struct Rep, 1953, 17, 392 25 P P Pfeiffer, PhD Thesis, Technische Hochschule, Karlsruhe 1961 26 C Miyake, H Takeuchi, H Ohya-Nishiguchi and S Imoto, Phys Status Solrdi A, 1982,74, 173 27 H Glaser, PhD Thesis, Technische Hochschule, Karlsruhe, 1961 28 R Scholder and P P Pfeiffer, Angew Chem Int Ed Engl, 1963 2,265 29 T S Bush, C R A Catlow and P D Battle, J Muter Chem, 1995, 5,1269 30 J M S Skakle, R A Howie and A R West, unpublished results Paper 6/03984G, Received 6th June, 1996 1942 J Muter Chem, 1996, 6(12), 1939-1942

ISSN:0959-9428    

DOI:10.1039/JM9960601939

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Synthesis and characterization of various MgO and related systems M. A. Aramendia," V. Borau, C. JimCnez, J. M. Marinas, A. Porras and F. J. Urbano Department of Organic Chemistry, Faculty of Sciences, University of Cbrdoba, Avda San Albert0 Magno sln, E-14004 Cbrdoba, Spain Various magnesium oxide catalysts have been prepared by thermal treatment of two different precursors: Mg(OH), and Mg5(OH),(C03),. An additional system based on MgO and doped with B203 was also prepared. The textural and acid-base properties of the catalysts were investigated. The synthesized solids were characterized from adsorption isotherms, X-ray diffraction (XRD), 'H MAS NMR, diffuse reflectance IR Fourier transform spectroscopy (DRIFT) and temperature-programmed desorption-mass spectrometry (TPD-MS) of adsorbed probe molecules (pyridine, 2,6-dimethylpyridine and carbon dioxide).The surface properties of the solids were found to be strongly influenced by the magnesium oxide preparation conditions (uiz.the precursor and calcination method used). Use of Mg(OH), as the precursor and in uucuo calcination provided highly basic solids (those with the highest proportions of strong basic sites). Various types of Lewis-acid sites were observed in the catalysts prepared in uacuo, probably as the result of the presence of Mg2+ cations of low coordination after the calcination. MgO, CaO and BaO were once regarded as catalytically inert materials, but are currently known to be highly active catalysts for certain base-catalysed reactions if properly activated.High- temperature heat treatment is required to obtain highly active catalysts. Currently, all these materials are typically basic solid catalysts, particulary MgO, which can be considered as a reference material among solid base catalysts.' Mixed oxides containing MgO were thoroughly studied re~ently.~.~The mixtures are often prepared in the hope of finding synergetic effects, i.e. to produce a material with properties surpassing those of a linear combination of its constituents. In most instances, the activity and selectivity enhancement resulting from admixing of single oxides is associ- ated with the reaction of surface defects (e.g. accessible metal cations) or the production of strong Brmsted acid sites as charge balancing cation^.^ In this sense, the addition of B203 (an acid solid) to basic MgO should give solids with different acid-base properties from their precursors. MgO is obtained mainly by thermal treatment of magnesium hydroxide or carbonate, and, more recently, by the sol-gel method.' The oxide morphology depends on the preparation conditions (pH, gelling agent, calcination rate and tempera- ture); however, the formation mechanism remains obscure, particulary as regards oxide hydroxylation.The thermal pretreatment gives rise to H,O and CO, evolution. Water evolution begins at about 400°C when Mg(OH)2 is heat-treated in uacuo. Carbon dioxide starts to evolve at a slightly higher temperature.' Basic sites appear upon heat treatment above 400"C, where the oxide surface is revealed by removal of H,O and CO,.However the stronger basic sites can only be produced by calcination in uacuo. Surface areas vary with the heat treatment conditions. Outgassing results in high surface areas relative to calcination under atmospheric conditions.' It therefore seems that in uucuo calcination leads to materials with enhanced properties: high surface area, stronger basic sites as well as greatly enhanced Lewis-basic sites. The presence of basic sites on MgO is attributed to surface 0,-ions. Magnesium oxide has a highly defective surface structure including steps, kinks, corners, etc., which provide O2-sites of low coordination.6 These differently coordinated 0,-sites must be responsible for basic sites of different strengths.The lower the coordination number of the 0,-site, the higher the strength of the basic sites.7 In this work we describe the synthesis of various magnesium oxide catalysts from two different precursors using two different calcination procedures (in a static atmosphere and in uacuo). A material based on magnesium oxide mixed with boron oxide was also prepared in the dilute binary oxide mode, i.e. it consisted of two components with the major oxide component (MgO) controlling the structure. All the catalysts were thor- oughly characterized by determining their textural properties, X-ray diffraction patterns, 'H MAS NMR spectra, diffuse reflectance IR spectra and acid-base properties in the first part of a comprehensive research program including a study of the activity of the catalyst in the dehydration/dehydrogen- ation of propan-2-01.Experimenta1 Catalyst synthesis The magnesium oxides used were prepared from two differ- ent precursors, namely: (a) MgS(OH),(CO3), * 4H@ (Merck Art. 5827), which yielded the catalysts MgO(1)AIR and MgO(1)VAC; and (b) Mg(OH), (Merck Art. 5870), which provided the solid MgO(I1)AIR. All the solids tested were obtained by calcination in a ceramic crucible, either in the air or in uucuo (hence AIR or VAC labels), by heating from room temperature to 600°C at a rate of 4°C min-' and then maintaining the final temperature for 2 h, after which the solids were allowed to cool to room temperature.Catalyst BM50 was prepared by suspending 29.0g of Mg(OH), and 0.35 g of B203 in 200 ml of distilled water and sonicating the mixture for 1h. The resulting solid was dried in a stove at 120°C for 2 h and calcined by using the above- described temperature programme. The final Mg:B ratio thus obtained was 50: 1. Thermal analysis of the precursors The thermal analysis of the precursors was performed on a Micromeritics temperature-programmed desorption-tempera- ture-programmed reaction instrument (TPD-TPR 2900) that was fitted to a VG Sensorlab quadrupole mass spectrometer from Fisons Instruments plc/VG quadrupole (East Sussex, UK) operating in the multiple ion monitoring (MIM) mode. A portion (ca. 50 mg) of precursor was placed in the middle of the reactor (1cm id, 20 cm length) and stabilized at 50 ml min-' at room temperature for 15 min.Then, the temperature was raised to 700°C at a rate of 5°C min-'. Water (m/z 18), J. Muter. Chem., 1996, 6(12), 1943-1949 1943 carbon dioxide (m/z 44) and carbon monoxide (m/z 28) were monitored in all instances Adsorption isotherms and surface areas The textural properties of the solids (specific surface area, pore volume and mean pore radius) were determined from nitrogen adsorption-desorption isotherms at liquid-nitrogen tempera- ture by using a Micromeritics ASAP-2000 instrument Surface areas were calculated by the Brunauer-Emmett-Teller (BET) method' while pore distributions were determined by the Barret-Joyner-Halenda (BJH) method' (adsorption branch, cylindrical pores open on one side only and adsorbed layer thickness calculated by the Halsey method) All samples were degassed at 350°C to 0 1 Pa prior to measurement X-Ray diffraction and 'H MAS NMR spectroscopy X-Ray diffraction patterns for the solids were recorded on a Siemens D-500 diffractometer equipped with an automatic control and data acquisition system (DACO-MP) PatterFs were run from nickel-filtered copper radiation (A= 1 5405 A) at 35 kV and 20 mA, the diffraction angle 28 being scanned at 2" min-l Room-temperature 'H MAS NMR measurements were car- ried out on a Bruker ACP 400 spectrometer Samples were heated in a nitrogen flow (50 ml min-') for 4 h at the reported temperature, then cooled in the flow and transferred to the sample holder in an environmental chamber (in nitrogen atmosphere) Spectra were acquired with a 90" pulse (5 ps) and the repetition time was 10s The rotation frequency was 3 5 kHz The chemical shifts were expressed relative to tetra- methylsilane (SiMe,) with the usual conventions Diffuse reflectance IR spectroscopic experiments Diffuse reflectance IR (DRIFT) experiments were carried out for probe molecules pre-adsorbed on the solids In these experiments, each catalyst was cleaned by passing an Ar stream at 50 ml min-l at 100°C for 30 min The solids were then saturated with the probe molecule and subsequently flushed with a stream of pure Ar (50ml min-') at the saturation temperature for 2 h in order to avoid physisorption Spectra were obtained at four different temperatures (50, 100, 200 and 300 "C) with a temperature levelling time of 20 min Full details of the procedure are described elsewhere DRIFT expenments were conducted on a Bomen MB-100 instrument with an environmental chamber from Spectra-Tech The instrument was operated at a resolution of 8 cm-' over the range 4000-400 cm-' to gather 256 scans Temperature-programmed desorption-mass spectrometry (TPD-MS) experiments TPD-MS expenments were carried out on the above-described Micromeritics TPD/TPR 2900-VG Sensorlab quadrupole mass spectrometer The optimum TPD conditions were as follows heating rate 10°C min-' and Ar flow-rate 50ml min-l The mass spectrometer, which was operated in the MIM mode, was programmed to perform 6 scans min-l The amines used as probe molecules in order to determine the acid properties of the solids were pyridine (pK =5 25) and 2,6-dimethylpyridine (pK =7 3) In a previous DRIFT study of the bands for the two amines in the region 1400-1700 cm-l, Marinas and co-workers" found pyridine to be adsorbed at Brarnsted- and Lewis-acid sites, and 2,6-dimethylpyridine to be adsorbed on the former site type only owing to the steric hindrance of its two methyl substituents The peaks used to quantify pyridine were the base peak (m/z79) and the secondary peak at m/z 52 (80% abundance) The MS peaks chosen for 2,6-dimethylpyridine were the base peak (m/z 107) and the secondary peak at m/z 66 (60% abundance) Calibration was 1944 J Mater Chem, 1996,6(12), 1943-1949 performed by injecting pulses of variable size (1-10 pl) of a 10-5-10-6 mol dm-3 amine solution in cyclohexane Carbon dioxide was the probe molecule used to determine the basic properties of the catalysts The gases used in the CO, TPD-MS experiments, CO, and 5% CO, in argon, were both supplied by Sociedad Espaiiola de Oxigeno S A (> 99 999%) Carbon dioxide was quantified by its base (m/z 44) and secondary peaks (m/z 12, 10% abundance) Calibration was performed by injecting variably sized pulses of pure CO, or 5% CO, in Ar In this way, two calibration graphs were obtained one encompassing the range 1-20 e-' by using pulses of 1-10 ml of 5% CO, in argon and the other spanning from 15 to 40 e-' with pulses of 0 5-1 ml of pure CO, Correlation coefficients better than 0 99 were always obtained Several repetitions of each experiment provided a mean error of 2% in peak areas Moreover, the activation energy for the desorption of the chemical species formed at each solid was calculated from the Kissinger equation12 where A and C are constants TPD-MS experiments were carried out at a variable heating rate (b)that shifted the peaks in the profile (Tmax)Activation energies were calculated from the slope of a plot of In (b/Tmax)us l/Tmax Prior to adsorption of any probe molecule, each catalyst was cleaned by passing an Ar stream at 110 "C at 50 ml min-l for 30 min The solids were then saturated by passing an amine-N, or C02-Ar stream (50ml min-') at 25 or 50"C, respectively Subsequently, a pure N, or Ar stream (50 ml min-l) was passed at the saturation temperature for 2 h in order to remove any physisorbed molecules Once a stable baseline was obtained, chemisorbed amine or CO, was desorbed by heating from the saturation temperature to 600°C in a programmed fashion The selected mass peaks were monitored throughout the process Each experiment used ca 100 mg of fresh catalyst Full details of the TPD-MS method and equipment are given elsewhere l3 UV-VIS experiments The acid-base properties of the materials used in this work were also studied by using a spectroscopic method based on the determination of the amount of adsorbate retained in monolayer form by the solid This determination was carried out by measuring, by UV-VIS spectroscopy, the amount of adsorbate that remained in the overlaying solution after equilibrium was reached The adsorbates used were pyridine and 2,6-dimethylpyridine in cyclohexane solutions of known concentration for determining acid sites, and benzoic acid for basic sites Full details are given elsewhere l5 Results and Discussion Thermal analysis of the precursors Fig 1 shows the thermal analysis of the precursors used in the synthesis of the solids For Mg,(OH),(CO,), .4H20 (A), water started to evolve at a low temperature (ca 200"C), with a maximum at 265°C Carbon dioxide evolution was detected in the gas phase at ca 400"C, and peaked at 442°C On the other hand, carbon monoxide was never detected in the gas phase For Mg(OH), (B), the main peak corresponds to water at ca 388°C A small carbon dioxide peak was detected at the same temperature, which suggests that magnesium hydroxide was slightly carbonated As for the hydroxycarbonate, no traces of carbon monoxide were detected As the calcination temperature used in the synthesis was [ IAl 0 II iI"1 I, \E 100 200 300 400 500 600 700d 009? 00 Ba M I L II 1 100 200 300 400 500 600 700 temperature/"C Fig.1 Thermal behaviour of the precursors Mg,(OH)2(C03)4 -4H,O (A) and Mg(OH)2 (B). Monitored masses correspond to water (m/z 18), carbon dioxide (m/z 44) and carbon monoxide (m/z 28). 600"C, we can expect the final solids to consist primarily of MgO, without any traces of the precursors, as confirmed by XRD measurements (see below).Adsorption isotherms and surface areas Nitrogen adsorption-desorption isotherms obtained for all solids are shown in Fig. 2. All exhibit closed hysteresis loops. The isotherms are type IV in the Brunauer, Demming, 150 150 -100100 -50 -50 r Lh I-0 0v) a, 0.0 0.5 1.o 0.0 0.5 1.o w 5 \ 43 225 ba225 150 150 75 75 0 0 0.0 0.5 1.0 0.0 0.5 1.o PIP, Fig. 2 Nitrogen adsorption-desorption isotherms for the solids at liquid-nitrogen temperature Demming and Teller (BDDT) classification for mesoporous solids.I6 The surface area, pore volume and mean pore radii for the solids are given in Table 1.The catalysts prepared from magnesium hydroxide had a higher surface area (almost double) than those obtained from magnesium hydroxycarbon- ate. That for solid BM50 was slightly lower than that for MgO(II)AIR, probably owing to the presence of a small amount of B,03. On the other hand, the solid calcined in uucuo had a higher surface area than that calcined in air, consistent with reported re~u1ts.l~ X-Ray diffraction (XRD) Fig. 3 shows the X-ray diffraction patterns for the four solids studied in this work. They are all similar and exhibit three characteristic peaks for MgO (periclase variety) at d =2.44, 2.11 and 1.49 nm (20= 36.7, 42.8 and 62.2", respectively). The pattern for MgO(I1)AIR included a small peak at d= 3.05 nm (28=29.3") that was identified as a different variety of MgO that can be obtained by dehydration of Mg(OH)2, and gives mixed periclase and spinel-type patterns.No peaks from the Mg(OH)2 diffraction pattern were detected. Acid-base properties The acid-base properties of the catalysts studied were determined by two different procedures, viz. temper-ature-programmed desorption-mass spectrometry of probe molecules adsorbed on the solids, and UV-VIS spectroscopy. The probe molecules used to determine acid sites were pyridine (py, pKb =5.25) and 2,6-dimethylpyridine (dmpy, pKb=7.3). Dmpy is known to be selectively adsorbed on Brrnsted-but not on Lewis-acid sites because of steric hindrance by two methyl groups; on the other hand, sterically unhindered py is adsorbed on both Brrnsted- and Lewis-acid sites.The difference between both profiles (py -dmpy) accounts Table 1 Textural properties of MgO-based catalysts. Specific surface area (SBET),pore volume (V,) and average pore radius (r,) for the solids obtained in this work catalyst precursor sBET/m2g-' b/ml g-' r,/A ~ MgO (1)AIR Mgo(l)VAC MgO(I1)AIR BM50 Mg, (OH )2(C03 )4 Mg5(oH)2(Co3)4 Mg(OH), Mg(oH kB203 60 69 119 104 0.22 0.22 0.39 0.34 148 118 94 100 BM50 * A t I= ,MgO(I)AIR \F 21 p! .--lu-MgO(1)VAC e 10 20 30 40 50 60 70 2eldegrees Fig. 3 X-Ray diffraction patterns for the solids J. Muter. Chern., 1996, 6(12), 1943-1949 1945 for Lewis sites Since the catalytic properties of Brmsted-acid sites may well be different from those of Lewis-acid sites, it is important to understand the factors that determine the strength and type of sites in metal oxide catalysts Py and dmpy TPD-MS profiles were obtained for all the the catalysts, the results are shown in Fig 4 and 5, respectively, and summarized in Table 2 Pyridine TPD-MS profiles (Fig 4) were similar for all the solids they included two desorption peaks at low temperatures (103 and 145°C) except for MgO( II)AIR, which exhibited a single, broad peak (T,,, = ~II -1.. 1 . . . 1 . . _I, _-., . . . 0 100 200 300 400 500 600 temperaturePC Fig. 4 Temperature-programmed desorption-mass spectrometry of pre-adsorbed pyndine over the catalysts studied in this work 0 100 200 300 400 500 600 ternperature/*C Fig.5 Temperature-programmed desorption-mass spectrometry of pre-adsorbed 2,6-dimethylpyndine over the catalysts studied in this work 122°C) In addition, the catalyst calcined zn uacuo, MgO(I)VAC, exhibited an additional desorption peak at a high temperature (Tmax=381 "C) that was not observed in the other catalysts As far as dmpy TPD-MS profiles are concerned, all catalysts had broad desorption peaks at low temperatures Thus, the solids prepared from magnesium hydroxycarbonate exhibited a desorption peak at 72 "C, while those prepared from magnesium hydroxide exhibited a desorption peak at 101"C No high-temperature desorption peak was detected for MgO(I)VAC, however The results obtained by integrating the areas under the desorption peaks lead to the Brsnsted, Lewis and overall acidity values for the solids given in Table 2, which also includes the overall acidity as calculated from UV-VIS spectra by titration of acid sites with py The results provided by the two methods are seemingly inconsistent However, because all the catalysts exhibit predominantly basic rather than acid properties, the differences can be considered negligible Moreover, since B203 is an acid solid, the results provided by the TPD-MS technique, which gives a slightly higher total acidity for BM50 than for the other solids, seem more reason- able Finally, recording of DRIFT spectra for adsorbed amines was attempted However, because of the low acidity of these catalysts, the results were rather poor, so their DRIFT spectra were excluded The most salient result of this study was the py desorption peak for MgO( 1)VAC at high temperature (Fig 4), which was absent from the corresponding dmpy TPD-MS profile (Fig 5) This peak is related to Lewis-acid sites of high strength It has been reported that, on a thoroughly degassed MgO surface, coordinatively unsaturated Mg2+ cations are exposed18 which exhibit Lewis-acidic character l9 Such Mg2+ cations of low coordination may be responsible for the new Lewis-acid sites present after calcination zn uacuo to produce MgO(1)VAC Base properties were determined, by using C02 (the seem- ingly most suitable choice on account of its acidic nature) as the probe molecule in TPD-MS experiments In addition, a UV-VIS spectroscopic method" was also used to determine the basicity of the solids using benzoic acid as the titrant In studying adsorbed carbon dioxide, carbonate species of different types such as unidentate carbonate, bidentate carbon- ate, carbonate ions and hydrogen carbonates were found to be formed depending on the adsorption conditions and surface structure For unidentate carbonate, the adsorbed species was bound to the surface through a bond between the C in C02 and surface O2 For bidentate carbonate, the adsorbed species was bound to the surface via two bonds one between the C in CO, and surface 02-and the other between an 0 in C02 and a surface Mg2+ ion The presence of hydrogen carbonates suggest that hydroxy groups on MgO also act as bases toward C02 In the TPD-MS experiments with adsorbed COz, the con- centration of basic sites was reflected in the peak area of the TPD-MS profile, and their strength in the temperature at which the CO, desorption peak appeared l4 Fig 6 shows the COz TPD-MS profiles obtained for all the catalysts As can be seen, up to three different peaks were obtained two small Table 2 Determination of the acidity of the catalysts obtained in this work by temperature-programmed desorptlon-mass spectrometry (TPD-MS) and UV-VIS spectroscopy total acidity Bransted acidity Lewis acidity PYIPOl g catalyst dmPY/Pmol g (TPD-MS) PY -dmPY/Pmol g(TPD-MS) uv-VISTPD-MS MgO(1)AIR MgO( 1I)AIR MgO(1)VAC BM50 25 30 27 28 4 5 4 8 29 47 35 37 31 46 36 28 1946 J Mater Chem ,1996, 6(12), 1943-1949 0.1 0.0 MgO( 1)VAC MgO( 1)AIR 1 I 1 I I I 100 200 300 400 500 600 700 temperature/"(= Fig.6 Temperature-programmed desorption-mass spectrometry of pre-adsorbed carbon dioxide over the catalysts studied in this work peaks at low and high temperatures (Tma,=139 and 598"C), and the main peak at about 355"C, shifted to a higher temperature in the MgO(1)VAC profile.The desorption peak detected at high temperature (T,,, =598 "C) was only present in those solids prepared from magnesium hydroxide [MgO( 1I)AIR and BM50l. Therefore, magnesium hydroxide yields solids with stronger basic sites. The appearance of three regions in the TPD-MS profile suggests the presence of three different types of adsorption sites differing in their adsorption strength for C02.Similar results were reported previously.20,21 Thus, Tsuji et aL2' ana-lysed the isotopic distribution of CO, desorbed from an MgO surface containing adsorbed C1*O, and assigned the peak at the lower temperature to bidentate carbonate. They also suggested that, for peaks 2 and 3, processes other than simple adsorption-desorption of CO, on one pair of Mg2+ 0,-sites were involved and concluded that the fraction of C02adsorbed as unidentate carbonate was rather small. Table 3 shows the results obtained by integration of the area under each carbon dioxide desorption peak, as well as the total basicity as determined by CO, TPD-MS and by UV-VIS spectroscopy by using benzoic acid as the titrant.The results Table 3 Determination of the basicity of the catalysts obtained in this work by C02temperature-programmed desorption-mass spectrometry (TPD-MS). Comparison with results provided by UV-VIS spec-troscopy using benzoic acid as the probe molecule TPD-MS/pmol g-' UV-VIS/pmol g -catalyst peak 1 peak 2 peak 3 total total MgO(1)AIR 31 169 -200 156 MgO(1)VAC 25 436 -436 407 MgO(I1)AIR 135 326 45 506 495 BM50 25 354 61 440 42 1 provided by both methods are quite consistent. The solids prepared from magnesium hydroxide are more basic than those prepared from magnesium hydroxycarbonate. In this sense, MgO(I1)AIR was found to have 2.5 times more basic sites than MgO( 1)AIR (its analogue reprepared from magnesium hydroxycarbonate).The presence of B,O, in BMSO made the final material slightly less basic than MgO(I1)AIR. Finally, calcination in uacuo produced a material with enhanced basicity as revealed by a comparison between MgO(1)AIR and MgO(1)VAC: the solid prepared in uacuo was 2.1 times more basic than that calcined in the air. Table 4 shows the CO, desorption activation energies obtained from the Kissinger equation.', The activation energy can be related to the basic strength of the site type.14 The results are quite revealing: once again, the effect of in uucuo calcination is clearly apparent when the activation energies obtained for the solid MgO(1)AIR are compared with those for MgO(1)VAC.Thus, MgO(1)AIR was not only the solid containing the fewest basic sites, but also the weakest (Ea=4 and 14 kJ mol-' for peaks 1 and 2, respectively). In uacuo calcination produces a larger number of stronger basic sites (Ea=45 kJ mol-I). However, the precursor also has a clear effect that is apparent on comparing the results obtained for MgO( 1)AIR (magnesium hydroxycarbonate as precursor) and MgO( 1I)AIR (magnesium hydroxide as precursor). Solids pre- pared from magnesium hydroxide exhibited a larger number of basic sites that were also of a higher strength (Table 4). It is very interesting to note that catalyst BM50 had the strongest basic sites (Ea=108 kJ mol-I). In addition to the C02 TPD-MS profile, DRIFT spectra were recorded for carbon dioxide pre-adsorbed on all samples.Fig. 7 and 8 show the pre-adsorbed C02 DRIFT spectra for MgO(1)AIR and MgO(I1)AIR respectively, in the range 2000-1000 cm-l. Adsorbed C02 gave strong bands in the region 1400-1550cm-', with a shoulder at an even higher I I 1800 1600 1400 1200 wavenumbedcm-' Fig. 7 DRIFT spectra of pre-adsorbed carbon dioxide over MgO( 1)AIR at variable temperatures Table 4 C02 desorption activation energies (E,) for the catalysts studied as determined from the Kissinger equation, and correlation coefficients (r) peak 1 peak 2 peak 3 catalyst E,/kJmol- r E,/kJ mol - r E, /k J mol - Y MgO( 1)AIR MgO( 1)VAC 4 45 0.985 0.997 14 - 0.995 - -- -- MgO( 1I)AIR BM50 34 26 0.997 0.987 41 39 0.999 0.970 51 108 0.989 0.945 J.Mater. Chem., 1996, 6(12), 1943-1949 1947 h gE cBE cE c 1800 1600 1400 1200 wavenurnberkm-' Fig. 8 DRIFT spectra of pre-adsorbed carbon dioxide over MgO( 1I)AIR at variable temperatures wavenumber Weak bands at ca 1070 and 1220cm-' were also observed The DRIFT spectra for pre-adsorbed C02 on solids prepared from the same precursor were similar However, the materials made from different precursors also gave different DRIFT spectra While some bands (Al, A2, B2, A3 and B3) appeared in both spectra, others only appeared in the MgO(I1)AIR DRIFT spectra (B1 and E2) Table 5 shows all possible interactions between CO, and the MgO surface, and assigned IR bands in the DRIFT spectra Bands between 1400 and 1550 cm-' and that appearing at ca 1070 cm-' are ascribed to surface unidentate carbonates According to Philipp et al, unidentate carbonate seems to be the dominant adsorption state of CO, at room temperature," which is consistent with our DRIFT spectra (Fig 7 and 8) The band shoulder at 1630-1660 cm-' is assigned to surface hydrogencarbonate 600°C 300°C 120°C I I I I I I I , 20 10 0 -10 6 Fig.9 'H MAS NMR spectra of the solid MgO(I1)AIR after heating for 4h at 120, 300,600 and 900 "C The strong bands in the DRIFT spectra (Fig 7 and 8) corresponding to unidentate carbonate indicate that, under our conditions, the main adsorption mode of CO, on MgO is unidentate carbonate, as reported previously by Philipp et a1 21 On the other hand, TSUJ~ et a1 ,20 based on TPD experiments, found not unidentate, but bidentate carbonate to be adsorbed on MgO However, most of the C02 desorbed in their experi- ments was in a low-temperature peak, while most of that in our experiments was desorbed over the second peak at moder- ate temperature (peak maximum at ca 350 "C)In the adsorbed state, the symmetry of adsorbed carbonate species is lowered and those species formed generally present two v(C0) bands to either side of 1415 cm-' (corresponding to free carbonate) It has been considered that the Av splitting characterizes the structures of the species formed ca 100, 300 and 400 cm-' for unidentate, bidentate and bridged species, respectively 22 23 Table 5 C02 adsorption sites and species formed over magnesium oxides Vibration modes and wavenumbers ~~ adsorption mode unidentate carbonate hydrogencarbonate bridged carbonate ?I1 bidentate carbonate IR bands A1 (C-0,, stretching) A2 (0,-C-0, sym stretching) A3 (OI-C-O, asym stretching) B1 (COH bending) B2 (01-C-O1 sym stretching) B3 (0,-C-0, asym stretching) C1 (C=O,, stretching) El (C=OlI stretching) E2 (O,-C-OI asym stretching) cm 1075 1393 1526 1220 1419 1650 1776 1625 1300 1948 J Muter Chem, 1996, 6(12), 1943-1949 However, there is some confusion in the literature about such a relation By comparing X-ray structural and IR spectral data, it was shown that bidentate carbonate species generally have Av >250 cm- and bridged carbonate species have Av <250 cm-', the only unidentate carbonate structurally characterized has Av =80 cm-' 24 The Av found for our solids (Fig 7 and 8) agree with values reported in the literature corresponding to unidentate carbonate species [Av =70 and 140 cm-' for MgO(1)AIR and MgO(II)AIR, respectively] In addition to unidentate carbonate, hydrogencarbonate bands were found in both solids, the clearer in the MgO(I1)AIR DRIFT spectra at 1220 cm-' The formation of hydrogen- carbonate species is related to the presence of OH sites acting as bases against C02 'H MAS NMR spectra recorded after heating the catalyst MgO( 1I)AIR at several temperatures for 4 h (Fig 9) show that even after heating at 900°C the band due to the surface OH is still present The presence of these OH sites may be crucial in some types of catalysts, as they can enhance metal-support interaction^^^ or facilitate some types of catalysed organic reactions such as the aldol addition of acetone 26 As far as BM50 is concerned, we have no experimental evidence of the physical state of boron in the solid However, solids BM50 and MgO(I1)AIR show different reactivity and selectivity patterns in propan-2-01 decomposition 27 Derouane et a1 and McKenzie et a1 ,from 27Al NMR studies of A120, doped Mg028 29 suggest that, although aluminium is probably well distributed throughout the lattice, calcination produces a concentration of A1 at the surface, however, this A13+ ion is in a tetrahedrally coordinated oxidic environment which is unfavourable for acid-catalysed reactions 28 29 Additional experiments are being carried out to reveal the role of boron in this catalyst Conclusions The above results allow one to draw some interesting con- clusions (2) The precursor used to prepare an MgO catalyst is crucial in order to obtain the desired textural and acid-base characteristics Thus, solids prepared from Mg(OH), exhibited higher BET surface areas and greater basicity than those obtained from magnesium hydroxycarbonate In addition, basic sites involving OH species are present in both solids even after calcination at 900°C This type of basic site is crucial for some types of catalysed organic reactions (zz) For the same precursors, the calcination method also influences the production of the desired solid Thus, zn uacuo calcination leads to solids with a slightly higher surface area and greatly enhanced basicity It also gives rise to a new type of Lewis-acid site that probably arises from low coordinated Mg2+ (m)Finally, the catalyst doped with B203 (BM50) exhibits a lower surface area than its related system MgO(II)AIR, as well as a slightly lower basicity due to the primarily acid nature of pure B20, The authors gratefully acknowledge funding of this research by the Consejeria de Educacibn y Ciencia de la Junta de Andalucia and the Direccion General de Investigacibn Cientifica y Tecnica (DGICyT) in the framework of project PB92-0816 The staff at the Mass Spectrometry and Nuclear Magnetic Resonance Services and Inorganic Department of the University of Cordoba are also acknowledged for their kind technical assistance in the experiments References 1 K Tanabe, M Misono, Y Ono and H Hattori, New Solid Acids and Bases, Elsevier, Amsterdam, 1989, vol 51 2 A M Youssef, L B Khalil and B S Girgis, Appl Catal, 1992, 81, 1 M Ueshima and Y Shimasaki, Chem Lett, 1992,1345 G Connell and J A Dumesic, J Catal, 1987, 105,285 T Lopez, I Garcia-Cruz and R Gomez, J Catal, 1991,127,75 V R Choudhary and M Y Pandit, Appl Catal, 1991,71,265 X D Peng and M A Barteau, Langmuir, 1991,7,1426 S Brunauer, P H Emmett and E J Teller, J Am Chem SOC, 1938,60,309 9 E P Barrett, L S Joyner and P P Halenda, J Am Chem Soc, 1951,73,373 10 F M Bautista, J M Campelo, A Garcia, D Luna, J M Marinas, A A Romero, J A Navio and M Macias, J Catal, 1994,145,107 11 F M Bautista, J M Campelo, A Garcia, D Luna, J M Marinas and M R Urbano, J Muter Chem ,1994,4,311 12 H E Kissinger, Anal Chem ,1957,29, 1702 13 M A Aramendia,V Borau, C Jimenez, J M Marinas, A Porras and F J Urbano, Rapid Commun Mass Spectrom ,1994,8,599 14 M A Aramendia, V Borau, C Jimenez, F Lafont, J M Marinas, A Porras and F J Urbano, Rapid Commun Mass Spectrom ,1995, 9,193 15 J M Campelo, A Garcia, J M Gutierrez, D Luna and J M Mamas, Can J Chem, 1983,61,544 16 S Brunauer, L S Demming, W E Demming and E Teller, J Am Chem SOC,1940,62,1723 17 T Izuka, H Hattori, Y Ohno, J Sohma and K Tanabe, J Catal , 1971,22,130 18 S Coluccia, Proc Symp Adsorption and Catalysis on Oxide Surfaces, London, 1984, ed M Che and G C Bond, Elsevier, Amsterdam, 1985,p 59 19 A Zecchina and F S Stone, J Catal, 1986,101,227 20 H Tsuji, T Shishido, A Okamura, Y Gao, H Hattori and H Kita, J Chem SOC, Faraday Trans, 1994,90,803 21 R Philipp, K Omata, A Aoki and K Fujimoto, J Catal, 1992, 134,422 22 J C Lavalley, Catal Today, 1996,27, 377 23 G Busca and V Lorenzelli, Muter Chem ,1982,7,89 24 A M Greenaway,T P Dasgupta, K C Koshyand G G Sadler, Spectrochim Acta, Part A, 1986,42,949 25 T Lopez, I Garcia-Cruz and R Gomez, Muter Chem Phys ,1994, 36,222 26 G Zhang, H Hattori and K Tanabe, Appl Catal, 1988,36,189 27 M A Aramendia, V Borau, C Jimenez, J M Marinas, A Porras and F J Urbano, J Catal ,1996,161,829 28 E G Derouane, V Jullien-Lardot, R J Davis, N Blom and P E Hojlund-Nielsen, in New Frontiers in Catalysis, Budapest, 1992,vol B,p 1031 29 A McKenzie, C T Fishel and R J Davis, J Catal, 1992, 138, 547 Paper 6/05425K, Received 2nd August, 1996 J Muter Chem ,1996, 6(12), 1943-1949 1949

ISSN:0959-9428    

DOI:10.1039/JM9960601943

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

"0NMR investigation of hafnia and ternary hafnium oxides Timothy J. Bastow,"Mark E. Smith*".band Harold J. Whitfield' "CSIRODivision of Materials Science and Technology, Private Bag 33, Rosebank MDC Clayton, Victoria 31 69, Australia bDepartment of Physics, University of Kent, Canterbury, Kent, UK CT2 7NR 'Department of Applied Physics, Royal Melbourne Institute of Technology, Box 2476W Melbourne, Victoria 3001, Australia Structural evolution in alkoxide-derived amorphous HfO, gels is followed by I7Omagic angle spinning (MAS) NMR in I7O-enriched samples, with crystallisation being identified readily at 500 "C. A series of I70-enriched alkali-metal and alkaline-earth- metal hafnates with sodium chloride- and perovskite-related structures are also examined by I7OMAS NMR spectroscopy.The number of oxygen resonances is related to the crystal structure. Comparison of I7ONMR data is made to isostructural zirconium- containing compounds with the isotropic chemical shifts (Hf/Zr) in the ratio 0.86 k0.02. I7ONMR data are also presented from Hf0,-SiO, and Hf0,-GeO, gel mixtures. Oxygen-17 is a very significant probe nucleus for solid-state NMR of oxides since it exhibits a large isotropic chemical shift range and a relatively small nuclear electric quadrupole moment. Initial studies demonstrated its large chemical shift range and showed that in ionic materials the static linewidths were narrow enough for magic angle spinning (MAS) of 3 kHz to be effective.'q2 The major drawback to the widespread use of 170NMR is its low natural abundance (0.037%) which usually necessitates isotopic enrichment for ready observation.I7O has been used mostly in specific applications on enriched samples such as examination of ceramic materials3 and the different layers in high-temperature superc~nductors.~~~ One of the most important attributes of NMR as a structural probe is its short-range nature, which means that investigation of amorphous materials is possible where conventional charac- terisation techniques are often limited in the information they can provide. Sol-gel produced materials are of great tech- nological interest, allowing the production of materials at relatively low temperatures (that would otherwise require application of very high temperatures, e.g.glassy Zr02-Si02 mixtures) and atomically mixed compounds. Structural evol- ution in a gel occurs uia a series of distinct amorphous states that differ in their atomic arrangement and it is these states that NMR can probe effectively. The properties of materials can differ widely depending on the local bonding arrangement that occurs and the behaviour can be determined at an early stage of their formation. Recently a series of sol-gel produced simple oxides including ZTO~,~Ti026 and La203,7 have been studied by I7O NMR spectroscopy. Hydrolysis of alkoxide precursors with I7O-enriched water is an extremely effective route for incorporation of the 170label into oxide gels. The detailed route to oxide formation varies widely with amorphous mixtures of tetragonal and monoclinic precursor oxides forming in ZrO,, mixtures of OTi3 and OTi, coordinations in TiO, gels and La203 forming from LaO(0H).The structural chemistry of each oxide is thus very different and needs to be studied separately in detail. In binary mixtures the variety of behaviour increases. I7O effec-tively monitors both metals since it is bonded to all such nuclei. In metal silicate gels, I7O NMR spectroscopy has been shown to be a much less ambiguous structural probe than 29Si NMR, with phase separation and atomic-scale mixing readily detectable in Ti02-SiO, Direct crystallisation of TiO, gels into rutile is made possible by the incorporation of small amounts of tin in the original gel and ''0NMR showed that the room-temperature gel had a different structure to pure Ti02 gels as the oxygen was already rutile-like in the tin- doped case, as indicated by the I7O chemical shift." Similarly the behaviour of ZrO, dissolved in an SiO, matrix is very different to TiO, with NMR showing that almost complete solid solution of zirconium in SiO, can occur." Hafnium is a very similar element in its chemical and physical properties to zirconium, with hafnia (HfO,) and zirconia (ZrO,) showing complete solid solution, consistent with their similar atomic radii.Both oxides are stable in the monoclinic form at room temperature, but transform on heating successively to tetragonal and cubic forms. In the present work 170-enriched preparations of HfOz gels, ternary hafnates and Hf02-SiO, are examined by I7O MAS NMR and a comparison is made to the previous work on zirconium analogues.The work is extended to Hf0,-GeO, gels. The sensitivity of I7O NMR as an atomic scale probe in such materials is examined. In crystalline structures of ionic com- pounds, such as perovskite- and NaC1-structured hafnates, I7O MAS NMR spectra can show high resolution. This observation is a consequence of the small quadrupole interaction (C,= e2qQ/h, where eQ is the nuclear electric quadrupole moment and eq is the maximum component of the electric field gradient tensor) in these ionic materials. The resonances become broader in the less ionic silicon- and germanium-containing compounds. Experimental Hafnia gel was prepared by dissolving hafnium isopropoxide, 2 mmol Hf(OPr'), *PrOH, in a mixed solvent of 20 cm3 toluene and 8 cm3 isopropyl alcohol, and then adding dropwise 0.5 ml of 10 atom% 170-enriched water while stirring continuously.This produced a sol which gelled rapidly and the gel was aged at room temperature for 24 h. The excess solvent was slowly evaporated at room temperature under a flow of dry nitrogen. The dry gel was then heated successively at 100, 300, 500 and 800°C for 2 h periods. The hafnia gel heated in nitrogen at 500°C and then at 800°C gave an X-ray powder diffraction (XRPD) pattern in agreement with monoclinic baddelyite-type HfO, (ASTM-JCPDS no. 34-104). '70-enriched Li,HfO, and Na2Hf03 were prepared by solid- state reaction of lithium and sodium carbonate, respectively, with I70-enriched hafnia gel by heating at 800 and 950°C J.Muter. Chem., 1996, 6(12), 1951-1955 1951 respectively for 6 h under a dry nitrogen atmosphere The correct stoichiometry of the final product was ensured by using amounts of Li2C03 or Na,CO, that correspond to the amount of hafnium initially added as Hf(OPr'),.PrOH The XRPD pattern of the Li,HfO, product was in excellent agreement with it being the correct single-phase product (ASTM-JCPDS no 23-1183) The XRPD peaks of Na,HfO, had a similar intensity distribution to those of Na,ZrO, (ASTM-JCPDS no 25-770) but with slightly shifted positions in accord with a slightly larger unit cell for the hafnate The alkaline-earth-metal hafnates were prepared by the metathetical reaction of Li,HfO, with an excess of a melt of a eutectic mixture of NaCl and the appropriate Group 2 chloride at 800 "C for 4 h in a nitrogen atmosphere The eutectic mixture was used because the minimum temperature that causes reac- tion is desirable as it minimises exchange of 170with H,O, COz or 0, even in a nominally inert atmosphere, and with the reaction vessel Thus 2 mmol dm-3 Li,HfO, was heated with a melt of 9 2 mmol dmP3 SrCl, and 9 4 mmol dm-, NaCl The melt was cooled and the unreacted halide extracted by dissolution in water The dried product (94% of theoretical yield) gave an XRPD pattern with very sharp peaks in accord with crystalline SrHfO, (ASTM-JCPDS no 44-991) CaHfO, and BaHfO, prepared similarly gave powder patterns in accord with ASTM-JCPDS no 36-1473 and 24-102 respectively A gel with stoichiometry HfSiO, was prepared by dissolving 2 29 mmol dm-, Hf(OPr'), *PrOH in 20 ml toluene and then adding an equimolar quantity of Si(OEt), dissolved in 8 ml isopropyl alcohol 170-enriched water (1cm3) was then added with continuous stirring A gel formed which was allowed to dry at room temperature under a stream of nitrogen and then heated for successive 2 h periods at 500 and 850°C The mass of the final product was 0 551 g, which is 93% of the theoretical yield of HfS104 XRPD studies of the product showed no evidence of crystallisation of the product Selected area electron diffraction (SAED) gave a nng pattern consistent with short- range order but no crystallisation To prepare hafnium germanate, HfGeO,, 8 mmol dmP3 Hf (OPr'), -PrOH was dissolved in toluene and an equimolar amount of Ge(OEt), in isopropyl alcohol was added 170-enriched H20 (1 cm3) was then added stirring continuously which produced a gel which was allowed to dry at room temperature under a stream of dry nitrogen The dried gel was then heated stepwise to 100, 300, 500 and 950°C XRPD studies of the product formed by heating at 950°C were in excellent agreement with crystalline HfGeO, (ASTM-JCPDS no 34-408) The X-ray peaks were not as sharp as those from the alkaline-earth-metal hafnates The XRPD of the product formed at 500 "C showed no evidence of long-range ordering X-Ray powder diffraction of these samples was performed on a Siemens Type F diffractometer using Ni-filtered Cu radiation Selected area electron diffraction patterns were obtained in a JEOL 2010 electron microscope 170MAS NMR spectroscopy was performed at 54 24 MHz on a Bruker MSL 400 spectrometer equipped with a 94T magnet The spectra from HfO, gel and binary crystalline hafnates were accumulated using a 7 mm double-bearing (DB) MAS NMR probe spinning at 4-5 kHz using 15 ps (ca 30" tip angle) pulses and a 5-10 s recycle delay For HfO, mixed with SiO, and GeO, a 4mm DB-MAS NMR probe spinning at >8 kHz was employed with the other acquisition parameters the same as for the 7 mm probe For the 7 mm probe a Si3N4 rotor was used but for the 4 rnm probe only ZrO, rotors were available and these give a 170background signal at 6 ca 385 which can be seen in Fig 3(b) and 4 (later) where the 4mm probe was used The rotor peak can be identified readily and does not interfere with the spectrum so that spectral subtraction was unnecessary There is little known about 170 relaxation times in such solids but conditions are chosen that practically allow an adequate signal-to-noise to be collected in reasonable time, which is often a trade-off between allowing complete relaxation and adding sufficient scans For BaHfOJ the I7O was determined as 32s which means some spectra observed here are partially saturated 29S1 MAS NMR spectra were accumulated by spinning at 4-5 kHz in a 7 mm DB-MAS probe using a 1 5 ps (ca 30") pulse and a 15 s recycle time at an operating frequency of 79 4 MHz Spectra were referenced externally to H20 (6 0, 170)and SiMe, (6 0, 29S1) Results and Discussion Hafnia The 170MAS NMR results from the enriched HfO, gels are shown in Fig 1 and summarised in Table 1 The gel at room temperature is characterised by two peaks at 6 245 and 336 with the latter resonance the stronger The spectrum remains unchanged after 2 h at lOO"C, and although it superficially appears similar at 300°C the peaks have moved to more positive shifts by ca 10 ppm (Table 1) After heating at 500 "C some narrow peaks appeared, indicating crystallisation, shifted to 6 267 2 and 336 8 and of changed relative intensity These narrow peaks sit on top of the residual peaks of the former broader resonances On further heating to 800°C the broad components have been completely removed and the two sharp Fig.1 170MAS NMR spectra of hafnia gel (a) dned at 25 "C and then heated to (b) 100"C, (c) 300 "C, (d) 500 "C, (e) 800 "C and (f)of unenriched bulk hafnia Table 1 Heat treatments and peak positions of HfOz gels sample heat treatment ii(170) H1 gel dried at 25 "C 245, 336f2 5 H2 H1+2h, 100°C 245, 335 k2 5 H3 H2+2 h, 300°C 258, 346 f2 5 H4 H3 +2 h, 500 "C 2672, 3368k04 H5 H4+2 h, 800 "C 267 2, 336 Ok 0 4 natural abundance HfO, 2669, 3362f04 1952 J Muter Chem, 1996, 6(12), 1951-1955 resonances and some spinning sidebands remain. A comparison with calcined monoclinic Hf0, containing I7O at natural abundance is given in Fig.l(f) with peaks at 6 266.9 and 336.2. The peaks formed when the gel crystallised and for normal bulk HfO, are consistent and agree with the previous natural- abundance study.' Of the peaks in the room-temperature gel at 6 245 and 336, the latter is very close to that from the final crystalline product, although the lines are much broader. In previous studies of the field dependence of the linewidth in similar ionic gels6v7 it has been demonstrated that the line broadening is caused by chemical shift dispersion in such ionic oxides, as a result of the disorder present in the samples in the gel state. However, it would be misleading to regard the gel spectrum as simply a broadened version of that from the crystal since the intensity ratio (ca.1 :2) is not the same in the crystalline material (1: 1). Heating to 100"C leaves the spec- trum and hence the structure unchanged from the room-temperature gel. However, at 300 "C the peaks have shifted significantly, which must be an indication that structural changes are taking place as a precursor to crystallisation. At 500 "C the much narrower resonances indicate that crystallis- ation has largely taken place, although some significant amorphous domains remain as broad underlying resonances are observable. The shifts indicate structural change at 300 "C, showing opposite trends for the two peaks. The lower shift peak has continued to become more positive whilst that at 6 ca.346 has shifted back to 6 337 at 500°C (Table 1). This observation indicates that 170NMR is a subtle probe of structural change in amorphous oxides and crystallisation. At 800°C all of the amorphous component has been removed, leaving the two resonances of crystalline monoclinic HfO,. X-Ray powder diffraction (XRD) shows no evidence of crystallisation at 500 "C, which is a consequence of the typically nanocrystalline sample dimension of such gel-produced oxides that result in particle-size broadening of the XRD reflections. NMR, being short-range in nature, shows crystallisation read- ily as all the local environments become the same causing the significant line narrowing. Assignment of the two resonances as 6 267 from OHf, and 6 336 from OHf, is in accordance with decreasing chemical shift corresponding to increasing coordination number.It is also interesting to compare the structural behaviour observed here with that of other pure oxides. For TiO, the initial gel was a mixture of OTi, and OTi, and increasing heat treatment gradually eliminated the OTi, environment, with the shift of the OTi, site becoming more anatase-like.6 The chemically very similar ZrO, again showed peaks in the gel that largely line up with the monoclinic lines. However, throughout a third peak from the tetragonal form could be observed for ZrO,. On crystallisation there were clearly tetragonal crystallites present. This is not the case for HfO,, despite a tetragonal form existing at higher tempera- tures.The difference in the structure of the gels between ZrO, and HfO, must be a reflection of the relative stability of the monoclinic and tetragonal forms in the two systems and the influence of surface stabilisation. These subtle differences could have a profound influence on the sintering behaviour of these oxides. Alkali-metal and alkaline-earth-metal hafnates The 170MAS NMR spectra from some alkali-metal and alkaline-earth-metal hafnates are shown in Fig. 2 and the isotropic chemical shifts are summarised in Table 2. The spectra are characterised by some very narrow resonances and accompanying sidebands. Such high-resolution solid-state 170 NMR spectra are possible in highly ionic systems as C, is small as the charge density around the oxygen ion becomes closely spherically symmetric.The (1/2, -1/2) transition from such systems is then dominantly broadened by chemical shift anisotropy and the sidebands observed in Fig. 2 do resemble h h. *c-r 1 I 400 I300 I200 I 100 I0 I 6 Fig.2 170 MAS NMR spectra of (a) Li,HfO,, (c) CaHfO,, (d) SrHfO, and (e) BaHfO, (b) Na2Hf03, such powder patterns, although accurate analysis would require both chemical shift anisotropy and quadrupole effects to be taken into account. Most spectra show two isotropic resonances that lie in the range 6 240-332. For these ternary hafnates their crystal structures have two inequivalent oxygen sites in the ratio 2 : 1 and, for sodiumt, calcium and strontium samples the spectra integrate to accurately reproduce this ratio.Given the long T' measured for BaHfO, the accurate 2: 1 ratios determined here indicate that either the T,s are very much shorter or more likely that the sites have very similar Tls and hence similar degrees of saturation. The separation of the two isotropic resonances in these compounds decreases, 18.6, 8.6 and 2 ppm for sodium, calcium and strontium respect- ively, which is related to decreasing distortion of the crystal structures (Table 2) away from cubic symmetry. Li,HfO, reson-ances show a large deviation from 2: 1 with the spectrum giving a ratio of ca. 4: 1. Although T' relaxation effects cannot definitely be ruled out, a large range of relaxation delays has been tried with no observed change in the spectrum, and a similar discrepancy in intensity was seen for the isostructural Li2Zr0313so that relaxation effects are unlikely to be causing this intensity variation.These layer structures may be suscep- tible to some systematic stacking faulting that could lead to this change of intensity, but for confirmation would need to be subjected to a detailed electron microscopy study. BaHfO, shows a single resonance as expected from a cubic system where all the oxygen sites have become equivalent. An intriguing observation is that the ratio of the isotropic 170 chemical shifts from isostructural zirconates and hafnates fall in a very limited range of 0.84-0.88 (Hf/Zr, Table 2). An extensive study of such shifts in ternary titanates and zirconates shows that the isotropic 170chemical shifts also give nearly constant ratios for the isotropic shifts of isostructural corn-pounds13 so that the titanium, zirconium and here hafnium 7 A detailed discussion of the structural chemistry and 23Na MAS NMR spectra of NaHfO, is given in ref.12. J. Muter. Chem., 1996, 6(12), 1951-1955 1953 Table 2 Crystal data on HfO, and ternary hafnates and 170 isotropic chemical shifts compound crystal system space group Hf02 monoclinic P21lC Li,HfO, monoclinic cc Na,HfO, CaHfO, monoclinic orthorhombic C2/m Pbnm SrHfO, orthorhombic Pbnm BaHfO, cubic Pm3m are having the strongest effect in determining the 170chemical shift with the alkali metal having a secondary effect Hf02-SO2 An Hf0,-Si02 gel consolidated at 500 "C and then 850 "C for 2 h penods gave a 29S1 NMR resonance at 6 -96 [Fig 3(a)] This indicates that silicon is probably in a Q3 environment (ze three bridging and one non-bridging oxygen bond) The 170NMR spectrum using MAS at 9 4 kHz shows four isotropic peaks at 6 343, 252, 139 and -6 4 [Fig 3(b), the shoulder at 6 cu 380 is from the ZrO, rotor] All peaks, even those associated with ionic hafnium oxide-like environments are quite broad, indicating that the material is still highly amorph- ous Fast spinning is advantageous as it largely removes spinning sidebands and causes efficient narrowing of even Si-0-Si resonances which have quite large C, values and these sites produce the structured resonance at 6 -64 The four resonances gwen above can be attributed to OHf,, OHf,, HfOSi and OSi2 respectively 170and 29S1 give a consistent picture of the structure which is amorphous and has silicon- and hafnium-rich parts, but there is some hafnium present in the silica-nch part as indicated by the presence of HfOSi bonding and the Q3 nature of the silicon environment It is also clear from this data that via this gel route single-phase HfSiO, has not formed, as all the oxygen would be present as Hf-0-Si linkages and the silicon resonance would be much more positive corresponding to a Qo unit Comparison of the 170peak positions from ZrOSi" and HfOSi gves a ratio of 090, close to the range for the crystalline hafnates, again showing that the metals are having some influence (although equating the peak position with the isotropic chemical shift is not as certain for such a resonance compared to the more ionic linkages) The effect of the silicon is to move the shift Fig.3 (a) 29S1 and (b) 170MAS NMR spectra of an equimolar Hf0,-SiO, gel heated to 850°C (the shoulder at 6 ca 385 is from the ZrO, rotor) 1954 J Muter Chern, 1996,6(12), 1951-1955 ~~~ structure type 6,so (l70) ratio 6,,, ("0) Hf/Zr baddelyite-t ype 267, 336 0 84 Li2Zr0,-type 237 8,249 2 0 85 Na,ZrO,-t ype GdFe0,-type 239 8, 258 4 292 7, 284 1 0 84 0 87 GdFe0,- type perovskite 296, 298 331 9 0 87 0 88 well outside the range associated with the crystalline alkali- metal and alkaline-earth-metal hafnates and zirconates Hf02-GeO, For comparison, gels in the system HfO2-GeO2 were formed, heating the gel successively from room temperature to 100, 300, 500 and 950°C, and subsequently at room temperature recording the 170MAS NMR spectrum (Fig 4) Again several broad 170 resonances are observed which are summarised in Table 3 The peaks at 6 ca 330 and 240 approximately agree with OHf3 and OHf, from pure Hf02 gels The identity of the resonance around 6 115 is less certain Given the stoichiometry A Fig.4 170 MAS NMR spectra of an equimolar Hf0,-GeO, gel (a)dned at room temperature and then heated to (b) 100 "C, (c) 300 "C, (d)500 "C and (e)950 "C (the peak at 6 ca 385 is from the ZrO, rotor) Table 3 Heat treatments and 170 NMR peak positions of Hf02 GeO, gel sample heat treatment W70)(k5 ppm) G1 dried at 25 "C 327,239, 112 G2 G3 G4 G5 G1+ 1 h, 100 "C G2 + 1 h, 300 "C G3 + 1h, 500 "C G4 +6 h, 950 "C 332, 243, 118 340,248, 118 340, 245, 115 357,295, 115 The peak at 6 ca 385 is from the ZrO, rotor of the gel an immediate assignment could be to GeOGe linkages, indicating that the system is completely phase separ- ated into GeO, and Hf02 domains.However, GeO, gels produce broader I7Oresonances with well defined quadrupolar structure and different shift.', With the magnetic field applied here and the MAS employed it is likely that such linkages would not be narrowed and hence are mostly lost from the MAS NMR spectrum. The XRPD pattern of the product heated to 950°C shows it contains a crystalline component of HfGeO,.This has the scheelite structure with GeO, units linked by hafniums and a possible assignment of the 6 115 peak is to HfOGe linkages. The shift is quite close to the position of HfOSi at 6 139. The gradual increase of the relative intensity of the 6 115 peak reflects an increase in the fraction of GeO, units linked by hafniums, which may be an indication that the sample is becoming more single phase. The HfOGe resonance also begins to show some structure and can be fitted to a quadrupole lineshape with C, =5.2 MHz, an asymmetry parameter of 0.65 and an isotropic chemical shift of 6 185. Such behaviour contrasts with the HfSiO, sample, which at 850°C still showed marked phase separation.This paper indicates that 170NMR is a very subtle atomic-scale probe of the structures of a wide range of hafnate-based materials. Conclusions Hf02 gels show two resonances that can be attributed to OHf, and OHf, that shift with heat treatment, reflecting structural development in the amorphous gel and show marked nar- rowing on crystallisation into the monoclinic form, but unlike ZrO, there was never any evidence of the tetragonal form being present. Crystalline alkali-metal and alkaline-earth-metal hafnates show very high-resolution spectra, with the number of resonances and their splitting reflecting the distortion of the crystal structure. Comparison of the isotropic 170chemical shifts of the isostructural hafnates and zirconates shows an almost constant ratio of 0.86f0.02.Binary gels of HfO, with SiO, and GeO, show there is a higher tendency to form a solid solution with GeO,. M.E.S. thanks the EPSRC for funding research into the development of 170characterisation of oxide materials through GRlJ2393 8. References 1 T. J. Bastow and S. N. Stuart, Chem. Phys., 1990,143,439. 2 E. Oldfield, C. Coretsopoulos, S. Yang, L. Reven, H. C. Lee, J. Shore, 0.H. Han, E. Ramli and D. Hinks, Phys. Rev. B, 1989, 40, 6832. 3 R. K. Harris, M. J. Leach and D. P. Thompson, Chem. Muter., 1992,4,260. 4 R. Dupree, Z. P. Han, A. P. Howes, D. MckPaul, M. E. Smith and S. Male, Physicu C, 1991, 175,269. 5 T. J. Bastow, M. E. Smith and H. J. Whitfield, J. Muter. Chem., 1992,2, 989. 6 T. J. Bastow, A. F. Moodie, M. E. Smith and H. J. Whitfield, J. Muter. Chem., 1993,3, 697. 7 F. Ali, M. E. Smith, S. Steuernagel and M. E. Smith, J. Muter. Chem., 1996,6,261. 8 M. E. Smith and H. J. Whitfield, J. Chem. Soc., Chem. Commun., 1994,723. 9 P. J. Dirken, M. E. Smith and H. J. Whitfield, J. Phys. Chem., 1995, 99,395. 10 T. J. Bastow, L. Murgaski, M. E. Smith and H. J. Whitfield, Muter. Lett., 1995,23, 117. 11 P. J. Dirken, R. Dupree and M. E. Smith, J. Muter. Chem., 1995, 5, 1261. 12 J. V. Hanna, M. E. Smith and H. J. Whitfield, J. Am. Chem. Soc., 1996,118,5772. 13 T. J. Bastow, P. J. Dirken, M. E. Smith and H. J. Whitfield, unpub- lished results. 14 T. J. Bastow, R. Dupree, M. E. Smith and H. J. Whitfield, unpub- lished results. Paper 6/04406I; Received 25th June, 1996 J. Matev. Chem., 1996, 6(12), 1951-1955 1955

ISSN:0959-9428    

DOI:10.1039/JM9960601951

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Characterisation and properties of the non-stoichiometric perovskite, Ca,Fe, -,Nb,O, (0.45< x < 0.65) Jose A. Chavez,"+ Terence C. Gibbband Anthony R. West" "University of Aberdeen, Department of Chemistry, Meston Walk, Aberdeen, UK AB24 3UE bUniversity of Leeds, Department of Chemistry, Leeds, UK LS2 9JT The cubic perovskite phase Ca2Fe2-,Nb,0, (0.45< x < 0.65) has been prepared by solid-state reaction at 1400 "C followed by rapid cooling to room temperature. On slower cooling, or heating below 1360 "C, it decomposes to a mixture of Ca2Fe205 and Ca2FeNb06, but the quenched single-phase material is kinetically stable up to at least 950 "C. The oxygen content of composition x =0.60, determined by thermogravimetry, can be varied over the range 5.55 <y < 5.89 by heat treatments under a range of conditions, including H2-Ar at 700 "C to obtain low y, and high-pressure oxygen, 420 bar at 950 "C, for high y.The main oxidation state of Fe, as determined by Mossbauer spectroscopy was Fe3+, but about 13% Fe4+ was present in the as-prepared sample, increasing to 42% under high-pressure 02.For most y values, the crystal structure was that of a cubic perovskite with Fe/Nb disorder on the B sites and some oxygen vacancies; at high y contents, an orthorhombic distortion gave a structure resembling that of Ca2FeNb06. Electrical properties range from semiconducting at high y consistent with Fe3+ /Fe4+ exchange, to insulating at low y; although there was no direct evidence of oxide ion conduction over the range 25-300 "C, the rapid adjustment of oxygen content y on changing the atmosphere at 700-900 "C indicated significant mobility of oxide ions.The phase is therefore likely to be a mixed oxide ion-electron conductor. The perovskite structure shows great diversity in its stoichi- ometry with a wide range of substitutions possible on the A and B cation sites, together with the possibility of both cation and anion vacancies. The brownmillerite structure of Ca2Fe205 can be regarded as derived from perovskite by ordering of oxide ion vacancies, resulting in a tetrahedral coordination for some Fe atoms. The possibility of high levels of oxide ion conduction in certain anion-deficient perovskites remains an intriguing possibility, especially for applications as fuel cell cathodes and catalysts where mixed oxide ion-electron conduc- tion may be desirable.Following the report of an extensive range of cubic, perovskite-like solid solutions in Nb-doped Ca2Fe205,' we have investigated their stoichiometry, structure and electrical properties. Our results are considerably different to those reported in ref. 1; two solid solution phases form, one based on orthorhombic Ca2FeNb06 and the other on a cubic perovskite structure stable over a more limited composition range than given in ref. 1. These latter results are reported here. Experimental Samples were prepared by solid-state reaction of CaCO,, Fe203 and Nb205. Reagents were dried, weighed out in the required proportions, mixed with an agate mortar and pestle and fired in Pt boats at 90O-135O0C, with a final firing at 1400°C for 20 h of the reground samples.Products were characterised by X-ray powder diffraction, Stoe STADI P diffractometer, Cu-Ka, radiation. Oxygen contents and vari- ations in oxygen content were determined by thermogravimetry (TG) in various atmospheres. Samples were also heated under O2 at pressures up to 420 bar and temperatures up to 950°C in a Morris HPS-5015E7 high-pressure furnace in order to attempt a modification of their oxygen content. The oxidation states of the Fe in some samples were estimated by Mossbauer spectroscopy. Data were collected at room temperature using a 57Co/Rh source matrix and isomer shifts were calibrated relative to metallic Fe.Electrical conductivity was determined t Permanent address: National University of Mexico, UNAM, IIM, Apdo Postal 70-360 Mexico, D.F. by ac impedance measurements on sintered pellets with elec- trodes fabricated from either Au or In/Ga in a two-terminal arrangement. The frequency range 30 mHz-10 MHz was scanned using a combination of Solartron 1250/1286 and Hewlett-Packard 4192 instrumentation; data were processed using in-house software. Results and Discussion Synthesis of Ca,Fe, -,Nb,O, Attempts to investigate the possible doping of Ca2Fe,05 with Nb involved a preliminary investigation of the relevant region of the CaO-Fe,O,-Nb,O, phase diagram and then focused on a study of the system Ca2Fe205-Ca2Nb207. A detailed phase-diagram study2 was made which indicated the occur- rence, stoichiometry range and thermodynamic stability of a cubic, perovskite-like phase.Its formula may be represented as Ca2Fe2-,Nb,0,. It forms, and is stable, only above ca. 1360"C, although it is readily preserved to room temperature by removing from the furnace at e.g. 1400°C and cooling in air over a period of a few minutes. This quenched, single-phase material is kinetically stable up to at least 950°C. At higher temperatures, e.g. 1200-13OO0C, or on slow cooling from ca. 14OO0C, Ca2Fe2-,Nb,0, decomposes to give a mixture of Ca,Fe205 and Ca2FeNb06 .2 Ca2Fe2- ,Nb,O, is a solid-solu- tion phase with composition 0.45 <x < 0.65; it appears not to have an ideal composition within this range.A variety of characterisation studies on composition x =0.60 were made and are reported here. Determination of oxygen content, y At the outset, the oxygen content of Ca2Fe2-,Nb,0, was not known; if the Fe were present exclusively as Fe3+, then the oxygen content would be y = 5 + x. If, however, some Fe2+ or Fe4+ formed during synthesis, then the oxygen content would differ. Two methods were used to assess oxygen content and the oxidation state of Fe: TG and Mossbauer spectroscopy. Agreement was generally excellent; unless otherwise noted, the quoted y values are based on TG results. J. Mater. Chem., 1996, 6(12), 1957-1961 1957 Miissbauer spectroscopy. Mossbauer spectra are shown in Fig 1 for (a)the related compound Ca,FeNbO,, which can be used as an Fe3+ standard and (b)as-prepared Ca,Fe, 4Nbo ,OY (z e x=O 6) Also shown, (c), is a spectrum of x=O 6 after treatment in high-pressure O2 (discussed later) Fitting param- eters for the Mossbauer spectra are listed in Table 1 The presence of a disordered mixture of cations in a perovsk- ite can have a major effect on the observed 57Fe Mossbauer spectrum3 This factor is more important than any small deviation from cubic symmetry in the (averaged) unit cell Thus, in SrLaFeSnO, where all the iron is present as Fe3+ (an S-state ion with no intrinsic quadrupole effect), the disordered Sr,La and Fe,Sn cation charges generate a substantial electric- field gradient at the Fe which is manifested as an apparent (if somewhat broadened) quadrupole doublet Similar behaviour has also been found in the series A2FeMO6 (A=Ca, Sr, Ba, M =Nb, Ta) where only the M,Fe cations are di~ordered,~ in the equivalent M =Sb compounds where partial Fe,Sb order is found, there is clear evidence for multiple Fe sites The larger nominal charge difference for Fe3+ /Nb5+ results in much larger apparent splittings in A2FeMO6 than for Fe3+ /Sn4+ in Sr LaFeSnO, The spectrum of Ca2FeNb0, [Fig l(a)] is similar to that of CazFe, 4Nbo 60y[Fig 1(b)],but the spectrum of the latter has an asymmetry at low velocity which is clear evidence that some of the iron is in a higher oxidation state than 3+ The curve-fit shown in Fig l(b) is not perfect as the component lines are clearly compound because of the disorder and deviate from the normal Lorentwan profile which has been assumed throughout, however, the use of a more complex treatment is not warranted in the absence of an appropriate model Thus, IIIIILI I -2 0 2 velocity/mm s-l Fig.1 Mossbauer spectra of (a) Ca2FeNb06, (b) Ca2Fel 4Nbo 60568 and (c) Ca2Fe1 4Nb0 6O5 89 Table 1 Mossbauer fittmg parameters" compound d/mm s-' d/mm s-' r/mm s-l % area Ca2FeNb06 +0385(1) 0606(2) 0451(3) 100 Ca2Fe, 4Nb0 605 68 +O 386(2) 0 647(3) 0 543(5) 87 5 -0342(5) 0216(8) 0300(15) 125 Ca2Fe, 4Nbo 605 89 +o 365(5) 0 283(5) 0 384( 1) 55 3 -0 007(4) 0 207(5) 0 327(1) 44 7 "y values calculated from TG data 1958 J Muter Chem, 1996, 6(12), 1957-1961 the standard deviations quoted in Table 1 for the isomer shift, 6, quadrupole splitting, A, and linewidth, I-', ignore a potentially greater systematic error The relative area of the Fe3+ component is, however, well defined by the asymmetry and is comparatively insensitive to the model adopted The similarity of the parameters for the Fe3+ component in Ca,FeNbO, and Ca,Fe, ,Nbo ,OY suggest that the Fe and Nb cations are similarly disordered in the two structures If the high oxidation state Fe component corre- sponds to Fe4+ (rather than Fe5+ which is also known in perovskites) then the peak area ratio associated with Fe3+ and the additional peak in Fig l(b) is 87 13 This then leads to a calculated oxygen content at x =0 6 of y =5 691 Thermogravimetry. TG studies of the same as-prepared material were carned out in an atmosphere of flowing Ar, the sample was heated to 900 "C, held at that temperature for 2 h and cooled, these conditions were chosen since studies5 on other Fe-containing materials had shown reduction of Fe4+ to occur under similar conditions The results (Fig 2) showed a significant mass loss at 400-600 "C, with a constant mass thereafter and no significant change, apart from a slight baseline drrft, on cooling Assuming that all the Fe4+ had been reduced to Fe3+ during TG, the oxygen content of the original sample was calculated to be y= 5 684, with an Fe3+ /Fe4+ ratio of 88 12, in good agreement with the Mossbauer value This provides confirmation that the second oxidation state is indeed Fe4+ Variation of oxygen content These TG results indicated that the oxygen content of the perovskite phase is capable of significant variation Two sets of expenments were therefore carned out to try and extend further the range of oxygen contents For both, the same batch of starting matenal was used as that for which the oxygen content had been determined, above The TG-determined value of the oxygen content, y= 5 684, was taken as the starting oxygen content In one set of expenments, samples were heated under high- pressure oxygen, the design and mode of operation of the Morns furnace was such that the maximum pressure was attained at the highest temperature in a particular heat-cool cycle, on subsequent cooling, at 1-2 "C min-', the pressure inside the sealed vessel gradually reduced unless the vessel was deliberately opened to the atmosphere (not in these experi- ments) The oxygen content of the products was determined by TG in Ar to 900°C as above and, for one sample, by Mossbauer spectroscopy Results are summarised in Table 2 for two heat treatments At the highest pressure used, 420 bar, the value obtained for the oxygen content was similar for the two techniques The Mossbauer spectrum [Fig l(c)] is clearly rather different to that of the starting material [Fig l(b)] Again, the model used for fitting may not be an entirely ., . . . , . . . , . . . , . . . 5-70 I -5.68 -0.1 : 0 -5.66 h. 0 E-0.2: 0 Y E: .5.64 0 -0.3 0 -0 5.62 Fig.2 Mass as a function of temperature for Ca2Fe, 4Nbo 60yin flowing Ar 0,heating, A,cooling Table 2 Data for Ca2Fe,,4Nbo,60, treated in high-pressure 0, annealingconditions a/A y from TG Fe3+/Fe4' 950°C, 2 h, 3.8452 5.891 58/42 P,,, z 420 bar 5.915" 600°C, 2 h, 3.8476 5.768 76/24 P,,, z 45 bar "Calculated from Mossbauer spectroscopy. accurate description of the phase overall, but the asymmetry is well defined. As well as a substantial increase in oxidation of Fe to the +4 oxidation state, there is a noticeable change in the isomer shift and quadrupole splitting parameters. Thus, the isomer shift of the Fe4+ component is much more positive whereas that of Fe3 has decreased. One possible explanation + is that an electron-hopping process is taking place on a timescale slightly longer than the Mossbauer timescale of lo-' s so that each Fe environment is partially averaged.The decrease in the average quadrupole splitting could then be due merely to the reduction in charge discrepancy between the B-site cations. Conductivity data that may support this notion are discussed later. The isomer shift for the component at lower velocity confirms the existence of a higher oxidation state. Although the associated quadrupole splitting is in prin- ciple anomalously small for both the high-spin and low-spin states of Fe4+, both of which might be predicted by analogy with high-spin Fe2 to show a large valence-electron contri- + bution, it should be remarked that such an effect has not been observed in any of the perovskite oxides studied to date.The reason for this remains unknown. In the second set of experiments, TG studies were carried out in a range of atmospheres and the change in mass or oxygen content monitored directly. In each case, samples were heated at 10°C min-' to 700-9OO0C, held isothermally for 2 h, then cooled. The starting material was again taken to have the composition x=O.6, y=5.684. On heating in air [Fig. 3(a)] a significant mass loss occurred over the tempera- ture range 400-600 "C, which was partially recovered on cooling, to give a final oxygen content at room temperature of y= 5.667, corresponding to a residual Fe4+ content of 9.6%. On comparing Fig. 3(a) with Fig.2, a similar pattern of behaviour is seen on heating but in Ar (Fig. 2) the mass loss occurs rather more rapidly and levels off, at a lower tempera- ture, at 7 =5.60; in air [Fig. 3(a)] the oxygen content at 900 "C appears to be slightly greater than y=5.60, indicating a residual amount (1-2%) of Fe4+. Under flowing oxygen [Fig. 3(b)] the sample gained mass over the range 400-500°C and then lost most of this mass at higher temperature. The latter loss was recovered on cooling, to give a final oxygen content at room temperature of 5.808, corresponding to 30.0% Fe4+. Under strongly reducing conditions [5%H2-95%N2, Fig. 3(c)] a continuous mass loss was observed over the range 25-65OoC, with most of the mass loss occurring between 400 and 600°C.No further mass change occurred on cooling, giving a final oxygen content of 5.549. To account for this, the Fe is assumed to be present as an Fe3+/Fe2+ mixture, in the ratio 93 :7. The above results demonstrate a significant variation in oxygen content, depending on heat treatment, associated with a variation in oxidation state of Fe. They also demonstrate that this variation is largely confined to low temperatures. At the temperature of synthesis, 1400 "C, extrapolation of the TG results indicates that Fe4+ is most unlikely to be present, even in an atmosphere of pure 02.On cooling in air, rapid uptake of oxygen, with oxidation of some Fe3+ to Fe4+, occurs below ca. 800°C and the final oxygen content will depend on the cooling rate and the atmosphere.0.1 : (a) 5.70 000 0 0 0 0 0 oo~o~ i 0 7 5.68 AAA. Ao-0.1 ; A0 15.66 A0 -0.2 A0 15.64A0 A0 -0.3 A0 PO 5.62......15.60 -0.4 0 200 400 600 800 lo00 5.84 A A A A A A A 0.6 1 0 0 5.80 Obo 0 0.4 A0 5.76 Y A0d 0 0 00.2 so 15.72 0 4 O' 15.68 L..... .......................... I 0 100 200 300 400 500 600 700 800 ................................ 5.70 0 (') 15.66 0 0 1 0 15.62-0.4 I Fig.3 Mass as a function of temperature for Ca2Fel,4Nbo.60, in flowing atmospheres of (a) air, (b) oxygen and (c) H2-N2 (5:95) mixture: 0,heating: A, cooling Crystallographic data X-Ray powder diffraction data were indexed on a simple, primitive cubic unit cell for most oxygen contents, although at high y, 35.8, a distortion to orthorhombic symmetry occurs.Values for the cubic cell parameter us. composition are shown in Fig. 4; an approximately linear decrease in a with increasing oxygen content occurs, consistent with a gradual increase in bond strength and lattice energy on progressive oxidation of Fe. Clearly, there is interstitial space available for the extra oxygen and the unit cell contracts when oxygen enters these sites due to the increase in Fe-0 bond strength and shortening 3.880 3.875 3.870 H-NZO Ar(8995)3.865 E air23.860 1 3.855 4 3.850 0 HOP* 3.845 0 3.8401, I.............................,I 5.4 5.5 5.6 5.7 5.8 5.9 6.0 Y Fig. 4 Cubic cell parameter a us.oxygen content y for Ca,Fe, 4Nb0.60y.*,pseudo-cubic parameter from averaged orthorhombic parameters. J. Muter. Chem., 1996, 6(12), 1957-1961 1959 0: of the Fe-0 bonds associated with the increased oxidation state of Fe Rietveld refinement was performed on XRD data for one composition, x =0 6, y =5 68, using the basic perovskite struc- ture as the starting model Fe and Nb were placed on the B sites with overall full occupancy Oxygen occupancies and oxygen temperature factors could not be refined independently, the oxygen occupancy was therefore fixed at the value given by the formula and the temperature factor refined A satisfac-tory refinement was obtained [Table 3, Fig 51, confirming the essential correctness of the cubic perovskite structure model Electrical properties The electncal properties of Ca2Fe, 4Nbo 60rwere evaluated using ac impedance measurements A typical impedance data set at 200 "C is shown in Fig 6 for y =5 549 The impedance complex plane plot [Fig 6(a)] shows a broad arc which extrapolates to, or close to, the ongin The low-frequency intercept on the 2' axis gives the total resistance of the sample On replotting the same data in the 2" and M" spectroscopic plot format [Fig 6 (b)],' it is clear that the M" spectrum is a double peak, the low-frequency component occurs at about the same frequency as the 2" peak whereas much of the higher frequency peak is beyond the available frequency range The data were fitted to the circuit shown in [Fig 6(c)], with values for the component parameters also listed, the quality of fit is shown by the solid curves in Fig 6(a), (b) The low-frequency peak in the M" spectrum, and the main peak in the 2" spectrum, dominates the total resistance of the sample but its capacitance is somewhat larger than that of the less resistive, high-frequency M" peak (since M" peak heights are inversely proportional to capacitance') Since both capacitance values are small, a few pF, we can conclude that the sample is electrically heterogeneous with two components, 1 and 2, in the approximate proportions, given by the ratio of their reciprocal capacitances, of 1 143 Since the resistances of the two regions are in the ratio 137 1, their resistivities are in the approximate ratio 20 1 We do not know the origin of this heterogeneity but it may be associated with a variation in oxygen content through the sample, especially as this particular sample prepared by H2-reduction was not subjected to pro- longed low-temperature annealing to ensure homogeneity Table 3 Atomic parameters for Ca,Fe, 4Nbo 605691 (e s d s in parentheses) atom occupancy x/a y/b zfc UV Ca 10 05 05 05 0046(2) Fe 07 0 0 0 0 0151(8) Nb 03 0 0 0 0 0151(8) 0 0 9467 0 0 05 0088(6) 80! -20t 1 10 20 30 40 50 60 70 80 90 100 110 120 2Bldeglees Fig.5 (a) X-Ray diffraction pattern for Ca,Fe, ,NbO 60y,synthesised in air at 1400°C for 20 h, (b) difference between observed and calculated profiles after Rietveld refinement 1960 J Mater Chem, 1996,6(12), 1957-1961 0 04 2 00 05 10 15 20 0 10-2 1oo 102 lo4 lo6 lo* frequency/Hz R1 R2-F7+T c1 Cl R1= 1 504 x106 R R2 = 1 096 xlOs R C1= 8 138 xlO-I2 F C2 = 5 701 xlO-"F A =4 415 x lo4 S A1 = 2 515 x10-" s B1= 2 948 x10" S Bl = 3 630 x10-" s nl = 0 3748 n2 = 0 6142 Fig.6 (a) Impedance complex plane plot and (b)M ,Z spectroscopic plot for the H,-reduced sample at 200 "C, (c) equivalent clrcuit used to fit data and parameters extracted for this data set Plots such as Fig 6 are typical of all samples studied here Resistances values were extracted from the low-frequency intercepts of the complex plane plots, on the 2' axis, and thus represent the total resistances of the samples These data are shown in Arrhenius conductivity format in Fig 7 and Table 4 A large variation in conductivities, depending on oxygen content, is seen The data fall approximately into two groups In the more oxidised samples, the Arrhenius plots are almost -1 0 15 20 25 30 35 1000 KIT Fig.7 Arrhenius plots for CazFe, ,Nbo 60y,the same sample was used throughout, with annealing treatments in the sequence (1)-(4) (1) as-prepared sample, 1400°C in air, (2) 420 bar 02,95OoC, (3) H2-N2 (5 95), 700°C, (4) air, 700°C 0/@,heating, A/V,cooling Table 4 Conductivity Arrhenius parameters activation energ y/eV annealing conditions heating cooling as-synthesised, air, 1400 "C, 20 h oxygen pressure, 950 "C, 2 h, P,,, HZ-NZ, 700 "C, 2 h z420 bar 0.26 0.28 1.37 0.27 0.29 0.38 air, 700 "C, 2 h 0.36 0.38 parallel, with activation energies in the range 0.26-0.38 eV; differences in conductivities are therefore attributed largely to differences in mobile carrier concentration, which is highest in the samples treated in high-pressure oxygen.For the reduced sample, the conductivity was many of orders of magnitude lower; in fact, it was too low to be measured (< S cm-l) below ca. 140"C, and had a much higher acti- vation energy. After heating in air to ca. 350°C during the impedance measurements, the data were not reproduced on subsequent cooling but came instead into the category of behaviour of the more oxidised samples; it appears likely that partial reoxidation of the reduced sample had occurred during the impedance measurements, consistent with the TG results [Fig.3(b)] which indicate that oxidation can occur above ca. 340 "C at a heating rate of 10 "C min-l. The impedance complex plane plots [Fig. 6(u)] showed no evidence of low-frequency polarisation effects, such as an 'electrode spike' which could have been attributed to double- layer effects associated with ionic transport. This, coupled with the observation that the activation energies for conduction of the more oxidised samples are quite low, indicates that the principal current carrier is electronic rather than ionic. The plots also showed no evidence for grain-boundary impedances, which would have had significantly higher capacitance values than those observed, in the pF range.The impedance data do, therefore, correspond to the bulk of the sample. The conductivity data (Fig. 7) may be interpreted as follows. The material is an electronic semiconductor. Its conductivity depends very much on the presence of Fe4+ ions which gives rise to a mixed 3 +/4 + valence state and allows the hopping of electrons (or holes) between adjacent Fe atoms in the perovskite structure. With increasing oxygen content, the Fe4+ concentration rises, as does the level of conductivity, which is therefore p-type. The conductivity appears to approach its limiting value in the sample prepared in high-pressure oxygen. Thus, the conductivities from data sets (1) and (2) differ by a factor of only about 3, as do the Fe4+ contents, and in the high-pressure-treated sample, nearly half the Fe ions are tetra- valent.The conductivity may well start to decrease if a further increase in Fe4+ content could be induced, since there would be less opportunity for Fe4+ /Fe3+ exchange to occur. In the H,-treated sample, the conductivity behaviour is totally different; the level of conductivity is very low and the activation energy is high. No positive holes (Fe4+ ions) are present; instead, the residual conductivity may be n-type owing to the small amount of Fe2+ present. In spite of the fact that the samples appear to equilibrate with the atmosphere and modify their oxygen contents rapidly, especially at ca. 400-600°C, there was no evidence for signifi- cant levels of oxide ion conduction in the temperature range 25 to ca. 300 "C.This is probably because the level of electronic conduction, to lo-, S cm-I at 25 "C, is several orders of magnitude higher than that which could reasonably be expected for oxide ion conduction. This situation may change at high temperatures (2400 "C) however, since the activation energy for oxide ion conduction is likely to be in the range 0.8-1.5 eV, several times larger than that for the electronic conduction measured here and hence, with increasing tempera- ture, the transport number of oxide ions should increase. J. A. C. thanks UNAM, Mexico for a scholarship. The high- pressure furnace was provided by EPSRC. We thank A. Herod for assistance with the Mossbauer measurements. References 1 I. J. Moraes, M. C. Terrile, 0. R. Nascimento, M. S. Li, R. H. P. Francisco and J. R. Lechat, Muter. Res. Bull, 1992,27, 523. 2 J. A. Chavez, PhD Thesis, Aberdeen University, 1995. 3 T. C. Gibb, J. Muter. Chem., 1992,2,415. 4 P. D. Battle, T. C. Gibb, A. J. Herod, S-H. Kim and P. H. Munns, J. Mater. Chem., 1995,5, 865. 5 M. A. Alario-Franco, J. M. Gonzalez-Calbet and M. Vallet-Reglo, J. Solid State Chem., 1983,49,219. 6 S. J. La Placa, J. F. Bringley, B. A. Scott and D. E. Cox, Actu Crystallogr., Sect. C, 1993,49, 1415. 7 P. D. Battle, T. C. Gibb and S. Nixon, J. Solid State Chem., 1989, 79, 75. 8 J. T. S. Irvine, D. C. Sinclair and A. R. West, Adu. Muter., 1990, 2, 132. Paper 6/04653C;Received 3rd July, 1996 J. Muter. Chern., 1996,6( 12), 1957-1961 1961

ISSN:0959-9428    

DOI:10.1039/JM9960601957

出版商:RSC    

年代:1996

数据来源: RSC