摘要:

Synthesis, crystal structure and magnetic properties of Ln, -,Sr,NiO, solid solutions (Ln =La, Nd, Sm and Gd; 1.0<x <1.67) Michael Jamest and J. Paul Attfield" Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge, UK CB2 1E Wand the Interdisciplinary Research Centre in Superconductivity, University of Cambridge, Madingley Road, Cambridge, UK CB3 OHE New members of the Ln,-,Sr,NiO,+d (Ln =La, Nd, Sm and Gd) solid solution series have been synthesized for x> 1.0. The upper limit of x is cu. 1.6 for La and 1.67 for Nd, Sm and Gd. The tetragonal K,NiF,-type crystal structures of these phases have been Rietveld refined using X-ray powder diffraction data. Thermogravimetry shows nickel to be in the +3 oxidation state. Magnetic susceptibility data reveal Curie-Weiss behaviour for temperatures in the range 6-300 K.An approximate phase diagram for the Ln,-xSr,Ni04+d system as a function of Ln3+ radius and x is presented. Recently there has been a great deal of interest in rare-earth- metal nickel oxide phases having the tetragonal K2NiF4 struc- ture (Fig. 1) which show structural, magnetic and electrical similarities to high- T, copper oxide materials. In particular, the properties of the Ln,-,Sr,NiO,+, solid solutions have been extensively studied. The La, -xSr,Ni0,+6 solid solution has been investigated by a number of over the range of solubility 0 <x <1.6. The Nd, -xSrxNiO,k solid solution (0<x< 1.6) has also been widely Chen et aL7 have investigated the structural, electrical and magnetic properties of Ln,-,Sr,NiO,+d for Ln=Pr (O<x< l.O), Sm (0.5<x<1.0) and Gd (x= 1.0).The stoichiometric S= 1/2 Ni3+ oxides LnSrNiO, are metallic as the 3d,~-~z and 3d,z orbitals are nearly degenerate and give rise to a quarter-filled a* band. This is in contrast to La,CuO, containing S= 1/2 Cu2+, in which the 3d,~-~~ and 3d,z orbitals are split by a pronounced Jahn-Teller distortion and the half-filled a(3dxzWyz)*band is further split by electron correlations to give a Mott-Hubbard insulator. La,-,SrxNiO,+d solid solutions with x >1 prepared under 150 bar oxygen pressure are reported to have 6 zO3 and the equilibrium: Ni4+ +02--Ni3++o-A is considered to lie to the right, like that in the La2~xSrxCu04 system: +cu3 +02-cu2++o-In an attempt to synthesise phases Ln, -,Sr,NiO,+, with the smaller cation Y3+ (x> 1.0) we previously observed that a monophasic sample was produced only for the composition with x=1.67, corresponding to a new, semiconducting, defect K2NiF4 phase, YSr,Ni,011.8 We' have also synthesized and characterised the isostructural members of this family of phases LnSr5Ni3011 (Ln=Dy, Ho, Er and Tm).' For lanthanide ions smaller than Tm3+ (i.e.Yb3+ and Lu3+) these K,NiF, com-pounds do not form, but rather the rhombohedra1 phases LnSr,NiO, having the K,CdCl, structure are produced."." In this paper we report the synthesis and characterization of some Ln, -xSr,NiO,id solid solutions containing the larger trivalent lanthanide ions of La, Nd, Sm and Gd for x >1.0.Experimental Sample preparation Polycrystalline samples Ln, -xSr,Ni0,+8 were synthesized from spectroscopic grade powders of strontium carbonate, t Address for correspondence: ANSTO, Lucas Heights Research Laboratories, Private Mail Bag 1, Menai, NSW 2234, Australia. nickel nitrate hexahydrate and the corresponding rare-earth- metal oxide, Ln203. Samples were produced for compositions Ln=La3+ (x=1.67), Nd3+ (x=l.O, 1.1, 1.2, 1.3, 1.4, 1.5 and 1.67), Sm3+ (x=1.67) and Gd3+ (x= 1.0, 1.33 and 1.67). Prior to weighing, the lanthanide oxides were pre-heated to 1000°C in air to decompose any carbonate material to the oxide. The powders were dissolved in dilute nitric acid, and an intimate mixture of the metal oxides was formed via the decomposition of a citric acid*thylene glycol gel.The residues were pelleted and sintered in a tube furnace at 1100 "C under flowing oxygen for up to one week with frequent regrinding and repelleting until no further reaction was evident by powder X-ray diffraction. I' I1 I I I I I I I *I .I I II 'I * 0 A C 0 B Fig. 1 Crystal structure of A,BO, with the tetragonal K,NiF, structure type J. Muter. Chew., 1996, 6(l), 57-62 Powder diffraction 1.O <X-Ray powder diffraction (XRPD) profiles were recorded on a Philips PW1710 diffractometer, utilising Cu-Kcr radiation. Data of sufficient quality for structural refinement were col-lected over 13d 28< 113", in steps of 0.025", with integration times of 12 s.These structural refinements were carried out by the Rietveld method12using the GSAS program13and a refined background function. Thermogravimetry Thermogravimetry of ca. 30 mg samples of Lno.33Srl.67Ni04-a (Ln =La, Nd, Sm and Gd) and GdSrNi04-a were carried out using a Stanton Redcroft STA 1500 simultaneous thermal analyser. The sample was reduced under a mixture of 5% hydrogen in nitrogen (flow rate 58 ml min-') over a tempera-ture range of 15-900 "C at a heating rate of 10"C min-'. Magnetic susceptibility measurements Magnetic susceptibilitieswere measured for Lno.33Sr,.,7Ni0,-(Ln=La, Sm and Gd) and Nd2-xSrxNi04-6(x=l.O, 1.1, 1.3, 1.4, 1.5 and 1.67) using a Quantum Design SQUID magnet-ometer under an applied field of 3.0T.Samples were cooled to 6 K in zero field and the magnetization was measured while warming to 300 K. Results Preparation An attempt to synthesize the composition Lao.33Sr,.67Ni04-a led to a near monophasic sample, with a small amount of the impurity phase Sr,Ni,O,, .I4 Single phase samples were obtained for Nd2-xSrxNi04-a(x = 1.0, 1.1, 1.2, 1.3, 1.4, 1.5 and 1.67), Sm0.33Sr1,67Ni04-6and Gd,-,SrxNi04-a (x= 1.0, 1.33 and 1.67). Structural characterisation using X-Ray powder diffraction Each of the above compositions was studied by XRPD. In the case of the sample corresponding to a nominal composition of Lao.33Srl.67Ni04-a,as well as those for Nd2-xSr,Ni04-b (1.0bx d 1.5), the X-ray diffraction patterns were collected over the reduced range of 10<28<70", with steps of 0.025" and integration times of 1 s.The peak positions were indexed on a K2NiF,-type cell, and the refined cell parameters are shown in Table 1. The structural parameters for Lno.3,Srl~67Ni04-6(Ln =La, Nd, Sm and Gd), as well as Gd, -xSrxNiO, (x = 1.0 and 1.33)were determined by Rietveld refinement, using the GSAS program, in space group 14/mmm, with Ln and Sr disordered over the K sites. A good fit was obtained using a pseudo-Voigt peak shape function. In the final stages, the occupancy of the O(2) site was constrained to be fully occupied, while the occupancy of the O(1) site was freely refined. A typical set of observed, calculated and difference diffraction profiles, for Sm0.33Sr1.67Ni04-a,is shown in Fig.2. The observed d-spac-ings and peak intensities from the XRPD pattern of Sm0.33Sr1.67Ni04-6are given in Table 2. Structural details and Table 1 Refined cell parameters for Nd,-,Sr,Ni04-, (l.O<x< lS), with e.s.d.s in parentheses X a/A CIA VIA3 1.o 3.801( 1) 12.293(3) 177.60(8) 1.1 1.2 1.3 3.802( 1) 3.804( 1) 3.806(1) 12.275(3) 12.260(3) 12.248(4) 177.44(8) 177.41(8) 177.42(9) 1.4 3.806( 1) 12.247(2) 177.41(7) 1.5 3.806(1) 12.259(2) 177.58(7) 58 J. Mater. Chem., 1996, 6(l), 57-62 v)E 0.0 3 0 cu 0.6 z \ .-2. 0.4 v)cQ) c. .-t 0.2 0.0 ' I I I I 1 I I 20 30 4a 50 60 70 Bo 90 100 110 2wdegrees Fig.2 Observed (points), calculated (full line) and difference XRPD patterns for Sm,,33Srl,67Ni04-6 Table 2 Observed d-spacings and peak intensities for Smo.33Sr1.67NiO4-, hkl d-spacing/A relative intensity 002 6.102 7 101 3.612 11 004 3.062 11 103 2.776 100 110 2.677 77 112 2.454 2 105 2.059 15 006 2.044 12 114 2.0 17 28 200 1.895 34 202 1.810 1 211 1.679 3 116 1.626 10 204 1.612 5 107 1.591 5 213 1.566 26 008 1.534 3 215 1.395 6 206 1.391 10 2201118 1.341 7 interatomic distances for each of the above phases, as deter-mined by Rietveld refinement, are given in Table 3. Thermogravimetry Thermogravimetry (TG) was carried out on each of the phases Lno.33Sr,.67Ni04-a(Ln=La, Nd, Sm and Gd). The oxygen contents of these phases, as determined from these results, were 3.70( l), 3.69( l), 3.70(1) and 3.71(1) for Ln= La, Nd, Sm and Gd, respectively.The calculated value, assuming the Ni to be in the + 3 state, is 3.67. The oxygen content for the Ln= La sample suggests that Ni"' is present in both the main phase and the impurity Sr5Ni4011.A typical reduction profile, that for Gd0.33Sr1.67Ni04-dis shown in Fig. 3(a); the plateau indi-cates the presence of a stable intermediate compound contain-ing Nil. These Nil oxides, Ln0.33Sr1.67Ni02.67,have previously been isolated for Ln =Y and Dy-Tm.15y16 TG was also carried out on the compound GdSrNiO, [Fig. 3(b)], which exhibited a final mass of 93.6% correspond-ing to an oxygen content of 3.97(1). A plateau suggesting a stable intermediate is evident at a sample mass of 94.8% in the reduction profile of this sample.Magnetic characterisation The molar susceptibilities of the samples Ln0.33Sr1.67Ni04-a (Ln=La, Sm and Gd) and Nd2-,SrxNi0,-a (x= 1.0, 1.1, 1.3, 1.4, 1.5 and 1.67)were calculated from magnetization measure-ments, made using a Quantum Design SQUID magnetometer. Table 3 Profile and structural parameters, and interatomic distances for Ln2-,Sr,Ni04-d (Ln =La, Nd, Sm and Gd, x = 1.67; Ln =Gd, x= 1.0 and 1.33), with e.s.d.s in parentheses La" Nd Sm Gd Gd Gd X 1.67 1.67 1.67 1.67 1.33 1.o celi data a/+ 3.8 198 (2) 3.8028( 1) 3.7977( 1) 3.7916( 1) 3.7882( 1) 3.7718( 1) 12.35 18(8) 12.3259( 3) 12.2998( 3) 12.2906( 3) 12.1883( 3) 12.2163 (4)44 VIA3 180.23( 3) 178.24( 1) 177.39( 1) 176.69( 1) 174.91( 1) 173.80( 1) R factors (YO) 8.4 3.6 3.2 2.7 2.4 2.1RWP 5.8 2.4 2.0 1.7 1.5 1.4RP 5.8 4.1 3.9 3.7 3.7 3.6RF atomic parametersb Ni Uiso/A2 0.010( 2) 0.005(1) O.O06( 1) 0.007 ( 1) O.O02( 1) 0.005( 1) Ln/Sr z 0.3596( 2) 0.3589( 1) 0.3592( 1) 0.3592( 1) 0.3604( 1) 0.3610( 1) uiso18 -0.005(1) 0.004( 1) 0.005(1) 0.006( 1) 0.002( 1) O.O05( 1) O(1 Uiso/A2 -0.005( 1) 0.004( 1) 0.004( 1) 0.006( 2) 0.009( 2) 0.014( 2) occupancy 0.90(2) 0.88(1) 0.91 (1) 0.90( 1) 1.o 1.o O(2) z 0.1554( 10) 0.1598( 4) 0.1598(4) 0.1594( 4) 0.1643( 4) 0.1682( 5) uiso/A2 -O.O05( 1) O.O04( 1) 0.09 ( 1) 0.013( 1) 0.012( 2) 0.026( 2) interatomic distances/A Ni-O( 1)( x 33) 1.910( 1) 1.901(1) 1.899( 1) 1.896( 1) 1.894( 1 )' 1.886( 1)' Ni-O(2)(x2) 1.919( 13) 1.970(5) 1.966( 5) 1.959( 5) 2.OO3( 5) 2.055 (6) mean Ni-0 1.91 3( 8) 1.927( 3) 1.924( 3) 1.920( 3) 1.930( 3) 1.942( 3) Ln/Sr-O( 1)( x 33) 2.580( 1) 2.577( 1) 2.570( 1) 2.567( 1) 2.546( 1)' 2.538( 1)' Ln/Sr-0(2)( x 1) 2.523( 13) 2.454( 5) 2.453( 5) 2.456( 5) 2.390( 5) 2.355(6) Ln/Sr-0(2)( x 4) 2.707( 1) 2.699( 1) 2.696( 1) 2.691( 1) 2.696( 1) 2.691 (1) mean Ln/Sr-0 2.634( 5) 2.621 (2) 2.616( 2) 2.613 (2) 2.595( 2) 2.586( 2) "X-Ray data 10<228<70"; integration time 1 s.Sample contains some Sr,Ni4OIl. bAtomic positions: Ni (2a) O,O,O; Ln/Sr (4e) O,O,z; 0(1) (44 0.5,0,0; and O(2) (4e) O,O,z.Occupancies: Ni 1.0; Ln:Sr 1-x/2:x/2; and O(2) 1.0. 'Coordination to O(1) x 4. The molar susceptibilities (xM) us.temperature of Lno.33Srl.67Ni04-afor Ln=La, Sm and Gd are shown in Fig. 4 and the inverse molar susceptibilities (xM-') of 102 Nd, -,Sr,NiO, -are shown in Fig. 5. Each composition shows Curie-Weiss-like behaviour down to 6 K that can be well fitted 100 by the equation x~=x~+C~/(T-O).The xo term allows for contributions from excited states to the ,E, ground state of 98 Ni3+.I7 Values of xo, the effective moment (peff),and the Weiss parameter (0) for these compositions are given in Table 4. 96 Discussion94 Phase formation and solid solution range 92 Takeda and coworker^,^ amongst others, have investigated the h s. 901 La, -,Sr,Ni04 system (prepared at 1300 "C under 1bar,. I,, ,I, III "'I u) 0 200 400 600 800 1000 flowing 0,; samples with xbl.0 were annealed at 500"C, 3 150 bar O,), and found a solid solution range of Obxb 1.6.E 2. GdSrNi04 Attempts to form this phase for xa 1.6 are.not reported. The Q refined lattice parameters for the main phase in the samp!e 5 101 with nominal cqmposition Lao~33Srl,67Ni0, [a =3.8198(2) A, 100 c =12.3518(8) A] compare very well ewith those tor the x = 1.6 end member of their series (a= 3.82 A, c= 12.34 A).99 Takeda and coworkers also conducted a study of the crystal 98 chemistry of the Nd,-,Sr,NiO,-, system,6 formed by the decomposition of the corresponding nitrates, heating the resi- 97 dues at 1200 "C for 24 h under flowing O,, and annealing at 600 "C under 150 bar 0, (x21.2). They attempted to synthesize 96 samples with nominal compositions up to x= 1.8, and found 95 a limit to the solid solution at x=1.4.The investigation described in this paper has found a more extensive range of94 solid solutions for Nd, -,Sr,Ni04 -up to x =1.67. This may 93 be in part due to the slightly different conditions used in the production of these samples. 0 200 400 600 800 1000 temperature/"C The solid solutions Sm2-,Sr,Ni04-a (0.5dxd1.0) and Gd2-,Sr,Ni0,-a (x= 1.0) have been prepared in air at 1300 "C Fig. 3 TG reduction profiles for (a) Gd0.33Sr1.67Ni04-d and by Chen et aL7 We find that single-phase solid solutions up to (b) GdSrNiO, x= 1.67 can be prepared in both systems. Their lattice param- J. Muter. Chem., 1996, 6(l),57-62 59 La0.3 3sr 1 .67Ni04 -6 ""I'"'I""l""I""I"" 0.002 0.001 0.000 0 50 100 150 200 250 300 Gd0.33Sr1 mNi04 -6 0.25 0.20 0.15 0.10 0.05 0.00 0 50 100 150 200 250 300 temperature/K Fig.4 Molar susceptibility (xM) us. temperature for (a) La0.33Sr1.67Ni04-d, (b)Sm0.33Sr1.67Ni04-d and (c)Gd0.33Sr1.67Ni04-6 eters for GdSrNiO, [a =3.7652( 2) A, c =12.2057( 8) A] ate comparable wijh those found in this study [a= 3.7718( 1)A, c= 12.2162(3)A]; and a trend of increasing unit-cell dimen- sions with Sr content is observed for the x =1.33 and x =1.67 members (Table 3). An approximate phase diagram of the Ln, -,Sr,NiO, system at temperatures 1100-1300 "C can be generated, based on the results of the above studies, as well as on those reported by Demazeau et a1." in their investigation of the phases LnSrNiO, (Ln=La, Nd, Sm, Eu and Gd).Fig. 6 shows such a plot, in which synthesized members are marked on a phase field of rare-earth-metal ionic radius'' vs. Sr content (x). The mixtures of compounds present outside the phase field of the Ln, -,Sr,NiO, -solid solution are also given. Fig. 6 indicates the following. (i) A range of solid solutions [Ln,-,Sr,Ni0,-6 (Ln=Nd, Sm and Gd)] has been shown to exist for the larger +3 rare-earth-metal ions up to x =1.67. The end member for 60 J. Muter. Chem., 1996, 6(l), 57-62 400-I 25 300-1z 7 200ri' 100 n-0 50 100 150 200 250 temperature/K Fig. 5 Inverse molar susceptibility (xM-') us. temperature for Nd, -,Sr,Ni04 -a( 1.Odxd 1.67) the Ln=La solid solution appears to be xx 1.6.These phases are usually tetragonal; however, the low Sr-doped members of the solid solutions for Ln =La,3 NdY6 Pr and Sm7 have been observed to crystallize at room temperature with an ortho- rhombically distorted structure, in space group Brnab. (ii) Samples containing the smaller +3 rare-earth-metal ions Ln =Y, Dy-Tm (0.994-1.027 A), form essentially point phases Lno.33Srl.67Ni03.6-,(LnSrSNi3011).8'9 (iii) Samples containing +3 rare-earth-metal ions smaller than Tm do not form a K,NiF,-type structure, but give a mixture of compounds, including the rhombohedra1 phase LnSr3Ni06.10'11 (iv) Compositions on the Sr-deficient side of the Ln, -,Sr,NiO,-d solid solution region give rise to secondary phases Ln,SrO, and NiO.Preparations on the Sr-rich side contain SrO and NiO, which react in slow-cooled samples to give SrSNi,Oll.l4 Formation of the phases Ln, -,Sr,NiO, can be supported by tolerance factor arguments. Poix2' has defined a tolerance factor (t) for K,NiF, structures at t=a,/2bb, where a, and Pb are the Ln-0 and Ni-0 distances in nine- and six-fold coordination, respectively. Over the phase field depicted in Fig. 6, t varies from 0.867 (Nd2Ni04) to 0.966 (Nd0.33Sr1,67Ni03,67)(based on the ionic radii of Shannon)," all values of which fall within the range 0.85 <t <1.02 observed for stable K,NiF, structures.20 While t for Nd2Ni0, (0.867) is not far above the limiting value for K2NiF, formation, each of the other lower-limit compositions show tolerance factors that would suggest a greater range of solid solution than is observed (e.g.Sml,sSro.5Ni04has t =0,900, GdSrNiO, has t = 0.941 and Dy0.33Sr1.67Ni03.67 has t =0.961). With the exception of Ln=La, the upper limit of solid solution is constant, and appears to be controlled by the vacancy model we have discussed else~here.~,' Crystal structure Rietveld refinement of the structures of Lno.33Srl,67Ni04-6 (Ln =La, Nd, Sm and Gd) and Gd2 -,Sr,NiO, (x =1.0 and 1.33) confirms that they adopt the tetragonal K,NiF,-type structure in space group I4/rnrnm. The fit to the Lao.33Sr,.67Ni04-, profile is poorer than the others, due to the restricted data range and the presence of Sr,Ni,O,,, the structure of which is ~nkn0wn.l~ This results in physically unrealistic negative temperature factors for some atoms. The Lno~33Srl~6,Ni04 cell parameters and volume (Table 3) decrease with decreasing Ln3 + radius.No ordering between Ln3+ and Sr2+ ions was noted for these phases despite the large contrast in X-ray scattering factors; this feature has also been observed in the later members (Ln=Dy-Tm, Y) of this serie~.~?~Rietveld refinement also reveals oxygen vacancies at the O(1) site of these structures, although at a slightly lower concentration than expected from TG data. This conflict Table 4 Values of xo, observed molar effective magnetic moment, pobr,calculated Ln3+ contribution, p(Ln3+), and Weiss constant, 8 for Lno.33Srl,67Ni04-6(Ln=La, Sm and Gd) and Nd,-xSr,Ni04-6 (x=l.O, 1.1, 1.3, 1.4, 1.5 and 1.67) composition La0.33Sr1.67Ni04-6 Sm0.3,Sr1.6,Ni04 -6 Gd0.33Sr1.67Ni04-6 NdSrNiO, -Nd0.9sr1.1Ni04-6 Nd0.7Sr1.3Ni04 -6 Nd0.6Sr1.4Ni04 -6 NdosSr1.5Ni04-6 Nd0.33Sr1.67Ni04-6 Xo/emu mol-' POtJSlPB /-4Ln3+Y/pB O.O002( 1) 0.56( 1) 0 -8.1(1) 0.0006( 1) 0.74( 1) 0.49b -7.6( 1) 0 4.84( 1) 4.58 -3.9(1) O.O022( 1) 2.62( 1) 3.62 -11.8(1) O.O031( 1) 2.27( 1) 3.43 -6.3(1) O.O033( 1) 1.95( 1) 3.03 -2.5(1) O.O026( 1) 1.70(1) 2.80 -1.7( 1) 0.001 9 ( 1) 1.49(1) 2.56 -1.6( 1) 0.001 1 (1) 1.54( 1) 2.08 -4.1(1) a Calculated lanthanide moment p(Ln3+), =(2 -x)pJ2, where pJ=g,(J[J + 1])1'2 for ground state J, L and 5' quantum numbers. First excited state of Sm3+ is thermally accessible.Experimental values are typically in the range 0.80-0.98 pB for 0.33 mol Sm3+ at room temperature and fall as the temperature is reduced. 1.16 55 1.12 ah 0.-5 1.08 .-E 1.04 r5$ 1.00 0.96J"""""""""" 0.0 0.4 0.8 !!? 1.2 1.6 2.0 Sr content,x Fig. 6 Approximate phase diagram for the Ln, -,SrxNiO, -system appears to result from correlations between the refined occu- pancies and Debye-Waller factors, which may reflect local disorder around the oxygen sites due to the size discrepancy between Sr2+ and Ln3+, as well as the usual insensitivity of X-ray diffraction to the light oxygen atoms. Oxygen content Samples of Ln,-,Sr,NiO,-, tend to contain excess oxygen up to 6 z0.1 for x= 0 and 0.l.3,6For x= 0.2-1 .O, 6 is found to be close to zero with Ln= La-Gd.3,6,7 Our TG data for x= 1.67 samples with La-Tm, Y show that 6= -0.33.Hence, the oxygen contents of Ln, -,Sr,NiO, + ,solid solutions prepared under 1 bar O2 vary as follows. At x=O, 6>0 as interstitial oxygens lead to the oxidation of Ni2+ to Ni3+. As x increases to 1.0, 6 rapidly declines to zero. For x> 1, oxygen vacancies are formed with 6=( 1-x)/2 as Ni3+ is the highest oxidation state normally attainable under 1 bar 02. Reduced phases The phases Lno.33Srl.67Ni04- (Ln = La, Nd, Sm and Gd), all show the same mechanism of reduction as was observed for the smaller lanthanides (Y, Dy-Tm) with the same intermediate compo~ition:~.~ (i) the reduction of the Ni3+ in the sample to Ni': Ln0.3,Sr1 .67Ni1''03.67 jLn0.33sr1 .67Ni'02.67 (ii)complete reduction to Ni metal, accompanied by decompo- sition of the phase: Lno~33Sr,~67Ni'02~67+$Ln203 +@rO + Ni The mode of reduction of GdSrNiO, [Fig.3(b)] differs from that of the Lno~33Sr,~67Ni0,-, compounds. A consistent mech- anism is: (i) reduction of Ni"' to Nit in the primary phase accompanied by the ex-solution of Gd203 and Ni metal up to the maximum x (= 1.67) of the solid solution: GdSrNiO, -+~d,.,,Sr,.,,Ni'O,.,, + $Gd203+ $Ni (calculated sample mass = 94.8%) (ii) complete decomposition with reduction to Ni metal: gGdo.33Srl.67Ni'02.67+&Gd203 4-SrO+ SNi (calculated sample mass = 93.5%). Experimentally, a plateau corresponding to the stable Ni' intermediate Gd0.33Sr1.67Ni02.67 is observed in this reduction profile at a sample mass of 94.8(1)%, and the final mass is 93.6( 1)Yo. This reduction provides further evidence that x= 1.67 marks the upper limit of Sr substitution in these phases.Magnetic properties None of the phases Lno,33Srl.67Ni04-, (Ln=La, Nd, Sm and Gd) show any evidence of a transition to long-range magnetic order. The observed effective moment for Ln = La of 0.56( l)pB is significantly lower than expected for low-spin Ni3+ (1.73 pB). This value is also much lower than those reported for other low-spin Ni"' compounds such as SrLaA10,5Nio~,0, (1.52 pB), SrLaGao~,Nio~,O, (1.48 pB) and La,Lio.,Nio,504 (1.93 PB).l7 This low moment may reflect either strong antiferromagnetic coupling with J/k >> 300 K or delocalisation of some Ni3 + spins within small regions of the nickel oxide planes, but the high concentration of oxygen vacancies prevents either long- range magnetic order or metallic conduction from occurring.Similar magnetic behaviour is observed in semiconducting Yo.33sr1.67Nio3.67(~~ffz0.6PB, 6= -10.8 K).9 For Ln0.33Sr1.67Ni04 -,phases containing paramagnetic Ln ions (Table4) the magnetic moment due to Ni"' cannot be estimated accurately. The observed moments for the Ln = Sm and Gd members are close to those predicted (the thermal population of higher states for Sm3+ have not been accounted for). The observed effective moments for the series Nd, -,Sr,NiO, -( 1.0< x< 1.67) generally follow the expected trend of decreasing with increasing Sr2+ content; however, their values are lower even than the predicted contribution for Nd3+, due to crystal-field effects for this ion.Conclusions The series of new nickel oxides Ln0.33Sr1~67Ni0,-, has been extended by the synthesis and characterisation of the members Ln=Nd, Sm and Gd. These phases are analagous to those described previously for Ln = Y, Dy-Tm, but differ in that they are at the upper end of a substantial solid solution range, as opposed to existing at essentially a single composition. The upper limit for Ln=La appears to be Lao.4Srl,6Ni04-,. Rietveld refinement using powder X-ray diffraction data shows that each of these phases crystallizes with a tetragonal (T-type) K2NiF, structure, in space group I4/rnrnrn.Thermogravimetry J. Mater. Chem., 1996, 6(l), 57-62 of the members of the series Ln0.33Sr1.,,Ni0,-, (Ln=Nd, Sm, and Gd) confirms that the nickel is present as Ni"', with one oxygen vacancy per rare-earth-metal ion. Susceptibility data over the range 6-300K from these materials show Curie- Weiss behaviour. References 1 J. Gopalakrishnan, G. Colsmann and B. Reuter, J. Solid State Chem., 1977,22,145. 2 K. Sreedhar and C. N. Rao, Muter. Res. Bull., 1990,25, 1235. 3 Y. Takeda, R. Kanno, M. Sakano, 0. Yamamoto, M. Takano, Y. Bando, H. Akinaga, K. Takita and J. B. Goodenough, Muter. Res. Bull., 1990,25,293. 4 B. W. Arbuckle, K. V. Ramanujachary, Z.Zhang and M. Greenblatt, J.Solid State Chem., 1990,88,278. 5 S. M. Doyle, M. P. Sridhar Kumar and D. McK. Paul, J. Phys.: Condens. Matter, 1992, 4, 3559. 6 Y. Takeda, M. Nishijima, N. Imanishi, R. Kanno, 0. Yamamoto and M. Takano, J. Solid State Chem., 1992,96,72. 7 S. C. Chen, K. V. Ramanujachary and M. Greenblatt, J.Solid State Chem., 1993,105,444. 8 M. James and J. P. Attfield, J. Solid State Chem., 1993,105,287. 9 M. James, J. P. Attfield and J. Rodriguez-Carvajal, Chem. Mater., 1995,7, 1448. 10 M. James and J. P. Attfield, J. Mater. Chem., 1994,4,575. 11 M. James and J. P. Attfield, Chem. Eur. J.,submitted. 12 H. M. Rietveld, J. Appl. Crystallogr., 1969,2,65. 13 A. C. Larson and R. B. Von Dreele, Los Alamos National Laboratory Report No. LA-UR-86-748,1987. 14 J. Lee and G. F. Holland, J. Solid State Chem., 1991,93,267. 15 M. James and J. P. Attfield, J. Chem. SOC.,Chem. Commun., 1994, 1185. 16 M. James and J. P. Attfield, Physica C, 1994,235-240 (11), 751. 17 S.-H. Byeon and G. Demazeau, Muter. Lett., 1991, 12, 158. 18 G. Demazeau, M. Pouchard and P. Hagenmuller, J. Solid State Chem., 1976,18,159. 19 R. D. Shannon, Acta Crystallogr., Sect. A, 1976,32, 751. 20 P. Poix, J. Solid State Chem., 1980,31,95. Paper 5/03683F; Received 8th June 1995 62 J. Muter. Chem., 1996, 6(l),57-62

ISSN:0959-9428    

DOI:10.1039/JM9960600057

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Metamagnetism in EuPdIn and EuAuIn Rainer Pottgen Max-Planck-Institut fur Festkorperforschung, Heisenbergstrasse 1, 0-70569 Stuttgart, Germany EuPdIn and EuAuIn were prepared by reaction of the elemental components in tantalum tubes. They crystallize with the TiNiSi structure in space group Pnma and with 2=4 formula units per unit cell. Both structures were refined from single-crystal X-ray data: a =748.30( 13) pm, b =447.20( 8) pm, c =853.50( 14) pm, V= 0.28562( 5) nm3, wR2 =0.0408,449 F2 values, 20 variables for EuPdIn; and a =755.2( 2) pm, b =472.2(1)pm, c =842.3( 2) pm, V= 0.3004( 2) nm3, wR2 =0.0470,524 F2 values, 20 variables for EuAuIn. Magnetic susceptibility measurements show Curie-Weiss behaviour above 40 K for both compounds with experimental magnetic moments of 7.6( 1)~B/Eu (EuPdIn) and 7.5( 1)~B/Eu (EuAuIn), which are close to that of the free Eu2+ ion, namely peff=7.94 ~B/Eu.At an external field strength of 0.01 T, EuPdIn and EuAuIn order antiferromagnetically at 13.0(5) and 21.0(5) K, respectively. Magnetization measurements at 5 K indicate metamagnetism with a magnetic moment of 5.9( 1) ~B/Eu at 5.5 T for both compounds. The critical fields amount to 3.1( 1) T for EuPdIn and 0.25(2) T for EuAuIn. Both compounds are good metallic conductors with specific resistivities of 23 (EuPdIn) and 96 psZ cm (EuAuIn) at room temperature. Recently, we reported on the crystal structures and physical properties of a series of equiatomic europium-transition metal- germanide~.'-~ We have now extended these investigations on the corresponding indium compounds.The crystal structure of EuPdIn had already been investigated by Cirafici et ~1.;~ however, these authors reported no standard deviations for the positional parameters and no atomic displacement param- eters. They assigned the TiNiSi type structure to EuPdIn and proposed an order between the palladium and indium atoms. Their refinement resulted in R =0.054 for 329 reflections with Fo>2a(Fo). We have redetermined the structure of EuPdIn. Our X-ray data unambiguously showed that the ordering between palladium and indium is opposite to the model proposed by Cirafici et aL6 We have also refined the structure of isotypic EuAuIn which confirms our ordering model deter- mined for the palladium compound.Herein we also report in detail on the magnetic susceptibility and electrical conductivity measurements of EuPdIn and EuAuIn. Magnetic susceptibility data for EuPdIn from 100 to 1OOOK have already been published by Cirafici et d6We have now investigated the low- temperature properties down to 4.2 K. Experimenta1 Starting materials for the preparation of EuPdIn and EuAuIn were ingots of europium (Johnson Matthey), gold wire (Degussa, diameter 1 .O mm), palladium powder (Degussa) and indium tear-drops (Johnson Matthey), all with stated purities >99.9%. The large europium ingots were cut into small pieces in a glovebox, so that there was no contact with the air prior to the reactions. The elemental components were mixed in the ideal 1 :1: 1 atomic ratio and sealed in tantalum tubes under an argon pressure of about 800 mbar.The argon was purified prior to reaction over molecular sieves, titanium sponge (900 K) and an oxisorb ~atalyst.~ The tantalum tubes were subsequently sealed in silica ampoules to prevent oxidation, initially heated at 1270 K for two days and then annealed for two more weeks at 970 K. The reactions resulted in compact buttons which could be separated readily from the tantalum tubes. For inductively coupled plasma-atomic emission spec-trometry' (ICP-AES) investigations, small pieces of the samples were dissolved in aqua regia and analysed in an ARL 3580 spectrometer. Modified Guinier powder patternsg of all samples were recorded with Cu-Kct, radiation using 5N silicon (a=543.07 pm) as an internal standard.The indexing of the diffraction lines was facilitated by intensity calculations" using the pos- itional parameters of the refined structures. The lattice con- stants (Table 1) were obtained by least-squares refinements of the Guinier powder data. Single-crystal intensity data were collected on a four-circle diffractometer (CAD4) with graphite-monochromated Ag-Ka radiation and a scintillation counter with pulse height discrimination. The magnetic susceptibilities of polycrystalline pieces were determined with a SQUID magnetometer (Quantum Design) between 4.2 and 300 K with magnetic flux densities up to 5.5 T. The specific resistivities were measured on small blocks ( 1 x 1x 1.6 mm3 for EuPdIn and 1 x 1.2x 1.2 mm3 for EuAuIn) with a conventional four-probe technique. The cooling and heating curves measured between 4.2 and 300K were essen- tially identical, even for different samples.Results Powders and single crystals of EuPdIn and EuAuIn are light grey and stable in air over long periods of time. No decompo- sition whatsoever was observed after several months. Single crystals exhibit metallic lustre, and they have an irregular platelet-like shape. All Guinier powder patterns showed single-phase products. The elemental analysis of the EuPdIn and EuAuIn samples with the ICP-AES technique gave the following atomic per- centages: Eu :Pd :In, 31.4( 7) :34.6( 7) :34.0( 7); and Eu :Au :In, 31.0(7): 34.3(7): 34.7( 7), which are close to those calculated for the ideal composition (Eu :T :In, 33.3 :33.3 :33.3). No impurity elements were observed.Lattice constants The lattice constants for EuPdIn and EuAuIn are given in Table 1. The values for EuPdIn refined in the present work are in agreement with the data given earlier by Cirafici et The cell volume of the gold compound is about 5% larger than the cell volume of the palladium compound. This is related to the difference in size between Pd and Au: the metallic radius of Au (144.2 pm for CN 12; CN=coordination number)" is somewhat larger than the metallic radius of Pd (137.6 pm for CN12)." J. Muter. Chem., 1996, 6(l),63-67 Table 1 Lattice constants of the orthorhombic compounds with TiNiSi structure (standard deviations in parentheses) compound alpm b/Pm c/Pm v/nm3 ref.EuPdIn 748.30( 13) 447.20(8) 853.50( 14) 0.28562( 5) this work EuPdIn" 748.0( 2) 447.0( 1) 853.0(2) 0.2852( 2) this work EuPdIn 747.4 446.4 852.9 0.2846 6 EuAuIn 755.2( 2) 472.2( 1) 842.3 (2) 0.3OO4( 2) this work EuAuIn" 755.9( 2) 472.3 (2) 840.4( 1) 0.3000( 1) this work These values were obtained on the four-circle diffractometer. Table 2 Crystal data and structure refinement for EuPdIn and EuAuIn empirical formula EuPdIn EuAuIn formula mass/g mol -' 373.18 463.75 T/K 293(2) 293(2) wavelength/pm 56.086 56.086 crystal system orthorhombic orthorhombic space group Pnma (no. 62) Pnma (no. 62) unit-cell dimensions see Table 1 see Table 1 z 4 4 calculated density/g cm-3 8.68 10.26 crystal size/pm3 75x125~225 40 x 75 x 90 absorption correction from $-scan data from $-scan data transmission ratio (max :min) 1:0.739 1:0.365 absorption coefficient/mm- 18.85 41.36 F (000) 632 764 0 range/degrees 2-23 2-24 scan type 40 40 range in hkl +lo, +6, +11 +lo, +6, f12 total reflections 1683 1024 independent reflections 449 (Rint =0.0188) 524 (Rint =0.0285) reflections with I >241) 446 (R,=0.0111) 493 (R,=0.0291) refinement method full-matrix least-squares on F2 full-matrix least-squares on F2 data/restraints/parameters 449/0/20 524/0/20 goodness-of-fit on F2 1.462 1.203 final R indices [I >2a(I)] R 1 =0.0169, wR2 =0.0407 R1= 0.0192, wR2 =0.0463 R indices (all data) R1=0.0175, ~R2=0.0408 R1= 0.0224, wR2 =0.0470 extinction coefficient 0.0241 (9) 0.0101 (4) largest diff.peak, hole/e nm-' 882, -1941 1376, -1697 Structure refinements Table 3 Atomic coordinates and isotropic displacement parameters (pm') for EuPdIn and EuAuIn Single crystals of EuPdIn and EuAuIn were isolated from the crushed samples prepared in the tantalum tubes. They were Wyckoff examined by Buerger precession photographs to establish both atom site X Y Z Ueqe symmetry and suitability for intensity data collection. The photographs showed orthorhombic Laue symmetry mmm and EuPdIn 4c 0.03494(4) 1/4 0.67695(3) 118(1)Euthe extinctions were compatible with space group Pnma (no. Pd 4c 0.27153( 6) 1/4 0.37356( 5) 128( 1) 62), in agreement with the previous investigation.6 Some In 4c 0.14172(5) 1/4 0.06247(4) 111(1) crystallographic data and experimental details for the data EuAuIn Eu 4c 0.02553( 6) 1/4 0.68298(5) 141 (1) collection are listed in Table 2.4c 0.27882(4) 1/4 0.38271(4) 138(1)The atomic positions of EuPdIn as obtained from the Au 4c 0.15551(8) 1/4 0.06506(7) 124(2)refinement of Cirafici et aL6 were taken as starting values and In the structure was refined using SHELXL93,I2 with anisotropic a U,, is defined as one third of the trace of the orthogonalized atomic displacement parameters for all atoms. The refinement Uijtensor. readily converged to R1= 0.0404 and wR2 =0.1022 for all data. However, the isotropic displacement parameters of the pal- the same order between the transition metal and indium as in ladium and indium atoms had values of 64(4) pm2 and 173(4) the palladium compound.pm2, respectively. Furthermore, the standard deviations of the Final difference Fourier analyses revealed no significant atomic coordinates of these two positions were rather high. residual peaks. The results concerning the refinements are This indicated that the assignment of Pd and In had to be summarized in Table 2. Atomic coordinates and interatomic changed. We exchanged the Pd and In position and refined distances are listed in Tables 3 and 4. Listings of the anisotropic the structure again. The refinement converged within a few displacement parameters and the observed and calculated cycles to R1=0.0175 and wR2=0.0408 for all data.The structure factors are availab1e.i isotropic displacement parameters of Pd and In (see Table 3) were now in the same range and the standard deviations for Magnetic and electrical properties the positional parameters decreased by a factor of three. These The temperature dependence of the inverse magnetic suscepti- data clearly showed that the order between Pd and In in bility of EuPdIn between 100 and 1OOOK had already been EuPdIn is opposite to that proposed by Cirafici et aL6 The positions thus obtained were taken as starting values t Details may be obtained from: Fachinformationszentrum Karlsruhe, for the refinement of the EuAuIn structure. This refinement D-76344 Eggenstein-Leopoldshafen, by quoting the registry numbers easily converged to the residual listed in Table2 and shows CSD-404044 (EuPdIn) and CSD-404043 (EuAuIn). 64 J.Muter. Chem., 1996, 6(l), 63-67 Table 4 Interatomic distances (pm), calculated with the powder lattice constants, in the structures of EuPdIn and EuAuIn [all distances shorter than 470 pm (Eu-In, Au-In, Pd-In), 440 pm (Eu-Eu, In-In) and 410 pm (Eu-Au, Eu-Pd, Au-Au, Pd-Pd) are listed]" EuPdIn EuAuIn Eu: 1 Pd 313.7 Eu: 1 Au 317.1 2 Pd 314.8 2 Au 325.4 2 Pd 323.2 2 Au 334.1 1 In 338.6 1 In 336.5 2 In 342.0 2 In 345.6 2 In 343.6 1 In 348.9 1 In 358.2 2 In 351.6 2 Eu 379.4 2 Eu 390.2 2 Eu 394.4 2 Eu 394.1 Pd: 1 In 282.3 Au: 1 In 283.3 1 In 282.7 2 In 286.0 2 In 283.2 1 In 287.9 1 Eu 313.7 1 Eu 317.1 2 Eu 314.8 2 Eu 325.4 2 Eu 323.2 2 Eu 334.1 In: 1 Pd 282.3 In: 1 Au 283.3 1 Pd 282.7 2 Au 286.0 2 Pd 283.2 1 Au 287.9 2 In 326.1 1 Eu 336.5 1 Eu 338.6 2 Eu 345.6 2 Eu 342.0 1 Eu 348.9 2 Eu 343.6 2 In 350.6 1 Eu 358.2 2 Eu 351.6 a Standard deviations all GO.1 pm. measured by Cirafici et aL6 These measurements showed Curie-Weiss behaviour with an experimental magnetic moment peXp=7.99 ,uB/Eu and a Weiss constant of 0=40 K6We have investigated the magnetic properties of EuPdIn down to 4.2 K.Between 40 K and room temperature, the temperature depen- dence of the inverse magnetic susceptibility [see Fig. l(a)] is almost linear, indicating Curie-Weiss behaviour in agreement with the high-temperature data of Cirafici et uL6 The magnetic moment obtained from the linear portion of the 1/x us.T plot of pexp=7.6(1)~B/Eu compares well with the theoretical effec- tive moment peff=7.94 pBfor the free Eu2+ ion. The extrapol- ation of the linear relation of 1/x us. T resulted in a Weiss constant of @=13(1)K. Below 25 K the susceptibility of EuPdIn becomes dependent on the external magnetic field. At 0.01 T, the x us. T plot [see Fig. l(b)] shows a pronounced maximum at 13.0(5) K with, however, an upturn of the suscep- tibility below about 8 K. In order to analyse this behaviour in more detail, we also measured the magnetization us. external field behaviour at 5 K [see Fig.l(c)]. The magnetization measurement indicates metamagnetism. The transition from the antiferromagnetic to the ferromagnetic state occurs at the critical field strength of 3.1( 1)T. At the highest obtainable external field strength of 5.5 T the saturation magnetic moment amounts to p exp(sm) = 5.9(1) ~B/Eu, only slightly smaller than the calculated value of pcalc(sm)=7.0 ~B/Eu according to pcalc(sm)=g x J pB.13Similar values have also been determined for ferromagnetic EuAuGe [p exp(sm)=6.0(1)pB/EuI1and met- amagnetic EuZnGe [p exp (sm) =5.8( 1) pB/Eu].4 The temperature dependence of the reciprocal susceptibility of EuAuIn is shown in Fig. 2(a). Above 40 K, a linear increase of the inverse susceptibility is observed with increasing tem- perature, indicating Curie-Weiss behaviour.The experimental magnetic moment obtained from the slope of this plot amounts to 7.5( 1),uB/Eu; the Weiss constant has a value of 22( 1) K. At 21.0( 5) K, antiferromagnetic ordering is observed at an external magnetic flux density of 0.01 T [see Fig. 2(b)]. A magnetization measurement at 5 K reveals metamagnetism [see Fig. 2(c)] with a low critical field of 0.25(2) T. The saturation magnetic moment at 5.5 T amounts to 5.9 pu,/Eu. The specific resistivity of EuPdIn and EuAuIn decreases 8 0 -..*35 I E 0 50 100 150 200 250 300 TIK 5 t TIK ..................*.* Fig. 1 Magnetic properties of EuPdIn. (a) Temperature dependence of the inverse magnetic susceptibility between 25 and 300 K measured at a magnetic flux density of 3 T; (b) temperature dependence of the susceptibility between 4.2 and 30 K measured at 0.01 T; and (c) mag- netization us.external magnetic flux density, B,,,, at 5 K. with decreasing temperature as is typical for metallic conduc- tors (see Fig. 3). The room temperature values of the specific resistivity amount to 23 and 96 pi2 cm, respectively. In com- parison to the specific resistivities of 10.54 and 2.24 cm for metallic palladium and gold, respe~tively,'~ both ternaries may be classified as good conductors. Although both compounds show magnetic order at low temperatures, no anomaly was observed in either temperature dependency. This may be due to the polycrystalline character of the small platelets cut from the reaction ingots.Discussion The TiNiSi-type crystal structures of EuPdIn and EuAuIn have been refined from single crystal X-ray data. The correct J. Muter. Chern., 1996, 6(l), 63-67 .-/-•0' I I I I 1 J 0 50 100 150 200 250 300 TIK 4-0 r 3.0 1 I-0E 2.5 3 \s 2.0 4 1.5 L 1.a 0.5 0 5 10 15 20 25 30 35 40 45 TIK 6-v I h E9 5-m 3 !24-2 2l .s 3-. . . 1 I , I I I I 0 1 2 3 4 5 $xtfl Fig. 2 Magnetic properties of EuAuIn. (a) Temperature dependence of the inverse magnetic susceptibility between 40 and 300 K measured at a magnetic flux density of 1 T; (b) temperature dependence of the susceptibility between 4.2 and 40 K measured at 0.01 T; and (c) mag- netization us.external magnetic flux density, BeXt,at 5 K. ordering between the transition metal and indium atoms could be determined from these refinements, especially for EuPdIn, where the difference in the scattering power between palladium and indium is quite small. We observed the opposite ordering between Pd and In to that observed by Cirafici et aL6, and we could prove this model by the structure refinement of isotypic EuAuIn. Exactly the same results have also been obtained by Dascoulidou-Gritner and Schuster for CaPt1n.l' The main difference in both models is the occurrence of too short Eu-In distances (in the range of the Eu-Pd distances of Table 4) in the refinement of Cirafici et aL6 In contrast to the TiNiSi-type structure,16 the coordination of the transition-metal and indium atoms in EuPdIn and EuAuIn is changed, resulting in a maximal separation of the Pd and Au atoms, avoiding Pd-Pd and Au-Au bonding.As has been recently shown by extended Hiickel band calculations for structurally related CaAl,Si,-type inter metallic^,'^ the more electronegative atoms (i.e. Pd and Au in the present com-66 J. Mater. Chem., 1996, 6(l), 63-67 20 16 12 8 E u 0 so 100 160 200 260 300 95 loo r 90 80 70 60 0 50 100 150 a0 250 300 TIK Fig. 3 Temperature dependence of the specific resistivity of EuPdIn (a) and EuAuIn (b) pounds) occupy the energetically favoured positions with the maximum distance in the rhombus (see Fig.5). EuPdIn and EuAuIn are the first two compounds in the family of the equiatomic EuTIn intermetallics. In the corre- sponding family of germanium compounds (ref. 4, and refs. therein), a very rich crystal chemistry with seven different structures up to now has been observed. This is different for the indium compounds. EuPtIn'* also crystallizes with the TiNiSi structure and EuAgIn18 adopts the orthorhombic struc- ture of CeCu, with a statistical distribution of Ag and In on the copper position. This was also observed for EuAgGe., A projection of the EuAuIn structure is shown in Fig. 4 and a perspective view is given in Fig. 5.l' The europium atoms have coordination number (CN) 15 with 5 Pd(Au), 6 In and 4 Eu atoms in their coordination shell.The average Eu-Pd and Eu-Au distances of 317.9 and 327.2 pm, respectively, are somewhat smaller than the corresponding sums of the metallic radii (for CN 12)11 of 341.8 and 348.4pm. The same holds true for the average Eu-In distances, which amount to 344.7 0 n Eu 00 Au 0 0 Fig.4 Crystal structure of EuAuIn projected onto the xz plane. All atoms are situated on mirror planes at y= 1/4 and y =3/4, indicated by thin and thick lines, respectively. The trigonal [Eu,In,] prisms around the gold atoms are outlined. A A Fig. 5 Perspective view of the EuAuIn structure along the y axis. The structure is shown in the same orientation as Fig. 4. 0, Eu; 0, Au; 0,In. The 3D [AuIn] network is outlined. (EuPdIn) and 346.6 pm (EuAuIn) and are much shorter than the average Eu-In distances of 365.0 and 361.7pm in the binary intermetallics Eu,In (Co,Si-type) and EuIn, (own type), respectively.20 The average Eu-Eu distances in both com-pounds are in the same range as in the corresponding ger- manium The palladium and gold atoms have CN 9 with four indium and five europium neighbours.The average Pd-In and Au-In distances of 282.9 and 285.8 pm are significantly shorter than the sum of the metallic radii (for CN 12)" of 303.9 and 310.5 pm, respectively. The indium atoms are much larger than the palladium and gold atoms, as reflected by CN 12 with four Pd(Au), two In and six Eu neighbours. The most remarkable difference in the structures of EuPdIn and EuAuIn is the In-In distance within the rhombus of the three-dimensional TIn networks (see Fig.5). While the In-In distance of 326.1 pm in EuPdIn is even smaller than the average In-In distance of 333.5 pm in elemental indium,,' the In-In distance of 350.6 pm in EuAuIn is much larger and may not be considered as a bonding contact. I thank Prof. Dr. A. Simon for his interest and steady support of this work. I am also grateful to Dr. R. K. Kremer for helpful discussions, to W. Rothenbach for taking the Guinier powder patterns, to E. Briicher for the susceptibility measurement, to N. Rollbuhler for the electrical conductivity measurement, to 0.Buresch for the ICP-AES analysis and to Dr. W. Gerhartz (Degussa) for generous gifts of palladium powder and gold wire. The Stiftung Stipendienfonds des Verbandes der Chemischen Industrie supported my research by a Liebig grant.References R. Pottgen, J. Muter. Chem., 1995,5, 505. R. Pottgen, Z. Naturforsch. Sect. B, 1995,50, 1071. R. Pottgen, Z. Naturforsch. Sect. B, 1995,50, 1181. R. Pottgen, Z. Kristallogr., in the press. R. Pottgen, R. K. Kremer, W. Schnelle, R. Mullmann and B. D. Mosel, J. Muter. Chem., in the press. 6 S. Cirafici, A. Palenzona and F. Canepa, J. Less-Common Met., 1985,107,179. 7 H. L. Krauss and H. Stach, Z. Anorg. Allg. Chem., 1969,366,34. 8 G. L. Moore, Inductively Coupled Plasma-Atomic Emission Spectrometry, Elsevier, Amsterdam, 1989. 9 A. Simon, J. Appl. Crystallogr., 1970,3, 11. 10 K. Yvon, W. Jeitschko and E. Parthe, J. Appl. Crystallogr., 1977, 10,73. 11 E. Teatum, K. Gschneidner Jr., and J. Waber, Rep. LA-2345, US Department of Commerce, Washington, DC, 1960. 12 G. M. Sheldrick, SHELXL93, Program for Crystal Structure Refinement, University of Gottingen, Germany, 1993. 13 A. Szytula and J. Leciejewicz, Handbook of Crystal Structures and Magnetic Properties of Rare Earth Intermetallics, CRC Press, Boca Raton, FL, 1994. 14 Handbook of Chemistry and Physics, ed. R. C. Weast, CRC Press, Boca Raton, FL, 59th ed., 1978. 15 K. Dascoulidou-Gritner and H-U. Schuster, Z. Anorg. Allg. Chem., 1994,620,1151. 16 C. B. Shoemaker and D. P. Shoemaker, Acta Crystallogr., 1965, 18,900. 17 C. Zheng, R. Hoffmann, R. Nesper and H-G. von Schnering, J. Am. Chem. SOC., 1986,108,1876. 18 R. Pottgen, unpublished results. 19 E. Keller, SCHAKAL92, Kristallographisches Institut, Universitat Freiburg, 1993. 20 M. L. Fornasini and S. Cirafici, Z. Kristallogr., 1990, 190,295. 21 J. Donohue, The Structures of the Elements, Wiley, New York, 1974. Paper 5104859A; Received 24th July, 1995 J. Mater. Chew., 1996, 6( l), 63-67 67

ISSN:0959-9428    

DOI:10.1039/JM9960600063

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Preparation and electrical properties of KCa, -xLaxNb3010 Daisuke Harnada," Masahiko Machida," Yoshiyuki Sugahara"" and Kazuyuki Kuroda*"*b "Departmentof Applied Chemistry, School of Science and Engineering, Waseda University, Ohkubo-3, Shinjuku-ku, Tokyo 169, Japan bKagami Memorial Laboratory for Materials Science and Technology, Waseda University, Nishi-waseda, Shinjuku-ku, Tokyo 169, Japan Electron-doped layered perovskite KCa, -,La,Nb3Ol0 was prepared from KNbO,, Nb20s, Nb, La203 and KCa2Nb3010 at 1200"C for 10 h under Ar. Essentially single-phase KCa, -xLa,Nb,Olo with plate-like morphology was obtained with 0 <x d 0.3, and lattice parameters increased with increasing x. All the La-doped products showed semiconducting behaviour. Linear (log p)-T relationships were clearly observed, but the resistivity behaviour deviates from the linear relationship at lower temperatures.A possible conduction mechanism is discussed. Recently, a series of compounds M[An-1Bn03n+l] (M is the interlayer cation, A is the cation in the layer structures, B is Nb or Nb/Ti, and n describes the thickness of the perovskite slab), whose structures are closely related to the Ruddlesden- Popper phase, have been reported.' They consist of perovskite- derived slab layers and interlayer cations, hence they can be considered as two-dimensional (2D) layered perovskites. The n value ranges from 2 to 7,lp3 and MA,Nb3Ol0 (n=3) and MANb207(n=2) are typical niobates in the series. Electrical properties of conductive 2D oxides have attracted increasing attention since the discovery of high-T, super- conducting cuprates possessing layered structure~.~ Since the structures of such superconducting cuprates are related to the perovskite struct~re,~the electrical properties of layered perovskite-related oxides with various B-site ions are of inter- est.The electrical properties of various Ruddlesden-Popper phases have been investigated extensively (B-site: Ti,6 Fe,7 V,8 Ni,9 Ru," Ir," RhI2). Furthermore, variations in electrical properties and electron-doping by induced oxygen-deficiencies have been reported.13 Niobium-containing oxides can be metallic,14 semiconduct- ing" or even superconducting,16-18 when the average valence of niobium is lower than +5.Among various structures, an electro-conductive Nb-based perovskite-related phase AXNbO3 is known (A =Sr,19-21 Ca,,, Ba,23v24 Eu ,25-31 S~-EU,,~Sr-Ca34). Thus, the electrical properties of electron- doped layered perovskite niobates are of interest. So far, d-electrons have been doped into layered perovskite niobates by intercalation of excess amounts of M ions (M =Li,35 H,36 Rb37). However, none of these reports have described the electrical properties of the d-electron-doped compounds. Furthermore, as far as we know, no studies have been reported on electron-doping by solid-solution formation. We report here the preparation of the d-electron-doped layered perovskite KCa, -xLa,Nb3010 and its electrical proper- ties in the temperature range 4-280 K.Trivalent lanthanum is selected as a foreign ion because Goparakrishnan and co- workers reported the successful substitution of La3 + for Ca2+ in K, -,Ca2 -xLa,Nb30,038 and KCa,-,La,Nb, -xTix01039 (note that the valence of niobium in these two compounds was +5 and no d-electrons were doped). The structures and electrical properties of the products are discussed. Experiment a1 Polycrystalline KCa, -xLaxNb3010 were prepared from KCa2Nb3010, KNb03, Nb205, Nb and La203 by solid-state reactions according to eqn. (1): x/2 KNb03+2/5x Nb205+ 1/5x Nb +x/2 La203+ (1-x/2) KCa2Nb3010 -+KCa,-,La,Nb,O,, (1) KCa2Nb3010 was synthesized by the solid-state reaction of stoichiometric amounts of K,C03, CaCO, and Nb205 at 1100"C for 20 h in air.KNb0, was obtained by the solid- state reaction of stoichiometric amounts of K2C03 and Nb205 at 1000°C for 1h in air. La203 was dehydrated by heating at 21000"C for more than 1 h before use. KCa,Nb,O,,, KNbO,, Nb205, Nb and La203 were thoroughly mixed and pressed into a button pellet at 59 MPa. The pellet in an alumina boat was placed in an alumina tube with Ti powder (to remove oxygen). After evacuation (cu. 8.5 x lo-, Pa), the alumina tube was filled with Ar. Then the pellet was heated at 1200°C for 10 h under an Ar atmosphere at a heating and cooling rate of 5 "C min-'. Crystalline phases were identified by a Mac Science MXP3 diffractometer (monochromated Cu-Ka radiation). Lattice par- ameters were refined by the non-linear least-squares method from the positions of 21 peaks that were reasonably indexed.The amounts of potassium, calcium, lanthanum and niobium were determined by inductively-coupled plasma emission spec- troscopy (ICP; Nippon Jarrell Ash ICAP-575 11). The mor- phology of the products was studied with a scanning electron microscope (SEM) equipped with an electron probe micro- analyser (JEOL, JXA-8600). The resistivity measurements were performed using the standard four-probe method in the range 4-280 K. Results and Discussion Table 1 summarizes the compositional analysis results. The compositions of the products are essentially consistent with the corresponding nominal ones. The loss of potassium is possible, and therefore an excess amount of potassium was used for the preparation of the potassium-containing layered perov~kites.l*~~*~'In the present study, however, no obvious potassium loss was observed.We used an excess amount (10%) of potassium, but no effect on the phase purity was observed. X-Ray diffraction (XRD) patterns of the products are shown in Fig. 1. When Odxd0.3, the products are essentially single phase, but very trace amounts of LaNbO, and/or an unknown phase are also detected. If 0.4<x, LaNbO, is obviously detected, and its amount increases with increasing x. The structure of KCa,Nb301, is reported to be pseudotetragonal J. Muter. Chern., 1996, 6(l), 69-72 Table 1 Compositions of the products determined by ICP sample, x= composition"Yb 0 0.1 0.2 0.3 0.4 0.5 0.6 a Compositions were normalized by setting the amount of niobium at 3.Amount of oxygen was set at 10. I I x=o.a x =0.7 X =0.6 x =0.5 X =0.4 X =0.3 x =0.2 x =0.1 KCazNIJS010 10 20 30 40 50 60 28/degrees Fig. 1 Powder XRD patterns of KCaz-,La,Nb3Ol0. a, LaNbO,; V,unknown phase. orthorhombic,' and XRD patterns reveal that the symmetry is maintained in all the products. Uma and Goparakrishnan prepared the solid-solution K, -$a2 -xLa,Nb3010 and reported the appearance of new peaks and marked changes of peak intensity ratios.38 The almost complete retention of the XRD profile in the present system may be ascribed to the presence of a constant amount of potassium ions in the interlayer space.The peaks ascribed to the KCa,Nb3010 structure shift to lower angles as x increases. The variation in lattice parameters with x is demonstrated in Fig. 2. All the three lattice parameters increase with increasing x,and seem to level off with larger x. The substitution of La3+ (ionic radius 0.136nm) for Ca2+ (ionic radius 0.134 nm) should be accompanied by the simul- taneous reduction of the same amount of niobium [NbSf (ionic radius 0.064 nm)+Nb4+ (ionic radius 0.068 nrn)];,, hence, both substitution and reduction should cause the increase in the lattice parameters. Since no obvious composi- tional changes occur during the preparation, the observed 0.392 2.98 2.97 0.39 increase in the lattice parameters indicates solid-solution for- mation.The behaviour of the LaNbO, peaks and that of the lattice parameters allows us to assume that the x limit for solid-solution formation is between 0.3 and 0.4 under the present experimental conditions. In the K, -xCa,-,La,Nb3010 system, a similar lattice expan- sion was reported with increasing x, and the phenomenon was ascribed to the replacement of Ca2+ with La3+.38 The increments of the lattice parameters from x=O to x=OS in the two systems are as follows: for this study, Aa =2.5( 1)pm, Ab=2.4(9) pm, Ac= 13.(3) pm; for the K,-,Ca2-xLa,Nb30,0 system, Aa =3.5( 2) pm, Ab =4.6(2) pm, Ac =19.(0) pm. Thus, although the effect of the reduction of niobium (Nb5+ +Nb4+) was absent in the K, -,Ca, -xLa,Nb30,0 system, the increments of the lattice parameters are larger in the K, -xCa2-xLa,Nb,0,0 system.The morphology of the products is shown in Fig. 3. For x= 0.1, the products consist of plate-like particles (ca.2 pm x 2 pm) only [Fig. 3(a)], and electron probe microanalysis (EPMA) indicated a homogeneous distribution of elements (K, Ca, La, Nb); these observations are consistent with the formation of essentially single-phase products with x <0.3. In contrast, if x =0.8, granular particles [Fig. 3(b), indicated by arrow] are observed besides the plate-like products. EPMA showed that the granular products contain mainly La and Nb. Taking the XRD results into account, the granular products appear to correspond to LaNbO,, the evident impurity at x 20.4. KCa2Nb3OIo was a white insulator, and La doping resulted in blue products.The blue colour darkened with increasing x. The resistivity at room temperature decreased drastically as x increased from ca. 105-106 (x=O.l) to ca. 103-104 (x=0.4), and further decreased to ca. lo2 mi2 m (x=0.6). The observed values are 10-1000 times larger than the reported values of Nal-,Sr,Nb03,33 if the same amount of Nb4+ is present in the KCa, -xLa,Nb,O,o and Na, -,Sr,NbO, units. 0*3ss2m94 0'384 0 0.1 0.2 0.3 X 0.4 0.5 0.6 2.93 -10pm Fig. 2 Variation in lattice parameters with x. 0,a; 0,b; 4, c. Fig. 3 SE micrographs of the products with (a) x =0.1 and (b) x =0.8 70 J. Mater. Chem., 1996, 6(l), 69-72 100 t 1 80 120 160 200 240 280 TIK Fig. 4 Temperature dependence of normalized resistivity of the samples (0.1 <x <0.6) in the higher temperature region.x =0.1 (O),0.2 (0), 0.3 (A),0.4 (A),0.5 (O), 0.6 (0). All the La-containing products exhibited a semiconducting temperature-dependence of their electrical properties (Fig. 4). Instead of the Arrhenius-type relationship for a thermally activated hopping conduction mechanism, however, a linear relationship is observed between the logarithm of the resistivity, log p, and the temperature, T,over a large temperature range. The slope in Fig. 4 decreases with increasing x, but becomes constant at 0.4<x, which is in agreement with the range of solid-solution formation estimated from the XRD results. Other layered transition metal oxides such as A,Nb, + 3,@3, +3m,43 Ruddlesden-Popper phases6-13 and phos- phate tungsten bron~es~~,~~ are metallic or semiconducting, but none of them shows a linear (log p)-T relationship.On the other hand, similar resistivity behaviour was reported for some mixed-valence oxides such as Fe30446 and Na, -,Sr,Nb03,33 and a possible mechanism for such behav- iour has been proposed based on the incoherent tunnelling of electrons between the nearest-neighbouring sites with an oscil- lating barrier (small polaron model with a vibrating barrier).47 If a simple harmonic oscillator model (with frequency o)is applied, the slope in Fig. 4 corresponds to 2a2kB/rno2, where the parameter a is related to the expanse of the wavefunction [kB, Boltzmann's constant; rn, mass of involved ions (niobium) (= 1.5 x kg)].From Fig. 4, w/a values can be calculated, and by assuming a constant frequency o for all the products (o=lo1, s-'), l/avalues (so-called site localization parameters) can be estimated (Table 2). The decrease in slope in Fig. 4 thus corresponds to the increase in l/.. The o/a values are lower than those calculated similarly for the Na, -,Sr,NbO, system [263 (x=0.2)-514 (x=0.6)],,, but are consistent with the reported values for various oxides and chalcogenide~.~~ The resistivities of the products with relatively higher con- ductivities (0.3 <x) are measured down to 4 K (Fig. 5). The resistivity shows typical semiconducting behaviour at lower temperature, and linear relationship is not observed below cu.20 (x=0.6)-80 K (x=0.3). H~rd~~ proposed that the transfer rate R (which dominates the conductivity) in the vibrating Table 2 Parameters related to the temperature dependence of resis-tivity in the high-temperature region 0.1 91 0.091 0.2 105 0.105 0.3 123 0.123 0.4 169 0.169 0.5 174 0.174 0.6 174 0.174 w= 1x 10l2 s-l throughout. Fig. 5 Temperature dependence of resistivity of the samples (0.3 G x G0.6) in the lower temperature region. x =0.3 (0),0.4 (O), 0.5 (0), 0.6 (A). barrier model is expressed as: R zC x exp (-2aSo) exp (-U/k,T) exp (2a2kBT/rno2) where So is the equilibrium width of the barrier, U is the activation energy for the displacement of sites and C is a constant.The term exp(-2aSo) (the overlap term) is nearly independent of the temperature, thus the other two exponential terms dominate the temperature dependence. The coincidence term, exp (-U/kBT), reflects the thermally activated prob- ability for coincidence and the tunnelling term, exp(2a2kBT/rno2), is due to the oscillation of the barrier (as discussed above). At high temperatures where U<<kBT,the transfer rate is subjected to the tunnelling term to give a linear (log p)-T relationship. At lower temperatures, the coincidence term also contributes to the transfer rate, and the (log p)-T relationship deviates from linearity. Similar behaviour was reported for other mixed-valence oxides of Ti,O,, -149 and vno,, -50 1* Conclusions We have demonstrated the preparation, structure and electrical properties of KCa, -xLa,Nb,Olo.With 0 <x <0.3, essentially single-phase KCa, ~xLa,Nb,Olo is prepared from KNbO,, Nb205, Nb, La203 and KCa2Nb3OI0 by solid-state reactions at 1200°C for 10 h under Ar. No obvious loss of elements is observed. Lattice parameters increase with x,indicating solid- solution formation. All the La-doped products show semicond- ucting behaviour. Linear (log p)-T relationships are observed in a relatively high temperature range, but log p did not follow a linear relationship at lower temperatures. A possible expla- nation for the resistivity behaviour is proposed based on the vibrating barrier model which gives the linear (log p)-T relationship by the incoherent tunnelling of electrons between the nearest-neighbouring sites.These results indicates that d-electrons can be doped in the layered perovskite structures by solid-solution formation, and the electrical properties reflect the mixed valency of niobium in KCa, -xLa,Nb30,0. References 1 M. Dion, M. Ganne and M. Tournoux, Rev. Chim. Miner., 1986, 23, 61. 2 A. J. Jacobson, J. W. Johnson and J. T. Lewandowski, Inorg. Chem., 1985, 24, 3727; A. J. Jacobson, J. T. Lewandowski and J. W. Johnson, J. Less Common Met., 1986,116, 137. 3 R. A. Mohan Ram and A. Clearfield, J. Solid State Chem., 1991, 94,45. 4 J. G. Bednorz and K. A. Muller, Z. Phys. B, 1986,64,189. 5 C. N. R. Rao and B. Raveau, Acc. Chem. Res., 1989,22,106.J. Muter. Chern., 1996,6( l), 69-72 6 7 8 S. Hayami, H. Yamamoto, Y. Sugahara and K. Kuroda, in preparation. Y. Takeda, K. Imayoshi, N. Imanishi, 0. Yamamoto and M. Takano, J. Mater. Chem., 1994,4, 19. F. Deslandes, A. I. Nazzal and J. B. Torrance, Physica C, 1991, 28 29 30 K. Ishikawa, G. Adachi, M. Tanida and J. Shiokawa, Bull. Chem. SOC. Jpn., 1981,54, 159. K. Ishikawa, G. Adachi, M. Hasegawa, K. Sato and J. Shiokawa, J. Electrochem. SOC., 1981,128, 1374. K. Ishikawa, G. Adachi and J. Shiokawa, Bull. Chem. SOC.Jpn., 179, 85. 1982,553317. 9 Y. Takeda, R. Kanno, M. Sakano, 0. Yamamoto, M. Takano, 31 K. Ishikawa, G. Adachi and J. Shiokawa, Mater. Res. Bull., 1983, Y. Bando, H. Akinaga, K. Takita and J. B. Goodenough, Mater. 18,653. 10 Res.Bull., 1990,25,293. Y. Maeno, H. Hashimoto, K. Yoshida, S. Nishizaki, T. Fujita, 32 B. Ellis, J. P. Doumerc, M. Pouchard and P. Hagenmuller, Mater. Res. Bull., 1984,19, 1237. 11 J. G. Bednorz and F. Lichtenberg, Nature, 1994,372,532. R. J. Cava, B. Batlogg, K. Kiyono, H. Takagi, J. J. Krajewski, 33 B. Ellis, J. P. Doumerc, M. Pouchard and P. Hagenmuller, Solid State Commun., 1984,51,913. W. F. Peck, Jr., L. W. Rupp, Jr. and C. H. Chen, Phys. Rev. B, 34 K. Isawa, R. Itti, J. Sugiyama, N. Koshizuka and H. Yamauchi, 12 1994,49,11890. T. Shimura, M. Itoh, Y. Inaguma and T. Nakamura, Phys. Rev. B, 35 Phys. Rev. B, 1993,48,7618. R. Jones and W. R. McKinnon, Solid State ionics, 1991,45, 173. 1994,49,5591. 36 P. Gomez-Romero, M. R. Palakin, N.Casaii and A. Fuertes, Solid 13 I. S. Kim, M. Itoh and T. Nakamura, J. Solid State Chem., 1992, State Ionics, 1993,63,424. 14 15 101, 77. R. J. Cava, B. Batlogg, J. J. Krajewski, P. Gammel, H. F. Poulsen, W. F. Peck, Jr., and L. W. Rupp, Jr., Nature, 1991,350,598. R. J. Cava, B. Batlogg, J. J. Krajewski, H. F. Poulsen, P. Gammel, W. F. Peck, Jr., and L. W. Rupp, Jr., Phys. Rev. B, 1991,44,6973. 37 38 39 A. R. Armstrong and P. A. Anderson, Inorg. Chem., 1994,33,4366. S. Uma and J. Goparakrishnan, J. Solid State Chem., 1993, 102, 332. J. Goparakrishnan, S. Uma and V. Bhat, Chem. Mater., 1993, 5, 132. 16 17 18 19 20 21 22 23 24 25 26 M. J. Geselbracht, T. J. Richardson and A. M. Stacy, Nature, 1990, 345, 324. M. A. Rzeznik, M. J. Geselbracht, M.S. Thompson and A. M. Stacy, Angew. Chem., int. Ed. Engl., 1993,32,254. J. Akimitsu, J. Amano, H. Sawa, 0. Nagase, K. Gyoda and M. Kogai, Jpn. J. Appl. Phys. Part 2, 1991,30, L1155. D. Ridgley and R. Ward, J. Am. Chem. SOC., 1955,77,6132. B. Hessen, S. A. Sunshine, T. Siegrist and R. Jimenez, Mater. Res. Bull., 1991,26, 85. K. Isawa, J. Sugiyama, K. Matsuura, A. Nozaki and H. Yamauchi, Phys. Rev. B, 1993,47,2849. M. Hervieu, F. Studer and B. Raveau, J. Solid State Chem., 1977, 22,273. R. R. Kreiser and R. Ward, J. Solid State Chem., 1970,1, 368. M. T. Casais, J. A. Alonso, I. Rasines and M. A. Hidalgo, Mater. Res. Bull., 1995,30,201. G. J. McCarthy and J. E. Greedan, Inorg. Chem., 1975,14,772. J. P. Fayolle, F. Studer, G. Desgardin and B. Raveau, J. Solid State Chem., 1975,13,57. 40 41 42 43 44 45 46 47 48 49 50 J. Goparakrishnan, V. Bhat and B. Raveau, Mater. Res. Bull., 1987, 22,413. J. Goparakrishnan and V. Bhat, Inorg. Chem., 1987,26,4299. R. D. Shannon, Acta Crystallogr., Sect. A, 1976,32,751. J. Kohler, G. Svensson and A. Simon, Angew. Chem., Int. Ed. Engl., 1992,31,1437. J. P. Giroult, M. Goreaud, Ph. Labbe, J. Provost and B. Raveau, Mater. Res. Bull., 1981,16,811. E. Canadell, I. E-I. Rachidi, E. Wang, M. Greenblatt and M-H. Whangbo, Inorg. Chem., 1989,28,2455. A. J. M. Kuipers and V. A. M. Brabers, Phys. Rev. B, 1979,20,594. W. R. McKinnon, C. M. Hurd and I. Shiozaki, J. Phys. C, 1981,14, L877; I. Shiozaki, C. M. Hurd, S. P. McAlister, W. R. McKinnon and P. Strobel, J. Phys. C, 1981,14,4641. C. M. Hurd, J. Phys. C, 1985,18,6487. A. D. Inglis, Y. Le Page, P. Strobel and C. M. Hurd, J. Phys. C, 1983,16, 317. A. D. Inglis, C. M. Hurd and P. Strobel, J. Phys. C, 1984,17,6801. 27 F. Studer, G. Allais and B. Raveau, J. Phys. Chem. Solids, 1980, 41, 1187. Paper 5105248C; Received 7th August, 1995 72 J. Mater. Chem., 1996, 6(l),69-72

ISSN:0959-9428    

DOI:10.1039/JM9960600069

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Dehydroxylation sequences of gibbsite and boehmite: study of differences between soak and flash calcination and of particle-size effects Victoria J. Ingram-Jones," Robert C. T. Slade,"" Thomas W. Davies,b Jennifer C. Southern" and Sylvain Salvadord "Departmentof Chemistry, University of Exeter, Stocker Road, Exeter, UK EX4 4QF bDepartmentof Chemical Engineering, University of Exeter, Stocker Road, Exeter, UK EX4 4QF "Alcan Chemicals Limited, Chalfont Park, Gerrards Cross, Buckinghamshire, UK SL9 OQB dEntreprise Malet, 30 Avenue de Larrieu, 31081 Toulouse Cedex, France The thermal transformation sequences of boehmite (y-AlOOH) and two grades of gibbsite [y-Al(OH),] upon soak and flash calcination are reported. The techniques used were X-ray diffraction (XRD), differential thermal analysis (DTA), Fourier- transform IR (FTIR) and 27Al magic-angle spinning (MAS) NMR spectroscopies.Boehmite undergoes the dehydroxylation sequence boehmite, y, 6,8, a-Al,O, under both soak and flash calcination. The dehydroxylation sequence of gibbsite, however, depends on the calcination method and the particle size of the feed material. Soak calcination of a fine gibbsite (ca. 0.5 pm) gave the dehydroxylation sequence gibbsite, x, K, a-A1203; with flash calcination the sequence gibbsite, x, y, 6,8, a-A1203 was observed. Soak calcination of coarse gibbsite (ca. 14 pm) gave both the dehydroxylation pathways (a):gibbsite, boehmite, y, 6,8, a-A1203and (b):gibbsite, x, K, a-A1203, and pathway (a)was predominant. Flash-calcined coarse gibbsite experiences a crossover between these routes (x-y) without formation of K-A~,O,.Flash calcines of gibbsite undergo this x-y phase change at ca. 800 "C. Gibbsite [y-Al(OH),] and boehmite (y-A100H) are commer- cially important materials marketed both as dried products and as activated aluminas which are produced by heating. Gibbsite is used, for example, as a flame retardant, paper additive and as a feedstock in the manufacture of other aluminium chemicals, including aluminium metal. Boehmite is used in the production of ceramic pieces and catalyst supports. Dehydroxylation (during calcination) of gibbsite and boehmite leads firstly to one or more transition aluminas with partially disordered structures. The activated alumina phases (x, K, y, 6, 8, etc.) are often used in catalytic and adsorption roles because of their unique surface chemistry; they possess a large surface area and large internal porosity.As the calcination temperature increases, the structures become more ordered until the final transformation to the extremely stable corundum (a-Al,03) form, which is itself used in abrasives, refractories, polishing, aluminous porcelain, technical and engineering cer- amics and catalyst supports. The crystal structures of gibbsite, boehmite and corundum are well known.'-3 Gibbsite is monoclinic (space group P2& a =8.64 A, b =5.07 A, c =9.72 A, p= 85"26').' The structure contains double layers of hydroxide ions [each layer is hexag- onal close packed (hcp)] with aluminium atoms in octahedral coordination inside the layers in a pattern of hexagonal rings.The double layers stack to give an ABBA anion seqvence. Boehmite is orthyrhombic (space group Cmcrn: a=2.87 A, b= 12.20 A, c =3.69 A)., The structure contains octahedrally coor- dinated aluminium atoms, with the octahedral units linking to form complex layers. Hydrogen atoms are present as interlayer hydroxy groups (i.e.attached to oxygens at the top and bottom of layers). The oxygens are approximately cubic close packfd (ccp). Cor!ndum is rhombohedra1 (space group R3c:a =4.76 A, c =12.99 A).3 The structure contains oxygen in an approximate hcp arrangement with aluminium atoms in two-thirds of the octahedral sites. The structures of the poorly crystalline, transition aluminas are less well known.Characteristic structures are believed to be as follows: x-A1203 and K-A~,O, are thought to contain hcp oxygen^.^ x-A1203 has a very diffuse powder X-ray diffrac- tion (XRD) pattern and contains a large number of stacking faults. y-A1203, 6-Al,03 and 8-A1203 are thought to be spinel- related with A1 atoms in both tetrahedral and octahedral site^.^.^ y-Al,03 may be ordered or disordered with respect to the aluminium arrangement, depending on the production conditions. Upon the calcination of gibbsite and boehmite, the proportion of tetrahedrally coordinated A1 atoms increases with calcination temperature and the A1 atoms are disordered over octahedral and tetrahedral sites. As the calcination tem- perature increases further, there is a progressive filling of octahedral sites until corundum is formed and all the A1 atoms are in octahedral environment^.^ Many of the previous of the thermal transformation sequences of highly crystalline gibbsite and boehmite samples have used soak calcination methods, i.e.the materials were raised to the reaction temperature at a slow rate and maintained there for a period ranging from minutes to hours. There is still disagreement on the sequence of formation of transitional alumina phases, their structures and the mechanism of dehydration. It is generally acceptedlg that boehmite transforms in air to give a-Al,O, uia y-A120, (ca. 450"C), 6-A1,0, (ca. 750°C) and 8-A1203 (ca. 1000°C).Gibbsite follows two paths: (a) uia boehmite (60-3OO0C), y-Al,O, (500-850 "C), 6-Al2O, (850-1050 "C) and 8-Al2O, ( 1050-1 150"C); and (b)uia x-A1203(300-500 "C) and K-A~,O, (800-1 150 "C).Both pathways are followed simultaneously to eventually give a-Al,O,. The extent to which each of these paths is followed depends upon a number of factors (including gibbsite particle size, moisture, heating rate, pressure, bed depth and the heating method). The generally accepted dehy- droxylation sequences of boehmite and gibbsite when soak calcined in air are summarised in Fig. 1 and 2, respectively. In addition to the soak calcination method, the dehydroxyl- ation sequences resulting from flash calcination can be ana- I I I 1 I I 0 200 400 600 800 lo00 1200 TIT Fig.1 Thermal transformation sequence for boehmite in air. Transition aluminas are denoted by the appropriate Greek symbols. J. Muter. Chem., 1996, 6(l),73-79 gibbsite +x bK-a (b) I I I I I I 200 300 600 800 loo0 1200 Fig. 2 Thermal transformation sequences for soak calcination of gibbs- ite in air lysed. A text concerning the flash calcination processing of minerals has recently been published.20 This technique involves heating a powdered solid very rapidly (in a fraction of a second) by plunging it into a stream of hot, flowing gas. The solid is maintained at the gas temperature for a short time (usually between fractions of a second and one minute) and then quenched to room temperature.Control of the process variables in flash calcination permits time resolution of the flash calcination process and results in calcines kinetically frozen at different stages of the reaction sequence. This study interrelates the results obtained from XRD, DTA, FTIR and 27Al MAS NMR studies carried out on soak and flash calcines of boehmite and gibbsite, and consequently reveals the different dehydroxylation sequences they follow under differing process variables. Experimenta1 Materials Highly crystalline boehmite (BACO Cera Hydrate) and two gibbsite samples (BACO FRF20 and BACO UF35E SD) were examined (materials from Alcan Chemicals Europe). FRF20 has a median particle size of cu. 14 pm, with that for UF35E SD being ca. 0.5 pm (determined by laser-light scattering).Calcination Soak calcination studies involved calcining gibbsite and boehmite, at a bed depth of 1cm in platinum crucibles, in a muffle furnace at temperatures between 300 and 1200°C with a heating rate of 200°C h-l, holding at the reaction tempera- ture for a period of 2 h and quench-cooling to room tempera- ture by removing the crucibles from the furnace whilst at the reaction temperature. Flash calcination was carried out in two different calciners, viz. the Exeter calciner EXON,,' which is electrically heated, and the propane-fired Toulouse calciner INSA.,, In the Exeter flash calciner (Fig. 3) the powdered starting material was carried into the furnace by an N, gas stream where it was rapidly heated (4700-15000°C s-') to a set reaction tempera- ture (1200°C maximum) by the hot N2 gas flowing through the reaction chamber.The powder was heated during the short time (residence time) taken for it to reach the adjustable water- cooled collection probe. Reaction quenching was achieved by injection of cold N, gas through the porous walls of the collection cup. Boehmite and gibbsite samples were recycled through the calciner until series of samples with residence times from 2.5 s (a single pass) to 10 s (4 passes) had been obtained for reaction temperatures over the range 900-1200 "C. The INSA laboratory flash calciner (Fig. 4) is described in detail elsewhere.,, Air was used as the carrier gas and was heated in the refractory tube by a propane burner.The powdered material was injected into the base of the reactor tube using compressed air. The suspension of powder particles and combustion gases was pulled through the reactor tube by an extraction fan to the exit, where it was mixed with a stream of quench air. The residence time in the reactor tube was determined by the gas flow rate. In these experiments the Fig. 3 A schematic diagram of the Exeter flash calciner (EXON). 1, Carrier gas inlet; 2, powder inlet; 3, vibrator; 4,screwfeed mechanism; 5, ceramic flow-straightener; 6, water-cooled feeder probe; 7, reaction chamber; 8, heating rod; 9, cold furnace gas inlet; 10, collection cup; 11, water-cooled collection probe; 12, quench gas; 13, vacuum. Fig. 4 Construction of the INSA flash calciner.1, Refractory steel calcining tube; 2, silica/alumina wool insulation; 3, propane burner; 4,clay suspension generator; 5, conveyor belt; 6, vibration corridor; 7, compressed air; 8, refractory steel injection tube; 9, calcine collecting filter; 10, exhaust fan; 11, air flow meter; *, temperature measurement points. 74 J. Muter. Chem., 1996, 6(l), 73-79 boehmite and gibbsite samples were passed through the INSA flash calciner for a duration of 0.6 s at reaction temperatures in the range 630-1030°C. Characterisation Powder X-ray diffraction (XRD) profiles were recorded using a computer-controlled diffractometer (rhilips PW 1050 goni- ometer, Cu-Ka radiation, I. =1.54178 A) incorporating data accumulated over multiple scans.Thermal analysis [thermogravimetry (TG) and differential thermal analysis (DTA)] was carried out on a Stanton Redcroft STA-781 instrument. Samples (ca. 15 mg) were heated from ambient temperature to 1300°C at 10°C min-' in flowing N2 gas and then held at 1300 "C until constant mass was achieved. The samples were dried at 110°C prior to use. An internal standard of a-alumina was used. The FTIR spectra in the 4000-400cm-' region were recorded (as KBr disks) on a Nicolet Magna 550 spectrometer for samples pre-dried at 110 "C. 27Al MAS NMR experiments were carried out at ambient temperature on a Varian VXR300 spectrometer operating at a Larmor frequency of 78.15 MHz and a spin rate of ca. 12 kHz. The use of high spin rates is essential in this work; lower rates (e.g.the more generally used 3 kHz frequency) result in spinning sidebands within the range of experimental interest and there- fore complicate the analysis.Relaxation delays of 0.5 s were more than sufficient to avoid saturation effects. Spectra were referenced to an external standard of aqueous AlCl, (1 mol drn-,) and 15" pulse widths were employed. Results and Discussion Boehmite Flash calcination of boehmite in the INSA calciner resulted in the gradual changes in structure shown in the XRD profiles of Fig. 5. After flash calcination at 630 "C (Fig. 5b) it is observed that a large amount of boehmite is present, but some dehydrox- ylation has resulted in the formation of a poorly crystalline y-Al,O, phase.On flash calcination at 730 'C (Fig. 5c) it can be seen that none of the boehmite phase remains and the broad peaks at 26=ca. 37.1, 39.6, 45.4, 60.8 and 68.1" indicate that only y-A1203 is present. The flash calcine produced at 830°C gives a nearly identical profile, but upon calcination at 930 and 1030°C (Fig. 5e and f) the formation of 6-Al,03 is indicated by the presence of peaks at 28= 19.4, 32.7, 46.6 and 62.3'. The y/6-A1203 phase mixture is poorly crystalline; the triplet of 6-A1203 (26 =45.6,46.6 and 47.6') is not well defined and appears as one broad peak. These changes in the structure of boehmite upon flash calcination correspond to those after soak calcination up to ca. 800°C. Further calcination of the flash calcines revealed an equivalent dehydroxylation sequence to that followed upon soak calcination: boehmite, y, 6, 8, a-Al,O,.This sequence corresponds with the previous literature on soak calcination*-" and thus will not be discussed further. The results obtained from boehmite calcination are import- ant with regard to the behaviour of gibbsite upon calcination because gibbsite may dehydroxylate to boehmite (and conse- quently follow its dehydroxylation sequence) depending upon the particle size and calcination method used. Gibbsi te XRD and Thermal Analysis. Thermal and XRD results showed that soak calcination of coarse gibbsite (FRF20) gave both the dehydroxylation pathways (a) gibbsite, boehmite, y, 6, 8, a-Al,03 and (b)gibbsite, x,K, a-AI2O3 [with pathway (a) being predominant], in accordance with the The DTA traces of flash calcines of coarse gibbsite (FRF20) showed the usual peaks at ca.300, 525 and 1235 "C (Fig. 6). xl xl .-i= v)c CI .-c I x0.05,, . 1 b h AL 0 10203040006077080 2@/degrees Fig.5 XRD profiles of a, boehmite; and of INSA flash calcines of boehmite prepared at: b, 630; c, 730; d, 830; e, 930 and f, 1030°C DTA\ -h$ 94-Y $ 92 -90-88 -1 TG Fig. 6 Thermal analysis of a flash calcine prepared from coarse gibbsite, FRF20, in EXON (4s at 900OC) These peaks correspond to the following phase changes: gibbsite to boehmite and x-A1203 (peaks overlap); boehmite to y-A1203; and 8-Al2O, to a-A1203, re~pectively.~'~~~ The first two phase changes are endothermic dehydroxylation reactions, whereas the latter peak is an exothermic phase change due to a structural transition from ccp 8-A1203 to hcp a-A1203.No peaks are observed for the sequence y-Al,03 through to 8-Al,O, (Fig. 2) because the oxygen anion array remains ccp with the only changes in structure being a redistribution of A1 atoms between octahedral and tetrahedral sites with very small enthalpy changes.24 In addition to the peaks discussed above, there is also an unusual exothermic peak at ca. 820°C. On further heating of a flash-calcined coarse gibbsite sample at J. Muter. Chem., 1996,6( l), 73-79 820°C for 5 h it was determined that the phases y, 6 and 8-A1203 were present and therefore it was evident that the x-A1203 present had undergone a phase change; the multiphasic nature of the product at 820°C made it difficult to determine which particular phase had resulted from the transition. Interpretation of the phase change became easier, however, on analysis of the less complicated dehydroxylation sequence of the flash-calcined fine gibbsite sample (UF35E SD).XRD revealed that soak calcination of fine gibbsite (UF35E SD) gave the dehydroxylation sequence: gibbsite, x, K, a-A1203. Flash calcination in EXON gave only the x-A1203 phase, whereas in INSA an unusual phase change from x-A1203 to y-Al,03 occurred (Fig. 7). This particular phase change has also been observed during the dehydroxylation of gibbsite in a fluidised bed calciner with suspension preheater~.,~ However, only XRD was used in that study and the temperature at which that phase change occurred was not reported.DTA of the INSA calcines prepared below 830 "Cand of all the EXON calcines showed that the x-y phase change occurs at ca. 810 "C (Fig. 8) and consequently identified the dehydroxylation sequence of flash-calcined fine gibbsite to be: x, y, 6, 8, a-Al,O,. Similar DTA traces of gibbsite calcines prepared in a cyclone have recently been reported.26 The phase transform- ation in this case was also observed to be at 810 "C but was described as being from a completely amorphous product to q-A1203 (XRD). The calcination product is, however, unlikely to be q-A1203 (a calcination product of bayerite), but as the XRD profile of q-Al,03 is nearly indistinguishable from that of y-Al,O, the results are similar to those presented here.Further XRD analysis of coarse gibbsite confirmed that flash- calcined coarse gibbsite experiences a crossover (x-y) between the soak calcination dehydroxylation routes shown in Fig. 2, and consequently no K-A~,O~ forms (Fig. 9). The absence of the K-A~~O~ phase in flash calcines of gibbsite is also evident on examination of the NMR spectra. 7 I 701 0 200 400 600 800 loo0 1200 1400 TI% Fig. 8 Thermal analysis of a flash calcine prepared from fine gibbsite, UF35E SD, in EXON (4 s at lO00"C) 0 200 400 600 800 lo00 1200 n0c Fig. 9 Dehydroxylation sequences for flash calcined coarse gibbsite [(a)+(c)] and flash calcined fine gibbsite [(c) only] "Al MAS NMR Spectroscopy.The 27Al MAS NMR spectra obtained from the gibbsite starting materials have a main peak at 6 4.7 (Fig. 10a). This, and a small shoulder at 6 -15, are fully consistent with six-coordinated ('octahedral', A106) aluminium cations, as shown from X-ray diffraction studies.' Soak calcines. The spectra of soak calcines obtained from both gibbsite samples are also very similar. As shown in Fig. lob, the spectrum for the soak calcine of FRF20 prepared at 400 "C (containing mainly boehmite and x-Al,O,) has peaks at 6 5.8 and 62.4, assigned to six- and four-coordinate Al, re~pectively.~~.~~The peaks for six-coordinate A1 sites in the different components of the calcine overlap to form one broad asymmetric peak.The disorder in the amorphous phases is evident from the broad and asymmetric lineshapes of the peaks. The upfield tails (Fig. lob-d) are due to a distribution of quadrupolar interaction parameters (a result of a range of 10 20 30 40 50 60 70 80 300 200 I00 0 -100 -200 -300 28ldegrees s Fig. 7 XRD patterns of: a, UF35E SD; and of INSA flash calcines of Fig. 10 27Al MAS NMR spectra of: a, coarse gibbsite; and of coarse UF35E SD prepared at b, 730; c, 830; and d, 1030°C gibbsite samples soak calcined at: b, 400; c, 700, and d, 1000°C 76 J. Muter. Chern., 1996,6(l), 73-79 bond lengths and angles), and therefore suggest not only a loss of long-range order but also a range of local environments. This could be due to the additional occupation of octahedral and tetrahedral interstitial sites (not occupied in ordered spinel structures) which introduces further disorder into the systems.These observations have been seen in other systems; upfield tails in 27Al NMR spectra are well known in aluminosilicate glasses, once a sufficiently rapid rotation is employed to remove overlapping spinning sideband^.^' The peak assigned to four- coordinate A1 arises solely from the presence of x-A1203. The spectrum for the calcine obtained at 700 "C (x-A1203 and y-Al,O,) is similar to the spectrum of the sample obtained at 400°C with peaks at 6 5.8 and 60.2, indicating six- and four-coordinate Al, respectively (Fig. 1Oc). The spectrum (Fig.10d) for the soak calcine of coarse gibbsite prepared at 1000 "C (K,6, 8-A1203) differs considerably from those previously described. It has a peak which corre- sponds to six-coordinate A1 as before, but which is now shifted upfield to 6 4.3. The peak is also significantly broader than the corresponding one in the previously described spectra, and has a very slight shoulder to higher field. Although these changes could be influenced by quadrupolar effects, it is also possible that they could indicate that the six-coordinate A1 atoms are now in more symmetrical environments compared to those in the previous calcine but that there are a range of these relatively symmetrical sites which cannot be dis-tinguished. In the shift region 6 30-70 some structure is apparent, with a peak at 6 63.9 and another at 6 43.9.The first peak is assigned to four-coordinate A1 sites. The feature at 6 43.9 is upfield of the region usually associated with AlO, sites and may arise from a highly defective structure which consequently changes in the shape and position of resonances arising from the tetrahedrally coordinate Al. For example, it has been suggested that intermediate peaks such as this may arise from highly distorted AlO, units with an excess of oxygen in the immediate environment.,' It is also possible that the peak could arise from defective (oxygen-deficient) A106 sites. There are, therefore, at least two possible physical origins for 'AlO,' sites. The differences in the spectrum discussed are assigned to the presence of K-A~,O, since they exist in the spectrum of a pure K-A~~O, phase prepared via soak calcination of fine gibbsite, but are not observed in the spectra of 6/8-Al,O, phases prepared from boehmite.To obtain a detailed characterisation of the A1 environments in K-A~,O,, a study and comparison of spectra of K-A~,O, samples at a range of magnetic fields is required. Flush calcines. The 27Al MAS NMR spectra obtained from the EXON flash calcines of both fine and coarse gibbsite differ considerably from those obtained from soak calcines. Fig. 1 la shows a spectrum, obtained from fine gibbsite flash calcines, with three distinct peaks at 6 4.6, 27.2 and 63.3, which can be assigned to six-, five- and four-coordinate Al, respectively.There is no shoulder observable at 6 43.9, confirming the absence of the K-A~,O, phase. Spectra obtained from the flash calcines appear to be almost identical to those obtained in a study on p-A1203 under identical conditions of spin speed and magnetic field ~trength.~ p-A1203, is produced by calcining gibbsite, UF35E SD, at 200°C in vucuo, and it is evident from its largely featureless XRD pattern that massive disorder exists in this material; the A1 sites are likely to possess extremely distorted geometries. The presence of a peak, at a position (6 27-29) very similar to the value obtained in this study, has also been reported for anodically formed amorphous alumina films,31 and was assigned to A10, sites. The presence of five- coordinated aluminium sites is often associated with highly disordered materials.This arouses interest because of the strong Lewis acidity which appears to be associated with such sites, suggesting the possibility of special catalytic activity.32 Peaks assigned to AlO, sites are present in all spectra of 1'"'1""1"'~1"~"'~'' 300 200 100 0 -100 -200 -300 6 Fig. 11 Representative "A1 MAS NMR spectra of a, all the fine gibbsite flash calcines prepared in EXON and those fine gibbsite flash calcines prepared in INSA below 830 "C; and of b, all the fine gibbsite flash calcines prepared in INSA above 830°C and all flash calcines of fine and coarse gibbsite after further heating at 810°C for 5 h gibbsite flash calcines prepared in the EXON flash calciner, but are only present in INSA flash calcines prepared below 830 "C.Upon flash calcination at higher temperatures this peak is not observed and the spectra resemble those of the defect spinel structure of y-A1203, i.e. AlO, and A106 sites in the approximate ratio 1: These changes were also observed upon further thermal treatment of the EXON flash calcines by soak calcination at 810°C for 5 h (Fig. llb). This reveals that the metastable AlO, sites only exist in the extremely amorph- ous, kinetically frozen x-Al,03 phase produced by flash calci- nation and not in further dehydroxylated samples, or in a x-A1203 phase produced by soak calcination methods. It is thought that the faster heating rate of flash calcination in the production of the kinetically frozen x-A1203 phase must be leaving additional disorder in the structure, with some A1 'frozen' in five-coordinated environments.FTIR spectroscopy. Coarse gibbsite. In the IR spectrum of coarse gibbsite (Fig. 12a) there were OH stretching vibrations at 3621, 3528, 3463, 3397 and 3376cm-'. The band at 3463 cm-' is associated with hydrogen bonding between adjac- ent layers, whilst the bands at 3528 and 3621 cm-' are thought to correspond to stronger hydrogen bonds stretched between hydroxy groups lying in the same plane.34 It was noticed that the peaks corresponding to the hydrogen bonds between hydroxy groups lying in the same plane (3621 and 3528 cm-') had a greater relative intensity than the peak associated with hydrogen bonding between the layers (3463 cm-') and the peaks at 3397 and 3376 cm-'.The A10-H in-plane bending vibrations are seen at 914, 966 and 1019 cm-1.35+36 The peak at 914cm-' is thought to correspond to an Al(0H)Al group free from hydrogen bonding.37 The out-of-plane, v(OH), bend- ing vibrations are observed at 812 and 743 cm-', whereas the vibrations due to the bending modes of AlO, units occur at 674, 570 and 515 cm-1.36 Fine gibbsite. In the FTIR spectrum of the fine gibbsite sample (Fig. 13a) 0-H stretching vibrations were again observed at 3620, 3530, 3468, 3392 cm-' as expected, but the relative intensities of these peaks differed from those of coarse gibbsite. The stretching frequencies 3620 cm- ' (hydrogen bonds between hydroxy groups lying in the same plane) and 3392cm-' are relatively less intense than the peaks at 3468 cm- ' (corresponding to hydrogen bonding between adjacent layers) and 3530 cm- (hydrogen bonding between OH groups in the same plane).J. Muter. Chem., 1996, 6(l), 73-79 77 Fig. 12 FTIR spectra of: a, coarse gibbsite; and of flash calcines of coarse gibbsite prepared in EXON at 1000"C with residence times of: b, 5; c, 7.5; and d, 10 s 4400 3800 3200 2600 2000 1400 800 200 wavenumber / cm-' Fig. 13 FTIR spectra of: a, fine gibbsite; and of flash calcines of fine gibbsite prepared in EXON at 900°C with residence times of: b, 5; c, 7.5; and d, 10 s Coarse gibbsite calcines. The IR spectra of the soak and flash calcines of coarse gibbsite (Fig.12b, c, d) show a gradual change in structure with increasing temperature and residence time. The peaks corresponding to 0-H bonding gradually merge to form a broad peak centred at 3485cm-', and the formation of low crystallinity boehmite is observed as indicated by a short-lived characteristic peak at 1080 cm- 'correspond-ing to a AlO-H bending ~ibration.~, The definition in the A1-0 stretching and bending vibrations (674, 570 and 515 cm-') is also gradually lost and replaced by two broad bands centred at 570 and 790 cm-'. These bands are due to a wide range of A1-0 stretching modes for AlO, and A10, units, re~pectively.~, In spectra of the soak calcines it was noticed that the broad peak centred at 570cm-' (AlO,) is always relatively more intense than the broad feature at 790cm-' (AIO,).Although the spectra obtained from the EXON flash calcines were identical to those of soak calcines in all other respects, it was observed that the band centred at 790cm-' was of similar intensity to that of the peak at 570 cm-'. Fine gibbsite calcines. The 0-H stretching region of the IR spectra of the soak and flash calcines of UF35E SD was observed to gradually lose its fine structure with further calcination, and a broad peak centred at 3472 cm-' conse-quently appeared (Fig. 13b, c, d). The peak characteristic of boehmite (1080 cm-') was not observed, providing further confirmation of a different dehydroxylation route to that for coarse gibbsite. At low wavenumbers, in contrast with the flash calcines of coarse gibbsite, the region centred around 770 cm-' (rather than 570 cm-') is more intense throughout the series of EXON flash calcines.From the spectra of INSA calcines, however, the peak at 770cm-' is only slightly more intense in the spectra obtained below 830 "C (Fig. 14). Upon flash calcination at higher temperature, the peak at lower wavelength shifts further into the far-IR region (604,584,571 cm-l) (thought to indicate more extensive three-dimensional oxide formation3') and also becomes the most intense part of the spectrum. Comparison of the FTIR and NMR spectra of soak and flash calcines of both gibbsites showed that the broad peak at ca. 770cm-' is only more intense than the broad peak at ca.560 cm-' for flash calcines where 27Al MAS NMR spectra reveals that A105 sites are present. These regions of the FTIR spectra have been assigned in the literature to the A1-0 stretching modes of A104 and AlO, units, re~pectively.~~ However, with the benefit of the evidence from our NMR 1500 loo0 500 0 wavenumber / cm-' Fig. 14 FTIR spectra of fine gibbsite flash calcines prepared in INSA at: a, 630°C (with some A105 sites occupied); and b, 830°C (no A10, sites occupied) 78 J. Muter. Chem., 1996, 6(l), 73-79 studies, it is evident that the spectral region around ca. 770cm-1 also contains features due to the presence of A10, units. 5 6 7 E. J. W. Verwey, Z. Kristallogr., 1935,91, 65. A. J. Leonard, P.N. Semaille and J. J. Fritpiat, Proc. Br. Ceram. Soc., 1969, 13, 103. R. C. T. Slade, J. C. Southern and I. M. Thompson, J. Muter. Chem., 1991,1,563. Conclusions 8 9 R. Tertian and D. Papee, J. Chim. Phys., 1958,55,341. B. C. Lippens and J. H. De Boer, Acta Crystallogr., 1964,17, 1312. Soak and flash calcines of boehmite and of two samples of gibbsite with different particle sizes have been characterised by XRD, DTA, FTIR and 27Al MAS NMR spectroscopy. Boehmite undergoes the dehydroxylation sequence: boehmite, y, 6, 8, a-Al,O, under both soak and flash calcination. 10 11 12 13 H. Dexpert, J. F. Larue, I. Mutin, B. Moraweck, Y. Bertaud and A. Renouprez, J. Met., 1985,37, 17. J. H. De Boer, J. M. H. Fortuin and J. J. Steggerda, J. Phys. Chem., 1954, 57, 170.H. Saalfeld, Neues Jahrb. Miner. Abh., 1960,95, 1. J. F. Brown, D. Clark and W. W.Elliott, J. Chem. SOC.,1953,84. Differences in the dehydroxylation sequences of gibbsite were 14 S. J. Wilson, Proc. Br. Ceram. Soc., 1979,28,281. observed with changes in calcination method and particle size. Upon soak calcination the dehydroxylation sequence of fine gibbsite was observed to be: gibbsite, x, K and a-Al,O,. Soak calcination of coarse gibbsite gave both the dehydration path- ways (a)gibbsite, boehmite, y, 6, 8, a-Al,O, and (b)gibbsite, x, K, a-Al,O,; pathway (a)was predominant. 15 16 17 18 19 20 G. W. Brindley and J. 0.Choe, Am. Mineral., 1961,46,771. T. Sato, J. Thermal Anal., 1987,32, 61. G.D. Chunkin and Y.L. Seleznev, Inorg. Muter., 1987,23, 371.M. K. B. Day and V. J. Hill, J. Phys. Chem., 1953,57,946. W. H. Gitzen, Alumina as a Ceramic Material, Am. Ceram. SOC., Columbus, OH, 1970, ch. 3. Flash Reaction Processes, ed. T. W. Davies, NATO AS1 Series An unusual phase change between the hcp x-A120, phase and the ccp y-A1203 phase occurred at ca. 800 "Cin the flash- calcined gibbsite samples. The dehydroxylation sequence of the fine gibbsite sample upon flash calcination was observed to be gibbsite, x, y, 6, 8, a-A1203, whereas for the coarse gibbsite sample both dehydroxylation pathways (a)gibbsite, x, 21 22 23 24 E282, Kluwer, Dordrecht, 1994. T. W. Davies, High Temp. Technol., 1984,2, 141. R. H. Meinhold, S. Salvador, T. W. Davies and R. C. T. Slade, Trans. Inst. Chem. Eng., 1994, part A., 72, 105.The Diferential Thermal Investigation of Clays, ed. R. C. Mackenzie, Mineralogical Society, London, 1957, ch. XII. B. A. Scott and W. H. Horsman, Trans. Br. Ceram. Soc., 1970, y, 6, 8, a-Al,O, and (b) gibbsite, boehmite, y, 6, 8, a-Al,O, were followed. ~c-Al,o, formation was not observed in the flash calcines and the x-A1203 phase produced by flash calci- nation was shown by 27Al MAS NMR and FTIR spectroscopy to have a considerably different short-range structure to that produced by soak calcination. 25 26 27 28 69,37. K. Yamada, T. Harato, S. Hamano and K. Horinouchi, J. Met., 1983, 35,22. B. P. Zolotovski, V. E. Lojko, V. M. Mastikhin and R. A. Buyanov, Kinet. Katal., 1990,31, 1014. R. H. Meinhold, R. C. T. Slade and R. H. Newman, Appl. Magn. Reson., 1993,4, 121.D. Muller, W. Gessner, H. J. Behrens and G. Scheler, Chem. Phys. We thank the EPSRC and Alcan Chemicals Limited for funding a CASE studentship for V.J.1-J. We thank the EPSRC Solid-state NMR Service (University of Durham) for recording spectra. We thank Alcan Chemicals Europe for provision of 29 30 Lett., 1981,79, 59. R. K. Sato, P. F. McMillan, P. Dennison and R. Dupree, J. Phys. Chem., 1991,95,4483. H. Schneider, L. Merwin and A. Sebald, J. Muter. Sci., 1992, 27, 805. commercial grade gibbsite and boehmite. We thank Derek Bridson, Michael Jones, Neville England and John Coote for their technical assistance with EXON. 31 32 33 I. Farnan, R. Dupree, A. J. Forty, Y. S. Jeong, G. E. Thompson and G. C. Wood, Philos. Mag. Lett., 1989,59, 189. F. R. Chen, J. G. Davis and J. J. Fripiat, J. Catal., 1992,133,263. R. Dupree, I. E. Farnan, A. J. Forty, S. El-Mashri and L. Bottyan, J. Phys. Colloq. C, 1983,8, 113. References 34 V. C. Farmer, The Infrared Spectra of Minerals, monograph no. 4, Mineralogical Society, London, 1974, p. 137. 1 H. D. Megaw, 2.Kristallogr., A, 1934,87, 185. 35 G. A. Dorsey, Jr., J. Electrochem. SOC.,1966,113, 169. 2 P. P. Reichertz and W.J. Yost, J. Chem. Phys., 1946,14,495. 36 Ph. Colomban, J. Muter. Sci., 1992,24, 3002. 3 R. E. Newnham and Y. M. de Haan, Z. Kristallogr., 1962,117,235. 37 D. Tilak, B. Tennakoon, W. Jones and J. M. Thomas, J. Chem. 4 H. Dexpert, E. Freund, E. Lesage and J. P. Lynch, in Metal Support Soc.,Faraday Trans. 1,1986,82,3081. & Metal Additive Egects in Catalysis, ed. B. Imelik, Elsevier, Amsterdam, 1982, p. 53. Paper 5/03618F; Received 6th June, 1995 J. Muter. Chem., 1996, 6(l), 73-79

ISSN:0959-9428    

DOI:10.1039/JM9960600073

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Hydrothermal syntheses and crystal structures of new layered tungsten(v1) methylphosphonates, M2(W03)3P03CH3(M = NH,, Rb, Cs) William T. A. Harrison,+" Laurie L. Dussack," John T. Vaughey,$" Thomas Vogtb and Allan J. Jacobson' "Departmentof Chemistry, University of Houston, Houston, Texas 77204-5641, USA bPhysics Department, Brookhaven National Laboratory, Upton, New York 11973, USA The hydrothermal syntheses of Cs2( WO,),PO,CH,, Rb2( W03),P03CH, and ( NH4),( W03),P03CH,, three new non- centrosymmetric, layered tungsten( vI)-containing phases, are described. The crystal structures of Cs2( W03),P03CH3 and (NH4)2( W0,),P03CH3 were refined by using Rietveld analysis of powder diffraction data. These phases are built up from hexagonal tungsten-oxide-like layers of vertex-sharing W06 groups, capped by P-CH, entities (as methylphosphonate groups) on one side of the tungsten/oxide layer.Caesium, rubidium or ammonium cations provide interlayer charge compensation for the anionic layers. IR, Raman and thermogravimetry (TG) data for these phases are also presented. CrTstal data: Cs2( W03),P03CH3, M, = 1055.36, trigonal, space group R3 (no. 146), a = 7.25791 (9) A,c = 20.2762(4)A,V= 925.00(4) A,, Z = 3, R, = 2.37%, R,= 3.12% (2943 neutron powder data). (NH4)2(W03)3P03CH,, M,= 825.63, trigonal, R3 (no. 146), a= 7.22851(7)A, c= 19.3471(3)A,V=875.48(3) A3, Z=3, R,=8.14%, R,= 10.12% (2665 X-ray powder data). We have now reported the syntheses and structures of several new, related, non-centrosymmetric layered materials based on a hexagonal motif of corner-sharing Vv06, MoV106 or Wv'06 0~tahedra.l~~This layer connectivity results in a characteristic array of three- and six-rings of octahedra (Fig.l), which is also found in three-dimensionally connected phases such as hexagonal tungsten oxide (HTO), W0,.6 In the new layered materials, these octahedral layers are capped, either on both faces of the MO, layer for the vanadium or just on one side, for the molybdenum and tungsten phase^.^.^^^ Other double-capped layered phases based on the same motif were reported earlier.7-" The capping species include selenium(1v) atoms (as part of pyramidal [ SeO3I2- selenite groups) in NH4( V02)3(Se03)2,1 K(V02)3(Se03)2,4 (NH4)2(M003)3se03,2 cs2(M003)3Se03,2 (NH4)2(W03)3Se035 and Cs2( WO,),SeO,.' In both Cs2(Mo0,),P0,CH,3 and Rb2(Mo0,),P0,CH,,3 P-CH, entities (as part of [O3PCH,I2- methylphosphonate groups) provide the capping.Alkali-metal or ammonium cations serve as the interlayer charge-balancing species in all these phases. In this paper we report the syntheses, structures and some properties of Cs2( W03),P03CH,, Rb2( WO,),PO,CH, and ( NH4),( W03),P03CH3, the tungsten(v1)-containing analogues of the layered alkali-metal molybdenum (VI) methyl-phosphonate materials Cs2( MoO,),PO,CH, and Rb2( Mo03),P03CH3 ., Synthesis and Initial Characterization Cs2( W03),P03CH3 was hydrothermally prepared from 0.873 g CsOH.H,O (5.2 mmol Cs), 0.6 g W03 (2.6 mmol W), 0.5 g 98% CH3P03H2 (5.2 mmol P), and 12 cm3 of H20 (initial Cs : W : P03CH, ratio = 2 : 1: 2).These reactants were sealed in a 23 cm3 PTFE-lined Parr bomb and heated to 200°C for 5 days. After overnight cooling, the pH of the final filtrate was 3.1. Fine white Cs2(W03),P03CH3 powder was recovered by vacuum filtration and air drying. The overall yield of solid product, based on tungsten, was ca. 87% by mass. Rb2( W03),P03CH, was synthesized from a mixture of 0.530 g 50% aqueous RbOH solution (2.59 mmol Rb), 0.6 g WO, (2.59 mmol W), 0.497 g 98% CH,PO,H, (5.18 mmol P) and 10 cm3 water. The initial mole ratio of reactants was 1: 1: 2 Rb: W: P. The mixture was sealed in a Parr bomb, heated to 200°C for 11 days, and allowed to cool to ambient temperature over 24 h. Fine white Rb2( W03)3P03CH3 powder (yield=30% based on W) was recovered from the filtrate (pH = 2.0) by vacuum filtration. Reactions with higher starting concentrations of Rb resulted in products contaminated with unreacted WO, (X-ray powder diffraction measurements). The best method to prepare pure, highly crystalline (NH4),(W0,),P03CH3 started from a mixture of 0.3 g (NH4)loW12041 (1.18 mmol W), 0.38 g NH4Cl (7.1 mmol NH4), 0.34 g CH3P03H2 (3.55 mmol P), 2.85 g 35% tetra- i, t Present address: Department of Chemistry, University of Western Fig.1 STRUPLO polyhedral view of a generic hexagonal tungsten Australia, Nedlands, WA 6907, Australia. oxide (HTO) octahedral layer, viewed down the crystallographic $ Present address: Department of Chemistry, University of Iowa, [OOl] direction, showing the infinite network of octahedral 3-and Ames, Iowa 50011, USA.6-rings J. Muter. Chew., 1996, 6(l), 81-87 ethylammonium hydroxide (TEAOH) and 6 cm3 water. The reactants were heated to 180°C for 5 days in a Parr bomb. Product recovery by vacuum filtration led to an 85% yield of white (NH4)2( W0,),P03CH, powder (pH of filtrate =6.6). Reactions at lower starting pHs (less TEAOH added to reaction) resulted in lower yields of (NH4)2( WO,),PO,CH, (e.g.47% at pH 3.8). X-Ray powder data for thoroughly ground samples of the title compounds were recorded on a Scintag XD_S 2000 autc- mated powder diffractometer [Cu-Ka radiation, A = 1.54178 A, T=25(2) "C]. They resulted in similar powder patterns which could be indexed on rhombohedrally centred hexagonal unit cells (see Table 6, later).These R-centred unit cells suggested that Cs2( W03),P03CH3, Rb2( WO,),PO,CH, and (NH,),( W0,),P03CH3 are isostructural with the three-layer molybdenum-containing materials Cs,( Mo0,),P0,CH3 and Rb2(Mo03)3P03CH3,3as was found to be the case when full profile refinements were carried out (uide infra). Since we could not obtain suitably sized single crystals of these materials for X-ray diffraction measurements, we pro- ceeded to refine their crystal structures by the Rietveld method. X-Ray data were used for (NH4)2(W03)3P03CH,. In order to maximize the refinement precision of the lighter atoms, neutron powder data were collected for Cs,( WO3),PO3CH,.Crystal Structure Refinement The crystal structure of Cs2( W03),P03CH3 was optimized by the Rietveld method using constant-wavelength powder neu- tron diffraction data. A ca. log quantity of ground white powder sample was loaded and sealed into a cylindrical vanadium sample can. Room temperature, high-resolution powder data were collected on the HRNPD diffractometer at the High Flux Beam Reactor (HFBR), Brookhaven National Laboratory. A data collection time of 48 h resulted in satisfac- tory copnting statistics (5" <28 < 165"; step size =0.05"; A = 1.8857 A, as precalibrated with A1203). A characteristic large background signal was observed for the Cs2( W03),P03CH3 sample, due to incoherent scattering from the methyl protons. The Cs2( W03),P03CH3 structure refinement was carried out using the program GSAS.12 The initial atomic model, in space group R3 (no. 146), was based on the structure of CS,(MOO~)~PO~CH~,~with W substituting for Mo.The calcu- lated Cs2( Mo03),P03CH3 proton position was not included in the initial refinement cycles for Cs2( W03),P03CH3. Coherent neutron scattering lengths (x m) of b(C)= 0.665, b( H)= -0.374, b(Cs)=0.542, b(W) ~0.477, b(P)=0.513 and b(0)=0.581 were assumed. The refinement proceeded in typical fashion, with profile (scale factor, detector zero-point correction, polynomial background descriptors, peak shape parameters) and atomic positional and thermal parameters added to the model as additional variables as the refinement converged.The single crystallographic proton site was located from a difference Fourier synthesis, and added to the atomic model as an H atom. A simple Gaussian model was found to be inadequate to describe the observed peak shape, and a significantly better fit was obtained by using a pseudo-Voigt Gaussian/Lorentzian model.', The structure of (NH4)2( W03),P03CH3 was refined by the Rietveld method using powder X-ray data (Scintag XDS5000, flat-plate sample geometry, 8-8 geometry, 20" <28 < 100'). The starting model for (NH4)2( W03),P03CH3 in space group R3 (no. 146) was taken from that of Cs2( W03),P03CH3 (N replacing Cs). The X-ray Rietveld refinement proceeded normally, and profile and atomic param- eters were added to the model as variables as the refinement progressed.The refinement converged to satis-factory residuals (program: GSAS), but the precision of the refined light-atom parameters is much poorer than the equivalent values obtained from the neutron refinement on 82 J. Muter. Chem., 1996, 6(l), 81-87 Table 1 Crystallographic parameters empirical formula 3csZp lo1Zcl H3 w3p1 1ZNZCIH 11 formula mass 1055.36 825.63 habit powder powder colour white white crystal system trigonal trigonal a14 7.25791 (9) 7.22851 (7) CIA 20.2762( 4) 19.3471( 3) VIA3 925.00( 4) 875.48(3) z 3 3 space group R3 (no. 146) R3 (no. 146) TIT 25(2) 25(2)ra$iation neutrons X-rays LlA 1.8857 1.5418 Pcalolg cm-3 5.68 4.70 powder data 2943 2665 parameters refined 51 29 2.37 8.14 3.12 10.12 5.84 3.84XL ~ ~ ~~ R, = 100 x CIyo-Cy,l/Xl yo 1.R, =100x [CW(yo -CY,)~/ZCW~O~]"~, where Cis a scale factor. Cs2( W03),P03CH3. No possible (NH4)2( W0,),PO3CH, proton positions could be located from Fourier maps. Crystal and experimental data for these refinements are listed in Table 1. Final observed, calculated and difference profiles for the Cs2( W03),P03CH3 and (NH4)2(W03)3P03CH3refinements are shown in Fig. 2 and 3, respectively. Physical Measurements Thermogravimetry (TG) data for Cs,( W03),P03CH3, Rb2( W03),P03CH, and (NH4)2( W03),P03CH, were col-lected on a DuPont 2950 analyser (ramp 5"Cmin-' under flowing 0, gas). The post-TG residues were analysed by powder X-ray diffraction (XRD). IR spectra (KBr pellet method) for the title compounds were recorded in the 400-4000 cm-' range on a Galaxy FTIR 5000 series spectrometer.Raman data were obtained using a coherent K-2 Kr+ ion laser excited at 406.7 nm. Data for Cs2(Mo0,),P0,CH3 (KBr pellet method) were accumu- lated at 1 s intervals for every wavenumber over the range 100-1700 cm-I (Spex 1403 double monochromator/ Hamamatsu 928 photomultiplier detection system). Results Crystal structures Final atomic positional and thermal parameters for Cs2(W0,),P03CH, are listed in Table 2, with selected bond distance/angle data in Table 3. Similar data for (NH4),(WO3),PO3CH3 are listed in Tables 4 and 5 Cs2( W0,),P03CH, and (NH4),( WO,),PO,CH, both crys- Table 2 Atomic positional/thermal parameters for Csz(W03),P03CH: atom X Y 2 uisoIA2 0 0 0.1733(4) 0.026( 2) 113 213 0.2291(5) 0.008(2) 0.0033( 9) 0.5189( 13) 0.05724( 26) 0.01 1 (2) -113 113 0.1871 (4) 0.018 (2) 0.0719( 9) 0.5410(18) --0.02355( 23) 0.026( 2) 0.2201 (14) 0.7921(14) 0.08817( 22) 0.010(2) -0.2030( 14) 0.5799(8) 0.05131 (23) 0.006(2) -0.0936(7) 0.4589( 13) 0.16167( 23) 0.003(2) -113 113 0.2727( 4) 0.013 (2) 0.5010( 15) 0.2034( 15) 0.2935( 4) 0.036(3) 1 I I I I I I 1 3000 l-cn -c3 0 2000,o ._2 cn 0)c) C.-iooa I I I I I I Ill I1 1 111 11111111 111111111 I .IIWIU 111111111ll11111 18IY 1111111111 m11.111 III.III 118111 III1111 I1 1 I‘ I1 a I I 1 I 1 I I I 20 40 60 80 100 120 140 160 2Wdegrees Fig.2 Final observed (crosses), calculated (line) and difference profile plots for the Rietveld refinement of Cs2(WO,),PO,CH, (neutron data) v)c c 0e 1000 .-a u) Q,c. c.-500 0 7. . -rr .* .I .. A. 11 ----11 I I I I I I I 1 20 30 40 50 60 70 80 90 100 2Bldegrees Fig. 3 Final observed (crosses), calculated (line) and difference profile plots for the Rietveld refinement of (NH,),( WO,),PO,CH, (X-ray data) tallize as isostructures of the M,( MoO,),PO,CH, (M =Cs, project into the interplanar region of the structure. The PCH, Rb) structure reported earlier., entity caps a ‘three-ring’ of O(4) atoms in the structure, while This structure type consists of anionic layers of vertex-the O(1) atoms are uncapped, resulting in short W( 1)=0( 1) sharing W06 octahedra and tetrahedral P03CH3 (methylphos- ‘0x0’ bonds.All the W-O(4)-P capping occurs on one side phonate) units, which are fused together by W-0-W and of the W/O sheet. Both O(2) and O(3)bridge adjacent tungsten W -0-P bonds. Two crystallographically distinct caesium atoms into three-rings. The overall sheet stoichiometry is or ammonium cations provide interlayer charge balancing. [(wo,)3P03CH3l2-. The crystal structure of Cs,( W03),P03CH, is illustrated in In Cs,( WO3),PO3CH3, the W06 octahedron is significantly Fig. 4-6. distorted from octahedr$ regularity: the tungsten atom is The layer motif in Cs2(W03),P03CH3 and displaced by do,,=0.28 A from the geometrical centre of its (NH,),( W03),P03CH3 consists of hexagonal tungsten-oxide- six oxygen atom neighbours.This distortion may be viewed like sheets of vertex-sharing WOs octahedra. Each W06 unit as a displacement of the tungsten atom from the centre of the shares four of its W-0 vertices with similar neighbours [via W06 octahedron towards an octahedral edge [atoms O(1) 2 x O(2) and 2 x 0(3)], with these W-0-W bonds roughly and O(3)] (Fig. 7), which results in two short (d < 1.8 A) W -0 aligned in the ab plane. The W( 1)-0(2) bond is canted from bonds, tjree intermediate length W-0 bonds, and one long the ab plane by ca. 19”, and the W( 1)-0(3) bond by ca. 4”. (d>2.2 A) W-0 vertex. The inter-octahedral W-O(3)-W’ This canting of the W(1)-0(2) vertex effectively forces a bonds thus show some degree of short-long bond length ‘three-ring’ trio of apical W( 1)-0(4) bonds closer together ‘alternation’, whereas the comparable W-O(2)- W’ bonds to allow them to be capped (see below).Three-rings and six- are not significantly different in length (Table 3). rings of octahedra result from this in-plane connectivity (Fig. The methylphosphonate group in Cs2( WO,),PO,CH, has 6). The two remaining apical W-0 bonds, to O(1) and 0(4), typical geometrical parameters (Table 3). The P and C atoms J. Muter. Chem., 1996, 6(l), 81-87 Table 3 Bond distances (A) and angles (degrees) for csZ( W03)3P03CH3 ~ ~~~ ~ ~ ~ ~~ ~ Cs( 1)-O( 1) x 3 3.2 19 (9) CS( 1)-O(2) x 3 3.196( 7) Cs(2)-O( 1) x 3 3.028(8) Cs(2)-O(2) x 3 3.228( 10) C~(2)-0(4)x 3 3.012( 7) W(1)-O( 1) 1.696( 7) 1.9 19( 13) W(1)--0(2) 2.004( 12) 1.765( 10) W( 1)-0(3) 2.040( 10) 2.205( 7) P(1)-0(4) x 3 1.594( 6) 1.735( 12) C( 1)-H( 1) x 3 1.174( 11) O(1)-W(1)-O(2) 99.7(5) O(1)-W( 1)-O(2) 95.9(5) O(1)-W(1)-0(3) 99.0(5) O(1)-W( 1)-0(3) 94.4( 5) O(l)-W(1)-0(4) 174.7(8) O(2)-w( 1)-O( 2) 83.0 (4) 0(2)-W(1)-0(3) 97.1(4) O(2)- W( 1)- O(3) 163.3( 4) O(2)-W( 1)-0(4) 84.60(31) O(2)-W( 1)-0(3) 164.9(4) 0(2)-W(1)-0(3) 86.7(4) O(2)- W (1)-O( 4) 81.37( 34) 0(3)-W(1)-0(3) 89.5(4) O(3)- W ( 1)-O(4) 83.59( 33) 0(3)-W(1)-0(4) 80.86(31) O(4)-P( 1)-C( 1) 108.9(4) O(4)- P( 1)- O(4) 110.0( 4) W(l)-0(2)-W(l) 133.2(4) W( 1)-O(3)-W( 1) 149.4( 4) W( 1)-0(4)-P( 1) 125.0(4) P( 1)-C( 1)-H( 1) 11 1.0( 5) Table 4 Atomic positional parameters for (NH,),( W03),P03CH3 atom X Y Z uiso /Az N(1) 0 0 0.1730 0.0088(5) “2) 1/3 213 0.229( 11) 0.0088(5) W(l) 0.0058(6) 0.5204(4) 0.064( 6) 0.0088(5) P(l) -1/3 1/3 0.199( 6) 0.0088(5) O(1) 0.061(4) 0.547( 5) -0.021(6) 0.0088(5) O(2) 0.223(8) 0.804( 9) 0.094( 6) 0.0088(5) O(3) -0.218(8) 0.582 (4) 0.054( 7) 0.0088(5) O(4) -0.098(4) 0.444(6) 0.165( 6) 0.0088(5) C(l) -1/3 113 0.2839( 29) 0.0088(5) Table 5 Selected bond distances (A) and angles (degrees) for (NH4)Z( W03)3P03CH3 N( 1)-O( 1) x 3 3.19(9) N( 1)-0(2) x 3 3.04( 7) N(2)-O( 1) x 3 2.95( 10) N(2)-0(2) x 3 3.05( 15) N(2)-0(4) x 3 2.98(8) W(1)--0(1) 1.685 (20) W( 1)-W) 1.94(6) W(1)--0(2) 1.94(5) W( 1)-0(3) 1.89(6) W(1)-0(3) 1.94(5) W(1)-0(4) 2.062( 19) P(1)-0(4) x 3 1.61 3( 23) P(1 )-C( 1) 1.64( 14) O(1)-W(1)-0(2) 98.4(15) O(1)-W( 1)-0(2) 99.2( 12) O(1)-W(1)-0(3) 93.0(13) O(1)-W( 1)-0(3) 92.0( 13) O( 1)-W( 1)-0(4) 171.6( 16) O(2)-W( 1)-0(2) 87.7( 17) 0(2)-W(1)-0(3) 96.8(6) 0(2)-W(1)-0(3) 167.6( 12) 0(2)-W(1)-0(4) 90.0(13) 0(2)-W(1)-0(3) 166.3(13) 0(2)-W(1)-0(3) 83.9(7) 0(2)-W(1)-0(4) 81.6(14) 0(3)-W(1)-0(3) 89.4(16) 0(3)-W(1)-0(4) 85.5(13) 0(3)-W(1)-0(4) 79.8(12) O(4)-P( 1)-0(4) 104.4( 13) O(4)-P(1)-C(1) 114.1(11) W( 1)-0(2)-W( 1) 133.0( 13) W( 1)-O(3)-W( 1) 147.9( 16) W( 1)-O(4)-P( 1) 132.2( 18) are on a three-fold axis, and three equivalent P(1)-0(4) bonds result, each of which bridges to a different WO, unit.The hydrogen atoms of the methyl group are close to being staggered with respect to the oxygen atoms of the PO, moiety [H(l)-C(l)-P(1)-0(4) torsion angle, 5=163”]. The methyl group points towards an octahedral six-ring in the next [( W03)3P03CH,]2-sheet, an! the minimum non-bonding C-H--.O separation is cu.2.7 A. The two crystallographically distinct cation sites (both with site symmetry 3) are both found in the interlayer region for these phases. Cs( 1) is six-coordinate to nearby oxygen atoms 84 J. Muter. Chem., 1996, 6(l), 81-87 Fig. 4 ORTEP view of the W03/P03CH3 building unit of Cs,( WO,),PO,CH,, showing the atom-labelling scheme Ia sin y Fig. 5 View down the [OlO] direction of the unit-cell packing of Cs2( WO3),PO3CH3, showing the ABC . . . sheet structure of (WO,),PO,CH, layers (Cs-0 contacts not shown for the interlayer Cs+ species) and forms a distorted trigonal prism with respect to three oxygen atoms in one adjacent (W0,),P03CH, layer, and three in the other. Cs(2) is nine-coordinate and makes six Cs-0 bonds to one adjacent (W03),P03CH3 sheet, and three Cs-0 bonds to the other.Bond valence sum14 (BVS) values for these cations {BVS [Cs( l)]=0.78, BVS [Cs(2)] = 1.56, expected value= 1.00 in both cases} suggest that the stability of this structure type is not crucially determined by the bonding requirements (size) of the guest cation. However, these species must have some role to play in stabilizing the HTO-type six- ring windows, and must be large enough to bridge adjacent (W0,),P03CH, layers. The (NH4),( W03),P03CH3 structure is substantially simi- lar to that of Csz(WO,),PO,CH,, but the modest precision of Fig.6 View down the [Ool] direction of part of one (W03),P0,CH3 layer in CS,(WO~)~PO~CH,,with selected atoms labelled. The two types of 3-ring may be seen: the capped 3-ring with W atoms linked by O(3) species, and the uncapped 3-ring, with O(2) atoms forming the W-0-W bridges. O(2)and O(3) species alternate around the 6-rings (cf. Fig. 1). 1 Fig. 7 Detail of the tungsten atom coordination in Csz( W03)3P0,CH, showing the displacement of the W atom from the geometrical centre of its octahedron towards the octahedral edge defined by atoms O(1) and O(3) (see text) the X-ray Rietveld refinement [the esds of the derived param- eters are about 4 times as large for (NH4)2(W03)3P03CH3 compared with those for Cs2( WO,),PO,CH,] makes more detailed comparisons difficult.It is clear that the ammonium cations in (NH4),( W03),P03CH3 occupy similar sites to their caesium counterparts in Cs,( W0,),P03CH3. However, no details regarding the hydrogen-bonding scheme in (NH,),( W0,),P03CH3, if any, could be elucidated in this study. Physical data TG for Cs,( WO,),PO,CH, (oxygen atmosphere) showed a gradual 0.4% mass loss to 500"C, then a sharp 0.6% mass loss at 600 "C. The white post-TG residue consisted of unidenti- fied phase(s). TG for Rb2( W0,),P03CH3 showed a slight (ca. 0.15%) mass loss to 53OoC,followed by a sharp loss at ca. 600 "C (overall loss= 1.3%). After heating to 650 "C, the post-TG residue was white; further heating to 800 "C resulted in a light-yellow phase.Powder XRD showed this yellow residue to consist of hexagonal WO," and other unidentified components. The decomposition path for Cs2( W03),P0,CH3 is unknown: however, it is thermally stable to ca. 500°C (powder XRD measurements on a post-thermally treated sample): the initial slight mass losses for Cs2( W03),P03CH, and Rb2( W03)3P0,CH3 are attributable to loss of surface absorbed water, apparent from the IR results. TG for (NH,)2(W0,),P03CH3 (ramp 5 "C min-' to 500 "C under 0,) showed a one-step 7.6% mass loss from 350 to 475 "C. The crystalline component of the off-white residue was hexagonal W03. Further heating to 680°C led to a light-yellow residue which contained both hexagonal and triclinic WO3.I6 Heating (NH4)2(W03)3P03CH3to 800 "C showed a 7.7% mass loss at 500 "C, a slight mass gain to 700 "C, and a second sharp mass loss at 700-720°C (overall mass loss= 8.5%).The light-green residue contains crystalline triclinic W03 and, by implication, glassy tungsten/phosphate components. The IR spectrum of (NH4),( W03),P03CH3 (Fig. 8) shows four characteristic NH, bands at 3335, 3160 and 3048 cm-' (vl and v3 modes), and a strong, sharp band at 1410 cm-' (v4 mode). The spectra of Cs2(W03),P0,CH, (Fig. 9) and Rb2(W03),P03CH, (Fig. 10) are featureless in these regions. All three spectra show complex, overlapping W06 modes in the 650-600 cm-' region, and four sharp peaks which may be correlated with P-0 phosphonate modes in the region between 1040 and 860 cm-'.Similar bands have been seen in previous studies of metal/oxygen clusters capped by alkylphos- phonate groups.I7 Small, sharp peaks at 2940 and 1415 cm-' are methyl C-H bending and stretching modes, and the small r a I I I I I I I I I 4000 3500 3000 2500 2000 1500 1000 500 wavenumberkm-1 Fig. 8 IR spectrum of (NH,),(WO,),PO,CH, 100 L'rl 80 A5 60 a,0 Yt Y.-E 40 -; c 20 Qdo rr) 0-w-I I 1 I I I 1. 4000 3500 3000 2500 2000 1500 1000 500 wavenumber/cm Fig. 9 IR spectrum of CS~(WO,)~PO,CH, J. Muter. Chem., 1996, 6(l), 81-87 85 .oo 80 60 40 20 0 4000 3500 3000 2500 ZOO0 1500 1000 500 wavenumber/cm-' Fig. 10 IR spectrum of Rb2(WO3),PO,CH3 sharp band at ca.1310 cm-' represents the P-C stretch, and is seen for all three samples. The Raman spectrum of Cs2( W03),P03CH3 (Fig. 11)corre-lates well with the IR spectrum of the same compound. Strong Raman bands may be assigned to more symmetric stretching modes for the W06 (626, 698 cm-') and P03C (932, 965 cm-') groups. The weaker peaks at 1310 and 1415 cm-' match the P-C and C-H modes seen in the IR spectrum. Raman bands below ca. 350 cm-' correspond to lattice (phonon) modes as seen in related phase^.^,^.^ Conclusions Compounds Cs2( W03)3P03CH,, Rb2( W03),P03CH, and (NH4)2( W03),P03CH3 have been prepared for the first time and structurally and physically characterized. They are non- centrosymmetric layered phases isostructural with their molyb- denum(v1)-containing analogues CS~(MOO,)~PO,CH, and R~,(MoO,)~PO,CH,.~A polyhedral view of this structure type is shown in Fig.12. Powder neutron diffraction was successful in elucidating the full structure, including the hydro- gen atom position in Cs2( WO3),PO3CH3. A three-layer (ABCABC ...) repeat motif in the c direction is observed in these structures: as with the M2( Mo0,),P03CH3 structures, this may be correlated with the steric requirements of the methylphosphonate group, which points towards a six-ring 200 400 600 800 1000 1200 1400 1600 A v/cm Fig. 11 Raman spectrum of Cs2( WO,),PO,CH, -a sin y Fig. 12 Polyhedral view of the Cs,( W03),P03CH3 structure, viewed down the [OlO] direction, showing the 3-layer repeat motif of the singly capped tungsten oxide layers.Cs and H atoms are omitted for clarity. window in each adjacent sheet.3 Because of the staggered stacking arrangement of the (W0,),P03CH, sheets in these structures, there are no pseudo-infinite channels comparable to the six-ring channels found in hexagonal WO3.I5 The distortion mode of the octahedral cation is different for the tungsten and molybdenum phases: in the M2( Mo0,),P0,CH3 materials, the Mo atom displacement inside the Moo6 octahedron is towards an octahedral face [so-called local ( 111) distortion], resulting in a (three short +three long) Mo-0 bond distance di?tribution in both CS~(MOO~)~PO~CH,[Aoct =0.34 A] and Rb2( Mo0,),P03CH, [Aoct =0.34 A].However, in Cs2(W03),P03CH3, the W atom displacement is towards an octahedral edge [local (110) distortion], and the W06 bonding situation described above results. Compounds Cs2( W03),P03CH3 , Rb2( W0,),P0,CH3 and (NH4),( WO3),PO3CH, complement the layered tungsten(v1) selenites (NH4)2( W03),Se03 and Cs2( WO,),SeO, .5 However, in the two-layer (ABAB ...) selenite phases, the tungsten atom displays a local [lll] distortion mode inside its oFygen atom octahedron [for (NH4)2( WO,),SeO,, d =0.20 A; for Cs,(WO,),SeO,, d =0.27 A], similar to the bonding situation for the molybdenum methylphosphonates. The displacement of a MoV' or Wvl do cation in octahedral coordination may be understood in terms of a second-order Jahn-Teller effect:I8 'spontaneous' distortion of the MOO, or WO, unit will remove degeneracies in the molecular energy levels which arise from overlap of the unoccupied d orbitals of the metal species with the filled p orbitals of the oxygen atom species.The smaller distortion (off-centre displacement of the W atoms) in the tungsten compound is consistent with the general observation that distortions around do transition metals increase with increasing cation charge, but decrease with increasing cation size.lg The magnitude and direction of the cation displacement inside the octahedron is much harder to predict from first principles, and may reflect a combination of second-order electronic effects, lattice stresses, and cation- cation repulsions, as discussed recently by Kunz and Brown.lg The layer separation (defined by the separation of the W atoms in the z direction) 'for SS~(WO~)~PO,CH, is 6.76 A, with a comparable value of 6.45 A for (NH,),(W03),P03CH3.This suggests that the cation plays some role in defining the interlayer separation in these tungsten-containing phases. Conversely, in the M2(Mo03),P03CH3 phases, the c unit-cell dimension is altnost identical for both the caesium and rubid- ium congenersY3 suggesting that interlayer packing consider- ations are most important for the Mo-containing materials. The various layered HTO-type phases are summarized in 86 J. Muter. Chern., 1996, 6(l), 81-87 Table 6 Summary of layered HTO-type phases formula a/A SrA13 (OH )6( p04) (HP04) KA13(OH)6 (S04)2a NaA13(OH)6(cr04)2 Ga3(0H)6(S04)(HS04).H20 KFe3(OH)6( 7.015( 3) 7.020( 2) 7.060( 3) 7.178 7.315( 2) 16.558( 6) 17.223(8) 17.25 (2) 17.170 17.224( 6) 705.7 735.0 744.6 766.1 798.2 7 8 9 10 8 K3(Sb02)3(P04)2 'nHZo 7.147( 1) 30.936( 6) 1368.5 11 K(V02)3(Se03)2 Rb(V02)3 (Se03 )Z NH4(vo2 )3 (Se03 )2 7.125(4) 7.131( 1) 7.137(3) 11.414(3) 11.459(6) 1 1.462( 4) 501.8(5) 504.6( 9) 505.7( 4) 4 20 1 T1,( MoO3),SeO3 7.2774( 5) 11.785(2) 540.5( 1) 21 Rb,( M00,)~Se0~ Cs,( Mo03),Se03 (NH4)2(Mo03)3Se03 7.283(4) 7.267( 2) 7.312(2) 11.964(8) 12.031 (3) 12.377(2) 549.6(8) 550.3( 3) 573.1(3) 21 2 2 Rb2(W03)3Se03 (NH4)2(W03)3Se03 cs2(W03)3Se03 7.2834( 4) 7.2291 (2) 7.2615(2) 11.965( 1) 12.1486(3) 12.5426( 3) 549.7 549.82( 3) 572.75( 3) 5 5 5 Cs,(MoO,),PO,CH, 7.304( 2) 20.02( 1) 924.7 3 Rbz( M003)3P03CH, 7.307( 2) 20.040( 4) 926.7 3 (NH4)2( W03)3P03CH3 Rb2( W03)3P03CH3 Cs,(W03)3PO3CH3 7.22851(7) 7.2483(5) 7.25791(9) 19.3471 (3) 19.3034( 6) 20.2762( 4) 875.48( 3) 878.30 925.00( 4) this work this work this work Various other complex alunite-type [KA13(OH)6( S04),-type] minerals also exist.Table 6. All these phases have heTagonal (trigonal) a and b unit-cell parame!ers of ca. 7.0-7.3 A, as does hexagonal W03 itself6(a=7.298 A). Their c unit-cell dimensions vary depending on the capping group, the layer-repeat motif, and the type of interlayer species. The phases reported earlier7-" are all capped on both faces of the octahedral sheets by tetrahedral groups, 3 4 5 6 W.T. A. Harrison, L. L. Dussack and A. J. Jacobson, Inorg. Chem., 1995,34,4774. W. T. A. Harrison, L. L. Dussack and A. J. Jacobson, Acta Crystallogr., Sect. C, 1995, in the press. W. T. A. Harrison, L. L. Dussack, T. Vogt and A. J. Jacobson, J. Solid State Chem., 1995, in press. B. Gerand, G. Nowogrocki, J. Guenot and M. Figlarz, J. Solid and all crystallize in the centrosymmetric space group R3m. They all consist of three-layer (ABCABC ...) repeat motifs in the c unit-cell direction. The much larger c unit-cell parameter observed for the potassium antimony phosphate hydrate phase," compared to the alunite-type [KAl,(OH),( SO,),] 7 8 9 State Chem., 1979,29,429. T. Kato, Mineral. J.,1987, 13,390. S. Menschetti and C. Sabelli, Neues Jahrb.Mineral Monatsh., 1976,406. Y. Cudennuc, A. Riou and A. Bonnin, Rev. Chim. Miner., 1980, 17, 1158. phases:-'' arises from the different arrangement of the inter- 10 G. Johannson, Ark. Kemi, 1963,20,343. layer K/H,O species in the former material. The new HTO- type phases1-' all crystallize in non-centrosymmetric space groups and may adopt double (vanadium-containing phases) or single (molybdenum, tungsten) capping as noted above in the introduction. 11 12 13 14 M. Tournoux, M. Ganne and Y. Piffard, J.Solid State Chem., 1992, 96, 141. A. C. Larson and R. B. Von Dreele, GSAS User Guide, Los Alamos National Laboratory, Los Alamos, New Mexico, USA, 1991. C. J. Howard, J.Appl. Crystallogr., 1982, 15,615. N. E. Brese and M. OKeeffe, Acta Crystallogr., Sect. B, 1991, 47, 192. We thank Paul Meloni and Roman Czernuszewicz for assist- 15 B. Gerand, G. Nowogrocki, J. Guenot and M. Figlarz, J. Solid ance in collecting the Raman data. This work was funded by the National Science Foundation (DMR9214804) and the Robert A. Welch Foundation (E-1207). The neutron diffraction experiments were supported by the Division of Materials 16 17 State Chem., 1979,29,429. R. Diehl, G. Brandt and E. Salje, Acta Crystallogr., Sect. B, 1978, 34, 1105. W. Kwak, M. T. Pope and T. F. Scully, J. Am. Chem. SOC.,1975, 97, 5735. Sciences, US Department of Energy, under contract no. DEA- AC02-76CH00016. 18 19 J. K. Burdett, Molecular Shapes, Wiley, New York, 1980. M. Kunz and I. D. Brown, J. Solid State Chem., 1995,115,395. 20 L. L. Dussack, W. T. A. Harrison and A. J. Jacobson, unpublished results. References 21 L. L. Dussack, W. T. A. Harrison and A. J. Jacobson, Muter. Res. 1 J. T. Vaughey, W. T. A. Harrison, L. L. Dussack and A. J. Bull., 1996, in press. Jacobson, Inorg. Chem., 1994,33,4370. 2 W. T. A. Harrison, L. L. Dussack and A. J. Jacobson, Inorg. Chem., Paper 5/03853G; Received 14th June, 1995 1994,33, 6043. J. Muter. Chem., 1996,6( l), 81-87

ISSN:0959-9428    

DOI:10.1039/JM9960600081

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Preparation and characterisation of mesoporous, high-surface-area zirconium(IV ) oxide? Michael J. Hudson* and James A. Knowles Department of Chemistry, University of Reading, PO Box 224, Whiteknights, Reading, Berkshire, UK RG6 2AD Mesoporous, high-surface-area zirconium (~v) oxide may be prepared by incorporation of cationic quaternary ammonium surfactants in the hydrous oxide and subsequent calcination of the inorganic/organic intermediate. The surfactants are incorporated by cation exchange at a pH above the isoelectric point of the hydrous oxide. Powder X-ray diffraction (XRD) patterns of the materials exhibit broad, low-angle reflections indicating some ordering of the structure on a scale of 2.5 nm and above. At room temperature there is no dependence of the scale of ordering on the chain length of the surfactant.However, after calcination to 723 K and above, the ordering is a linear function of the chain length. A scaffolding, rather than a templating, mechanism is invoked to explain this ordering. Calcination of the materials to 723-973 K results in the formation of a mesoporous zirconium(1v) oxide. Materials prepared using n-alkyltrimethylammonium hydroxides of chain lengths C14 and c16exhibit an increase in the pore size distribution with the chain length of the surfactant. BET surface areas of the calcined materials are 240-360 m2 8-l depending upon the chain length of the incorporated surfactant. These values compare favourably with those of calcined ZrO, aerogels. Calcination of the materials to above 1023 K results in crystallisation to tetragonal zirconium(1v) oxide. Crystallisation is complete by 1323 K.It is suggested that the high surface :volume ratio of the zirconium(1v) oxide results in surface-energy stabilisation of the tetragonal as opposed to the monoclinic phase. Recently, there has been interest in mesoporous silicates with well defined pore sizes, prepared by templating with surfactants and subsequent calcination.'.' The mechanism of formation of the silicate and aluminosilicate structures, the MCM-type materials, has been illuminated by the extension of the templat- ing technique to other oxide systems such as W03 and PbO using cationic and anionic surfactants, re~pectively.~ As part of a study to extend further knowledge of these systems, we have attempted to prepare pseudo-template zirconium(1v) oxide using a series of cationic surfactants such as alkyltrime- thylammonium cations, and to characterise the materials formed.Templated structures do not seem to form in this system; however, some ordering does occur which is explained by a proposed scaffolding mechanism. Solutions of ZrIV contain cationic Zr species, the interaction of which with anionic surfactants (such as dodecyl sulfate) results in rapid precipitation of the salt. Interaction of cationic surfactants and freshly precipitated hydrous zirconium(1v) oxide under basic conditions above the isoelectric point (IEP) of the hydrous oxide results in the incorporation of the cations by ion exchange. Hydrous oxides can exchange both anions and cations, depending upon the pH of the s~lution.~.~ At a pH above the IEP of the oxide (i.e.the point at which the net charge on the oxide surface is zero), hydrous oxides act as cation exchangers. The IEP of hydrous zirconium(1v) oxide is 6-8 in aqueous solution, depending upon the other ions in solution. The site of cation exchange is the hydroxy groups present on the surface of the hydrous oxide. The cation exchange process in basic solution may be represented by eqn (1). pH >IEP M-OH M-O-+H+ pH =IEP Under highly basic conditions, there is a degree of dissolution and reprecipitation of hydrous zirconium(1v) oxide, allowing a degree of ordering to develop over time.For this reason we concentrated on the cationic surfactant system; other workers t Preliminary communication: M. J. Hudson and J. A. Knowles, J. Chem. SOC.,Chem. Commun., 1995, 2083. are beginning to investigate the corresponding zirconium(1v) oxide/anionic surfactant system.6 In addition to interest in templated structures, mesoporous, high-surface-area zirconium(1v) oxide has potential uses as a catalyst and catalyst support. It has been used as a support for rhodium-catalysed hydrogenation of carbon oxides to hydrocarbons7 and for copper(I1) oxide catalysts in the forma- tion of methanol from synthesis gas8 Zirconium(1v) oxide is an active catalyst for the isomerization and hydrogenation of b~t-l-ene,~and is a superacid when sulfate-promoted.1° Zirconium(1v) oxide for catalytic applications is prepared by calcination of hydrous zirconium(1v) oxide to temperatures typically above 700K, or by high temperature hydrolysis of zirconium tetrachloride.Hydrous zirconium(1v) oxide has been prepared by both sol-gel methods and by precipitation using alkoxides and salts as precursors. Sol-gel methods of synthesis depend upon hydrolysis and condensation of the metal precur- sor. The relative rates of the processes determine whether a gel or particulate precipitate is formed. The products consist of Zr-0-Zr frameworks with solvent molecules (usually alcohol and/or water) filling the internal pores of the frame- work. Calcined hydrous zirconium(1v) oxide has low surface area and porosity since the drying of a hydrous metal oxide causes stress on the metal oxide framework by the capillary pressure at the liquid-vapour interface in the pores of the material.This stress tends to collapse the pores, reducing the surface area of the dried gel. The problem has been over- come by the preparation of zirconium(1v) oxide Aerogels are high-surface-area, high porosity materials pre- pared by sol-gel synthesis and subsequent supercritical drying. Extraction of the pore liquid by supercritical fluids reduces the collapse of the gel structure since supercritical fluids do not form liquid-vapour interfaces and do not exert drying stress on the frame~ork.l~*'~ Experimental Preparation The first step to the preparation of the mesoporous zir- conium(1v) oxide samples was the preparation of surfactant- J.Muter. Chem., 1996, 6(l), 89-95 exchanged hydrous zirconium(1v) oxide. Two methods were used to prepare the precursors. In the first method, solutions of Zr"' and alkyltrimethylammonium halide were made basic by the addition of aqueous ammonia. In the second method, solutions of Zr" and alkyltrimethylammonium halide were made basic by addition of the appropriate alkyltrimethylam- monium hydroxide. The source of Zr" in all cases was aqueous solutions of zirconyl chloride octahydrate (ZrOCl2.8H2O). The cations exchanged into the hydrous zirconium(1v) oxide were alkyltrimethylammonium cations. The various preparations differed in the aliphatic hydrocarbon chain length of the surfactant used (C8, cl0, Cl2, Cl4, c16or c18)and the inclusion of an auxiliary organic species (pyrene). A typical preparation of cetyltrimethylammonium (ctma, c16 chain)-exchanged zir- conium(1v) oxide using the first method is described below.Method 1. To an aqueous solution of ZrOC12-8H20 (0.1mol dm-3 in Zr", 100 cm3, Aldrich) was added an aqueous solution of cetyltrimethylammonium chloride (0.1 mol 86 cm3, Aldrich). The combined solution was stirred for 15 min (pH = 0.63). Aqueous ammonia (0.880 g ~m-~) was slowly added with continuous stirring to pH 11.48. Hydrous zirconium(1v) oxide precipitated as a gelatinous solid a few moments after the addition of the base. The mixture was stirred for 60min and then placed in a bath thermostatically maintained at 363 K.Immediately the reaction mixture had reached thermal equilibrium, the reaction flask was sealed. The reaction mixture was maintained at this temperature and continuously stirred for 90 h. After this time, the mixture was allowed to cool, the solid was filtered off under suction and washed with water and acetone until free of surfactant. The samples were washed repeatedly after filtering to ensure that no precipitated alkyltri- methylammonium salts were present. The white powder was dried at 333 K for 20 h. Yield was 2.17 g (79% based on the composition below). Thermogravimetric analysis indi- cated that the formula of this material approximated to ZrOl~87(Octma)o,26~0.7H20(ctma =hexadecyltrimethylam-monium), with the hydration of the material depending to a certain extent upon the ambient moisture.This sample was labelled ZC161a, the parts of the label referring to zirconium(1v) oxide, the chain length of the incorporated alkyltrimethylam- monium cation, the method of preparation (Method 1 or 2) and the sample number, respectively. See Table 1 for a selection of materials prepared in this study. The alkyltrimethylammon- ium cations used were: (28, octyltrimethylammonium; Cl0, decyltrimet hylammonium; C 12, dodecyltrimet hylammonium; C14, tetradecyltrimethylammonium; c16, hexadecyltrimethyl- ammonium; and c18, octadecyltrimethylammonium. Preparation including auxiliary organic species. A hexadecyl- trimethylammonium-exchanged zirconium(1v) oxide was pre- pared using the above method but with the addition of an auxiliary organic species, pyrene, in a mole ratio of 1:1 ctma :pyrene.The material was labelled zc161P. Powder XRD data indicated the presence of crystalline pyrene in the product. Table 1 Thermogravimetry of the amount of incorporated alkyltrime- thylammonium species in the materials as a function of chain length initial reaction reaction Zr :surfactant label PH temperature/K time/h ratio of product zc,1 11.65 11.63 363 363 92 92 1 :0.082 1 :0.106 11.62 363 90 1 :0.131 11.65 363 92 1 :0.183 11.36 363 90 1:0.260 11.48 363 94 1 :0.251 11.63 363 92 1 :0.341 11.76 363 92 1 :0.175 11.67 363 91 1 :0.194 Thermogravimetry (TG) showed that calcination of the material to 723 K removed the organic species.The results of surface area (BET) analysis of the calcined material are pre- sented below. Method 2. This preparation was similar to method 1 but, in this case, the base used was alkyltrimethylammonium hydrox- ide, which was prepared by stirring a 25 mass% solution of alkyltrimethylammonium bromide in water with an anion exchanger in the hydroxide form (Duolite A113, Aldrich). The stock ZrIV solution (50 cm3, 0.1 mol dm-3) was stirred with a solution of the alkyltrimethylammonium halide (25 cm3, 0.1 mol dmP3). The appropriate alkyltrimethylammonium hydroxide solution was added to pH 11.6. The rest of the preparation is as in method 1.The samples prepared using this method were zirconium(1v) oxide/tetradecyltrimethyl-ammonium and zirconium(1v) oxide/hexadecyltrimethyl-ammonium. Calcination of the materials The materials were calcined at temperatures of 723-973 K in static air. Typically ca. 0.5 g of the exchanged oxide was placed in an open platinum crucible. Dwell times were 120 min in all cases. Ramp rates appeared to have little or no affect on the properties of the materials. After calcination the materials were allowed to cool to room temperature in a desiccator over anhydrous calcium chloride. Charac terisa tion The a1 k yl trime t h ylammonium-exchanged hydrous metal oxides were studied using X-ray powder diffraction (XRD; Spectrolab CPS Series 3000 120, using Ni-filtered Cu-Ka radiation), simultaneous thermogravimetry-differential ther-mal analysis (TG-DTA; Stanton-Redcroft STA 1000 using lidless, platinum pans with recalcined alumina as DTA refer- ence) and Fourier transform infrared spectroscopy (FTIR; Perkin-Elmer 1100B spectrophotometer).TG-DTA was used to estimate the amount of alkyltrimethylammonium incorpor- ated in the zirconium(1v) oxide. This technique has been used in the determination of incorporated surfactant in studies of WO,/surfactant mesostructures.' All experiments were per- formed in static air at a ramp rate of 20K min-' to ensure the complete oxidation of the organic species. FTIR spectra were recorded on the materials in the form of KBr discs before and after calcination.The nitrogen adsorption isotherms and BET surface areas of the calcined materials were determined at 77 K by means of a Micromeritics Gemini 2370 surface area analyser. All adsorption experiments were performed in tripli- cate. BET surface areas and other BET parameters were calculated from five points in the relative pressure range 0.10-0.30 assuming a cross-sectional area of 0.162 nm2 for the nitrogen molecule. The total pore volume of the materials (V,) was estimated from the upper plateau in the nitrogen adsorp- tion isotherm, or from the amount of nitrogen adsorbed at P/Po=0.95, assuming a liquid density of nitrogen at 77 K of 0.808 g cm-3. The mesopore size distribution was calculated from the desorption branch of the isotherm by the Barrett, Joyner and Halenda (BJH) method." Results and Discussion FTIR spectroscopy Fig. 1 shows the FTIR spectrum of ZC161a before calcination.The broad bands between 3000 and 3500 cm-' in the spectrum are due to 0-H stretches of water associated with the material. The two sharp bands at 2800-2900 cm-' are due to C-H stretches of the hydrocarbon chain of incorporated hexadecyltrimethylammonium (ctma).16 These bands provide 90 J. Mater. Chem., 1996, 6(l),89-95 Fig. 1 FTIR spectrum of ZC,,la at room temperature direct evidence of the incorporation of alkyltrimethylammon- ium cation into the hydrous oxide. The broad bands at ca. 1600 cm-' in both spectra are due to the scissors mode of associated water. The series of sharp bands at ca.1500cm-1 are due to C-H deformations of the incorporated ctma. The broad band centred on 450 cm-' is due to Zr-0 stretches of the zirconium(1v) oxide lattice. Calcination to 723 K removed the hydrocarbon chain from the material. TG-DTA Fig. 2 shows typical TG-DTA results for ZrOl~,,(Octma),~,,~0.7H20(ZC,,la). Mass loss up to 447 K corresponds to the loss of loosely bound water (6.8%).Between 447 K and 891 K there are three regions of mass loss (34.2% in total) associated with three exothermic DTA peaks. These correspond to different stages in the oxidation of the alkyltrime- thylammonium species. Above 891 K there are no further mass losses. At this point the residue (58.00/) is anhydrous ZrO,. A small exothermic DTA peak at ca.1013 K is due to crystallis- ation of the material (see powder XRD results, below). The TG-DTA results for ZC161a are typical of the other Zr02/alkyltrimethylammoniumsamples, with the amounts of alkyltrimethylammonium incorporated depending upon the preparative conditions and the chain length of the exchanging surfactant. The amount of incorporated alkyltrimethylammon- ium cation increases regularly with chain length. The maximum amounts of alkyltrimethylammonium incorporated (ratio of Zr : octadecyltrimethylammonium, < 0.34) are consistent with literature values of the cation-exchange capacity of hydrous zirconium(1v) oxide: one exchangeable proton for every three or four Zr atoms.' Hydrous zirconium(1v) oxide prepared using aqueous ammonia as the base in the absence of alkyltrimethylammon- ium exhibits a small mass loss associated with a weakly 1101-f 12 4 n 90 2 70 Fig.2 Simultaneous TG-DTA results for ZC,,la in static air, ramp rate 20 K min-' exothermic DTA peak at 733-813 K.This peak is assigned to the oxidation of ammonium species incorporated into the hydrous oxide. The event is distinguishable from the events associated with alkyltrimethylammonium oxidation. Neverthe- less it is likely that some NH4+ is retained in the materials prepared using method 1. Isothermal TG showed that calcination of the exchanged materials at 723 K for 2 h eliminated the majority of organic species in the materials. For nitrogen adsorption studies, the materials were typically calcined at 723 K for 2 h, a temperature consistent with obtaining the highest achievable surface area at the lowest organic content.Powder XRD Powder XRD patterns of the uncalcined, alkyltrimethylam- monium-exchanged materials exhibited single broad reflections at low 28 values. Similar powder XRD patterns with single, broad, low-angle reflections have previously been observed in templated silicates.' In the silicate case, the single reflection was assigned to the hkl= 100 reflection of a hexagonal cell with the absence of higher order Bragg reflections owing to the absence of high order in the mesostructure. The powder patterns of zirconium(1v) oxide samples before calcination exhibit single, broad, low-angle reflections with d-spacings shown in Table 2.The values are lower than those reported in the literature for templated materials2 and are independent of the hydrocarbon chain length of the incorpor- ated surfactant. There are no higher order Bragg reflections observed in the powder patterns, either because the pore walls are amorphous or because of a lack of correspondence between the structures of adjacent pores. After calcination of the materials to 723 K, the powder XRD patterns exhibit similar, single, broad peaks at low 28 with the scale of the ordering depending directly upon the chain length of the incorporated alkyltrimethylammonium cation. Fig. 3 shows the powder XRD patterns for ZC81, ZC141 and ZC181. The derived d-spacings, as a function of chain length of exchanged alkyltrimethylammonium cation, are shown in Fig.4. The points lie along a straight line, indicating that the scale of ordering of the structure of the materials after calci- nation is directly dependent upon the size of the alkyl group of the incorporated cation. See the General Discussion for an interpretation of these results. Calcination at higher temperatures Powder XRD results for ZC161a calcined at 873 K, 1023K and 1323 K are shown in Fig. 5(a), (b) and (c), respectively. The low-angle reflection disappears by 1023 K as the material crystallises. Crystallisation is complete by 1323 K, with 21 observed reflections indexed to a tetragonal (ptimitive) unit cell (refited cell parameters: a = 5.0838(5) A and c = 5.1887(9)A, where the number in brackets is the standard deviation in the last significant figure).The crystallisation occurs at a lower temperature (1013 K by DTA) than is the Table2 d-Spacing of the low-angle powder XRD reflections as a function of chain length of the alkyltrimethylammonium cation for the materials before calcination label chain length d-spacing/A 8 19.15 10 18.20 12 33.70 14 25.71 16 35.23 16 33.42 18 28.18 14 17.21 16 32.76 J. Muter. Chern., 1996, 6(l), 89-95 I1""1""1""1""l""l""l""l""I 012345678 2e/degrees Fig. 3 Powder XRD patterns of ZC81, ZC,,1 and ZC161a after calci-nation at 723 K in air 3.4 3 3.2 3.0 c9 2.8.-12.6 b 2.4 2.2 1 2.0 ? 1 I I I I I 1 6 8 10 12 14 16 18 20 surfactant chain length Fig.4 d-Spacing of the low-angle powder XRD reflections as a function of chain length of alkyltrimethylammonium cation for the materials after calcination in air at 723 K case with other types of hydrous zirconium(1v) oxide. It is likely that the tetragonal form is stabilised by the higher surface energy arising from the thin walls of the porous material, leading to a higher surface:volume ratio than in other ZrO, materials.17 Nitrogen adsorption studies In order to investigate surface areas and porosities, the mate-rials were calcined at various temperatures. Isothermal TG analysis shows that calcination of the composite materials at 723 K for 2 h removes >99% of the incorporated organic phase.All nitrogen adsorption isotherms on calcined zirconium(1v) oxide/alkyltrimethylammonium materials exhibited type IV behaviour, typical of a mesoporous material." The desorption branch of the isotherm exhibited various types of hysteresis, depending upon the calcination temperature. Calcination at 723 K invariably resulted in type H2 hysteresis. Typical iso-therms of calcined ZCl,la and calcined ZC,,l are shown in Fig. 6. Isotherms such as these are similar to those of calcined zirconium(~v)oxide aerogels that have been previously reported in the 1iterat~re.l~The isotherms are unlike those of templated silicates,20with the absence of the characteristic step indicating a sharp pore size distribution.This suggests a distinct class of materials formed in this system, whose adsorp-tion properties are more akin to those of calcined aerogels. The derived BET surface areas of the calcined materials varied with the calcination' temperature and with the method of preparation. In general, the surface areas of the materials I """"",'"""'~~r~,,,,, I ,,,,,,, I 0 10 20 30 40 50 60 70 80 90 100 I....r....l....l....l....l.,..l...~...l.,..l....l1 ww~~'..I.~~.1"~I"""~I~~"I"~~I""I"~~I"'~~ 0 10 20 30 40 50 60 70 80 90 100 I....I....I .... I....I....l....l....I....I....I.... ktl dll Ill 19% om 1.5M 2%' 11124U 1.815 113 I799 311 I151 222 1535 u)( 1477 400 1.197 311 1x1 331 I I77 31I 20) 1.161 402 I156 410 1.141 224 1.131 200 J 311 422 4U 1052 * __1 I Fig. 5 Powder XRD patterns of ZCldla at room temperature after calcination at (a) 873, (b) 1023, (c)1323K calcined above 723 K were lower than those calcined at the lower temperature.This is consistent with the results of BET surface area studies on calcined ZrO, aerogels.21The maximum reproducible surface area for a zirconium(1v) oxide prepared in this study is 360 m2 g-' for zirconium(1v) oxide/hexadecyl-trimethylammonium and auxiliary pyrene (ZCI61P) calcined at 723 K. This value compares favourably with the BET sur-face areas of both precipitated zirconium(1v) oxide and zir-conium(1v) oxide aerogels calcined at these temperatures." As far as we are aware, these values are the highest observed BET surface areas of zirconium(1v)oxide calcined at these tempera-tures.Table 3 gives values of the BET parameters and total pore volume (5)for the materials prepared in this study. Calculated BJH pore size distributions (desorption branch) for the materials prepared by method 1 do not show a consistent variation in pore size with chain length of incorpor-ated alkyltrimethylammonium cation. Typical pore size distri-butions for ZC121, ZC161a and ZCI81 calcined at 723K are shown in Fig. 7(u). It seems that C18 chains, in contrast to shorter chains, increase the pore size of the calcined product. This is probably due to the large steric effects of the bulky chain expanding the pores in the forming metal oxide.BJH pore size distributions of materials prepared by method 2 92 J. Muter. Chern., 1996,6(l), 89-95 I3.0 I c I-%-I a 1.o "E0:8 0.0 0.2 0.4 0.6 0.8 1.0 0.5 300 200 100 0.0 0.2 0.4 0.6 0.8 1.0 PIPo Fig. 6 (a) Nitrogen adsorption isotherm at 77 K on ZCl,la calcined at 723 K. 0, adsorption; 0, desorption; (b) nitrogen adsorption isotherm at 77 K on ZC,,l calcined at 723 K in air. (ZC142and ZC162) are shown in Fig. 7(b),together with the results for ZCl& The chain length does, in this case, seem to determine the pore size distribution of the calcined material. In this case only alkyltrimethylammonium cations are available for incorporation during the formation of the gel (there are no NH4+ cations added).As a result, alkyltrimethylammonium cations are directly incorporated as the gel forms, directing the formation of the gel around the surfactant leading to a more completely ordered gel. General Discussion The results presented in this paper show that there is no evidence that the zirconium(1v) oxide/alkyltrimethylammon-ium system is similar to the templated silicate system, i.e. forming a relatively ordered, templated oxide at room tempera- ture. The chain length of the incorporated alkyltrimethylam- monium cation does not affect the powder XRD pattern of the zirconium(1v) oxide/surfactant complex in a predictable way that could be assigned to mesostructural ordering of the metal oxide around the organic template at room temperature.0.0 10 500 3.0 2.5 2.0 1.5 1.o 0.5 0.0 10 100 500 pore diameterlA Fig. 7 (a) Barrett, Joyner and Halenda (BJH) pore size distributions for zirconium(1v) oxide materials calcined at 723 K as a function of the chain length of incorporated alkyltrimethylammonium cations; (b)pore size distributions for zirconium(1v) oxides calcined at 723 K as a function of chain length of incorporated alkyltrimethylammonium cations. C14 and c16 materials were prepared using the appropriate alkyltrimethylammonium hydroxide as base. However, XRD results obtained after calcination show that the ordered components of the material are of a size determined by the incorporated alkyltrimethylammonium cations.We propose a scaffolding-type mechanism, together with a drying control effect of the incorporated surfactants in order to explain the increase in surface area, compared with zirconium(1v) oxide formed in the absence of the surfactant. The apparent conflict in the above observations is due to the scaffolding-type mech- anism occurring only in the ordered part of the material (observable by XRD). The drying control effect (nitrogen adsorption studies) occurs throughout the material, in both ordered and disordered regions. Table 3 BET parameters and total pore volume (V,) of the calcined zirconium(1v) oxides as a function of calcination temperature and chain length of the incorporated alkyltrimethylammonium cation calcination BET surface nm( BET)(label temperature/K area/m2 g-' mmol g- c(BET) v,/cm3 g-' 723 247 f 1 2.54 80 0.27 723 175f2 1.79 102 0.32 723 238 f 3 2.45 87 0.30 723 310f5 3.18 83 0.47 723 274+ 1 2.8 1 102 0.24 723 300f 1 3.04 64 0.29 723 312k3 3.16 71 0.34 973 164k 1 1.68 65 0.24 723 329f4 3.39 58 0.33 723 360f 4 3.69 58 0.44 823 264 2 2.82 70 0.35 723 313+3 3.40 86 0.60 723 270 k2 2.77 53 0.31 723 333f3 3.42 69 0.43 J.Mater. Chern., 1996, 6(l), 89-95 93 Controlled drying effect leading to higher-surface-area zirconium(IV ) oxide When a hydrous zirconium(1v) oxide gel is dried, the surface tension of the water contained in the pores within the Zr-0-Zr network tends to collapse the network, reducing both the pore volume and surface area of the dried gel.The collapse of the network may be prevented by supercritical drying of the hydrous gel. Since supercritical fluids have no liquid-gas interface they do not exert a surface tension on the Zr-0-Zr network. Aerogels prepared in this manner are known for many metal oxide systems. Typically, they have high surface areas and pore volume both before and after calcination." In the system described here, incorporation of alkyltrimethyl- ammonium ions into the zirconium(1v) oxide gel by cation exchange and subsequent calcination also lead to high-surface- area, high porosity solids. Surfactants incorporated into hydrous zirconium(1v) oxide act as chemically bound, drying control chemical additives; similar systems are effective in reducing the shrinkage of gels.22*23 Surfactants in aqueous solution reduce the surface tension of the solution; the efficiency of this reduction, above the critical micelle concentration (CMC), depends, in an approximately linear fashion, upon the chain length of the hydrocarbon chain in a homologous series of surfactants.Neglecting effects due to angle of contact, the capillary pressure of the pore liquid is proportional to the surface tension and surface :volume ratio of the ~api1lary.l~ The incorporated surfactants reduce the surface tension of pore water with which they come into contact, regardless of the ordering of the surfactant cations in the material. As a result, the surface area of the calcined mesoporous material is predicted to be proportional to the chain length of the incor- porated surfactant.The results presented here reflect this prediction only for the materials with incorporated surfactants of chain length C12-C18, the shorter-chain materials have surface areas higher than expected. The longer-chain-length surfactants are more strongly incorporated in zirconium(1v) oxide (see results of thermal analysis) reducing the surface tension less than might be expected. In contrast, the shorter- chain-length surfactants are less strongly incorporated into zirconium(1v) oxide, with the consequence of more interaction hydrous zirconium( IV) oxide exchanged with ammonium cations I90h RNMeJ stage 1 363K km4+ stage 2 at the pore water-air interface, consequently reducing the surface tension further.Proposed scaffolding mechanism, ordering during calcination The alkyltrimethylammonium cations exchange into the struc- ture of the zirconium(1v) hydrogel during or after it is formed, but they do not directly determine the scale of the ordering at room temperature. However, as heating proceeds and water is lost, the drying material shrinks under the influence of capillary pressure until the steric interaction between the incorporated organic cation and the Zr-0-Zr network prevents further shrinkage. The result is a scaffolding, rather than a templating, effect with the scale of ordering after calcination determined by the effective size of the incorporated cation.The proposed scaffolding mechanism is a templating-like mechanism and may be considered a more general case of the templating mechanism described by Kresge and coworkers.'.2 A schematic of the proposed controlled-drying and scaffolding mechanism shows a cross-section of the zirconium(1v) oxide gel (Fig. 8). Stage 1 involves incorporation of alkyltrimethylammonium cations in hydrous zirconium(1v) oxide, some ammonium ions and water molecules are retained. Stage 2 involves initial heating of the organic/inorganic material. The water molecules are lost on drying, the structure shrinks but the drying-control effect of the surfactant reduces the tension on the porous zirconium(1v) oxide. The dehydroxylation reactions of the hydrous zirconium(1v) oxide proceed, giving a partially ordered, scaffolded structure prevented from collapsing by the surfactant.During calcination, in stage 3, the organic species are burnt off, leaving mesoporous zirconium (~v) oxide. Conclusion Zirconium(1v) oxides incorporating alkyltrimethylammonium cations of different chain length were prepared. BET surface areas of the calcined materials (723 K) are 240-360 m2 g-', depending upon the chain length of incorporated surfactant. These surface areas are higher than those reported in the literature for hydrous zirconium(1v) oxide and zirconium(1v) oxide aerogels calcined at similar temperatures. The calcined stage 3 calcination 973> TM > 723 -402.-30 calcination T >I023K tetragonal zirconium( rv) oxide Fig. 8 Schematic diagram of the proposed scaffolding mechanism in long-chain alkyltrimethylammonium-incorporatedhydrous zirconium(rv) oxide. See text for a fuller explanation. 94 J. Mater. Chem., 1996, 6(l),89-95 materials were mesoporous, and the nitrogen adsorption properties of the calcined materials are more similar to those of aerogels than to templated structures. The large surface areas have been related to the controlled-drying effect that invokes the presence of surfactant molecules exchanged on to 3 4 U. Ciesla, D. Demuth, R. Leon, P. Petroff, G. Stucky, K. Unger and F. Schuth, J. Chem. SOC.,Chem. Commun., 1994,1387. A. Ruvarac, Group I V Hydrous Oxides-Synthetic Ion Exchangers, in Inorganic Ion Exchange Materials, ed.A. Clearfield, CRC Press, Boca Raton, FL, 1982, ch. 5, p. 141. G. A. Parks, Chem. Rev., 1965,65,177. the interior of the porous hydrogel, leading to a reduction in F. Schuth, personal communication, 1995. surface tension of the capillary water with the result of reducing the stress experienced by the gel framework during drying. The evidence does not support a templating mechanism as occurs in silicates; instead, after calcination there is an ordering of regions of the materials owing to a scaffolding mechanism. 10 11 T. Izuka, Y. Tanaka and K. Tanabe, J. Catal., 1982,76, 1. Y. Amenomiya, Appl. Catal., 1987,30, 57. G. M. Pajonk and A. El Tanany, React. Kinet. Catal. Lett., 1992, 47, 167. M.Hino and K. Arata, J. Chem. Soc., Chem. Commun., 1980,851. D. A. Ward and E. I. KO,Chem. Mater., 1993,5,956. The d-spacing of the calcined zirconium(1v) oxide is a linear function of the chain length of the incorporated surfactant. Calcination of these zirconium(1v) oxides to higher tempera- tures results in crystallisation to tetragonal zirconium(1v) oxide. Crystallisation begins at ca. 1023 K (by DTA) and is complete by 1323 K. It is suggested that the high surface:volume ratio 12 13 14 15 S. J. Teichner, G. A. Nicolaon, M. A. Vicarini and G. E. E. Gardes, Adv. Colloid Interface Sci., 1976,5,245. C. J. Brinker and G. W. Scherer, Sol-Gel Science, The Physics and Chemistry of Sol-Gel Processing, Academic, New York, 1990. S. S. Prakash, C. J. Brinker, A. J. Hurd and S. M. Rao, Nature, 1995,374,439. E. P. Barrett, L. G. Joyner and P. P. Halenda, J. Am. Chem. SOC., of the mesoporous, high-surface-area zirconium(rv) oxide results in surface-energy stabilisation of the tetragonal as opposed to the monoclinic phase. 16 17 1951,73,373. K. Nakamoto, Infrared and Raman Spectra of Inorganic and Coordination Compounds, 4th edn., Wiley, New York, 1986. A. F. Wells, Structural Inorganic Chemistry, 5th edn., Clarendon We wish to thank the European Commission under the Brite 18 Press, Oxford, 1984, p. 542. S. J. Gregg and K. S. W. Sing, Adsorption, Surface Area and Euram program for funding this work (grant no. BRE2.CT92.0198). 19 20 Porosity, 2nd edn., Academic, London, 1991. M. Schneider and A. Baiker, J. Mater. Chem., 1992,2,587. P. J. Branton, P. G. Hall and K. S. W. Sing, J. Chem. Soc., Chem. Commun., 1993,724. References 21 22 L. K. Campbell, B. K. Na and E. I. KO,Chem. Mater., 1992,4,1329. W. D. Klingery and J. Francl, J. Am. Ceram. SOC.,1954,37,596. 1 C. T. Kresge, M. E. Leonovicz, W. J. Roth and J. C. Vartuli, US Pat., 5098684,1992. 23 J. Zarzycki, M. Prassas and J. Phalippou, J. Mater. Sci., 1982, 17, 3371. 2 C. T. Kresge, M. E. Leonovicz, W. J. Roth, J. C. Vartuli and J. S. Beck, Nature, 1992,359,710. Paper 5/040065;Received 21st June, 1995 J. Mater. Chem., 1996, 6(l), 89-95

ISSN:0959-9428    

DOI:10.1039/JM9960600089

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Synthesis and catalytic properties of perovskite-related phases in the La-Sr-Co-Cu-Ru-0 system Yasutake Teraoka," Hiroshi Nii," Shuichi Kagawa," Kjell Janssonb and Mats Nygrenb "Departmentof Applied Chemistry, Faculty of Engineering, Nagasaki University, Nagasaki 852, Japan bDepartment of Inorganic Chemistry, Arrhenius Laboratory, Stockholm University, S-10691 Stockholm, Sweden A series of oxides of the composition La,~,Sr,~,Co, -,,Cu,Ru,O, with the perovskite structure were synthesized and characterized by their X-ray diffraction (XRD) patterns and by element analysis in a transmission electron microscope (TEM) equipped with an energy-dispersive spectrometer (EDS). The XRD and TEM-EDS studies confirmed that almost monophasic samples with the perovskite structure having rhombohedral (hexagonal) symmetry were obtained for 0<y <0.20, although the simultaneous formation of a by product with the K,NiF,-type structure was observed for high y values.The lattice parameters and rhombohedral distortion increased monotonically with increasing y value due to the fact that the average ionic radius of the substituting Cu/Ru pair is larger than that of the host Co ion. Temperature-programmed desorption (TPD) studies showed that the substitution of the Cu/Ru pair for Co caused a decrease in the amount of oxygen desorbed from the bulk of the oxide, and the catalytic activity for CO oxidation and CO-NO reactions decreased. The Cu/Ru substitution did, however, selectivity increase the N, formation and decreased that of N20 in the CO-NO reactions.Synthesis of K,NiF,-type oxides with the overall compositions (Lao.8Sro~2),Coo~50Cuo~50~zRuz04with z =0.05 and 0.10 was also attempted. In both cases, tetragonal K2NiF4-type oxides having a composition close to (Lao,8Sr,~,)2(Co,Cu)o~97R~o~0304were obtained, implying that the solubility of Ru into the given K2NiF4-type framework is very low. Perovskite-type oxides (ABO,) are an important class of compounds in the realm of metal oxides, and they have been extensively studied with respect to their synthesis, structure and various physical and chemical properties.'.2 If the basic criteria of charge neutrality and geometric (ionic size) factors are satisfied, one can prepare many ternary or multicomponent oxides with the perovskite-type structure.Proper combinations of metal cations can stabilize cation or anion vacancies in the perovskite framework and anomalous formal oxidation states of metal cations can be obtained. These structurally versatile features offer a wide range of possibilities, not only in designing new perovskite compounds but also in tailoring their chemical and physical properties. In catalytic applications, ternary perovskite-type oxides of the composition ABO,, and quaternary ones such as A,-,A,'BO,, are most frequently although a few articles dealing with quinary A, -,Ax'Bl -,BY'03 perov-skite systems have been p~blished.~.~ Very recently, Skoglundh et aL8 reported the catalytic properties of a six-component perovskite catalyst of an entirely new composition, La, -,SrXAll -,,Cu,Ru,O,.This compound is reported to be active for the catalytic conversion of NO and CO to N, and CO,, respectively, and might be used as an active ingredient of a three-way catalyst (TWC) for automobiles. This result has triggered the present investigation concerning the syn- thesis and catalytic properties of the Co-based analogue, La,~,Sr,~,Co, -,,Cu,Ru,O,, partly because Co-based perov- skites are generally more catalytically active than Al-based ones. We have also attempted to synthesize K2NiF4- type (K-type) oxides of the nominal composition (La0.8Sr0.2)2C00.5Cu0.5 -zRUzo4* Experimenta1 Synthesis Perovskite-type and K-type compounds of compositions La0.,Sr0.,Co, -,,Cu,Ru,O,, with 0 dy d0.20 at intervals of -zR~z04,0.05, and (Lao.8Sro~2)2Coo~5Cuo~5 with z =0.05 and 0.10, were synthesized from p.u.grade La(NO,), 6H20, Sr(N03),, C0(N03), *6H20, CU(NO,)~ .3H20 and RuCl,. Appropiate proportions of the starting chemicals were dis-solved in water and this solution was evaporated to dryness at 120"C, and then heat-treated at 550°C for several hours until evolution of brown NO, gas was no longer observed. The resulting powder was ground, pelletized and then calcined at 900°C for 24 h, at 1000°C for 24 h and at 1100°C for 24 h. Perovskite-type oxides of the formula Lao,8Sro.2-Co, ~,,Cu,Ru,O, are hereafter designated as CCRlOOy, i.e. Lao~8Sro~2Coo~9Cuo~05Ruo~0503is abbreviated as CCR5. Considering the charge neutrality and variable valencies of the B-site cations, it is easily recognized that both the perov- skite-type and K-type oxides may have non-stoichiometric compositions. In this article, however, nominal formulae with the stoichiometric composition of oxygen are used, because the exact oxygen content has not yet been determined. Characterization X-Ray powder diffraction (XRD) patterns were recorded in a Guinier-Hagg focusing camera, using Cu-Kal radiation and Si as an internal standard.The obtained films were evaluated with a computerized scan system.' A portion of each sample was ground in butanol and then dispersed on a holey carbon film supported on an Ni grid, and analysed in a transmission electron microscope (TEM; JEOL 2000FX) equipped with an energy-dispersive spectrometer (EDS; LINK AN10000).The metal content of the samples was determined by the spot meas- uring technique and at least ten individual grains were ana- lysed. The obtained mean values of x and y in La, -,Sr,Co, -,,Cu,Ru,0, were within f0.01, in agreement with the weighed-in perovskite compositions. Oxygen desorption and catalytic activity La,.8Sr,.2Col -2,Cu,Ru,03 samples with y =0, 0.05 and 0.10 synthesized at 900°C for 48 h were used for these measurements. Temperature-programmed desorption (TPD) studies of oxygen desorption were performed as follows. (i) After being J. Muter. Chem., 1996, 6(l),97-102 granulated to 42-80 mesh sized particles and mounted in a quartz-glass reactor, each sample (0.5 g) was exposed to vacuum for 30 min and then to 13 kPa of oxygen for 30 min at 800"C, and allowed to cool to room temperature in the same atmosphere.(ii) Then, after changing the atmosphere to helium with a stream rate of 30 cm3 min-', the samples were heated at a constant rate of 10°C min-', and the oxygen release was monitored with a thermal conductivity detector (TCD). The CO oxidation and CO-NO reaction studies were carried out in a fixed-bed flow reactor, over 0.2 g of catalyst granules (20-60 mesh) fed at a rate of 30 cm3 min-' with the following gases: (i) CO (0.5 ~01%) and 0, (5%) balanced with He; (ii) CO (0.5%) and NO (0.5%) balanced with He. After reaching steady-state conditions at each temperature, the gas composition before and after passing through the catalyst bed was analysed by a TCD-type gas chromatograph (Shimadzu GC-8A) with a Porapak Q column for separating CO, and N,O and a 5A molecular sieve column for separating O,, N,, NO and CO.Results and Discussion Preparation and crystal structure of La,~,Sr,~,Co,-~,,Cu,,Ru,,03 La,~,Sr,~,Co, -,,Cu,Ru,O, samples with 0 <y d0.20 were syn- thesized at 900, 1000 and 1100°C. When calcined at 900°C, monophasic samples were obtained, but the diffraction peaks were too diffuse to allow accurate determinations of the lattice parameters. As the calcination temperature increased, the peaks became sharper and better resolved, most probably due to an increase in crystallinity and/or crystallite size.At 1100 "C a small additional peak assignable to a phase of the composition with the K,NiF,-type structure was observed. Typical XRD patterns of La,,8Sr,.,Col -,,CuyRu,03 prepared at 900, 1000 and 1100°C are shown in Fig. 1. The EDS spot analysis of individual grains of samples prepared at 1100 "C revealed the occurrence of both perovskite and K-type grains, but the observed composition of the great majority of the grains agreed well with the weighed-in perovskite composition. This indi- cates that almost monophasic samples of the composition Lao$r,~,Col -z,Cu,Ru,03 (0dy <0.20) were formed. The amount of K-type phase tended to increase with increasing y value. In these perovskite- and K-type oxides, the oxidation states of the B cations are basically trivalent and divalent, respectively. As the normal valence state of Cu in these types of oxides is 2+, an increasing amount of IS-type phase with increasing overall content of Cu is expected. The structural considerations outlined below are based on samples calcined at 1100 "C, which yielded sharp and well resolved diffraction peaks; this in turn allowed accurate determination of the lattice parameters.The samples calcined at 900°C were those used for oxygen desorption and catalytic measurements, because calcination at higher temperatures reduces the specific surface area. The XRD patterns of La,.8Sro.,Col -zyCu,Ru,O, (0<y <0.20) were satisfactorily indexed on the basis of rhombohedral (hexagonal) unit cells similar to the ones reported for LaCo03" and Lao$ro.lCo031' (see also Table 1).Attempts were made to synthesize perovskite-based oxides with this type of hexagonal unit cell for y=0.3, 0.4 and 0.5, but their XRD patterns could not be indexed on this basis. The main phases found Ft y=O.3 and 0.4 could be indextd with cubic [a: 3.9144( 6) A] and vrthorhombic [a=5.588(1)A, b =5.5995(9) A and c =7.862( 1) A] unit cells, respectively, and the product with y=O.5 yielded a complex diffractrogram. It is thus concluded that it is possible to replace Co with Cu/Ru up to a y value of 0.2 in the rhombohedral unit cell of Lao.,Sro.,Co,- ,,CuyRu,O,. The lattice parameters of Lao.,Sr,,,Col -z,Cu,Ru,03 98 J. Muter. Chem., 1996, 6(l), 97-102 I I I I I I 20 25 30 35 40 45 28(Cu-KaJdegrees Fig.1 XRD patterns of Lao~,Sro.zCoo,,Cuo,lRuo,103calcined at (a) 900°C for 48 h, (b) 1000°C for 24 h, and (c) 1100°C for 24 h. The indices are based on the hexagonal unit cell; 0,Si (internal standard) and V,K,NiF,-type oxide. Table 1 XRD data for Lao.8Sro.2Coo,8Cyo.1Ruo.103synthesized at 1100"C [aH =5.4755(4)A, CH =13.217( 1) A] 1 0 2 3.8501 3.8527 12 1 1 0 2.7374 2.7377 100 1 0 4 2.7112 2.71 10 95 1 1 3 2.3251 2.23 17 1 2 0 2 2.2309 2.2028 20 0 0 6 2.2025 2.2028 6 2 0 4 1.9265 1.9263 66 2 1 2 1.7289 1.7298 3 1 1 6 1.7159 1.7163 3 3 0 0 1.5805 1.5806 17 2 1 4 1.5757 1.5754 42 1 0 8 1.5602 1.5602 15 2 2 0 1.3691 1.3689 15 2 0 8 1.3558 1.3555 11 3 1 4 1.2222 1.2219 10 2 1 8 1.2151 1.2149 11 (0<y G0.20) are given in Table 2. As seen in Fig.2, the hexagonal unit cell axes and volume increase monotonically with increasing y value. The formal oxidation states of the Co, Cu and Ru ions in the present system are not known (see also below) but the normal oxidation states of these ions in similar perovskite- based structures are three, two and four or three, two and five, Table 2 Lattice parameters of La,,,,Sr,,,,Co, -,,Cu,Ru,03 sythesized at 1100"C lattice parameters' sample y value aH/A cH /A vH/A3 a,/degrees CCRO 0 5.4457(2) 13.161 5( 8) 338.54 5.403 60.53 CCR5 0.05 5.4595(4) 13.192( 2) 340.54 5.410 60.60 CCRlO 0.10 5.47554 4) 13.217( 1) 343.19 5.423 60.65 CCR15 0.15 5.4911(6) 13.247( 2) 345.94 5.436 60.67 CCR20 0.20 5.5 106 (6) 13.281( 1) 349.31 5.452 60.72 aH,cH and VH are the lattice parameters and cell volume of the hexagonal unit cell and a, and Q, are the rhombohedral cell edge and angle, respectively. 13.161.~ I I I I~ 5.52 5.50 5.48 5.46 5.44 0.00 0.05 0.10 0.15 0.20 Y in La0.8Sr0.2c01-2$UfiUp3 Fig.2 Hexagonal lattice parameters as functions of y in ~a,~,~r,,,~o,~,,~u,~u,~,.(a) a axis, (b) c axis, (c) unit-cell volume. respectively. The most stable oxidation numbers of the Co, Cu and Ru ions are, however, three, two and four, respectively and octahedrally coordinate? Co3+, Cu2+ and Ru4+ ions have radii of 0.61, 0.73 and 0.62 A, respectively.'2 This implies that the increase in the lattice parameters can be ascribed to the fact that the average ionic radius of the substituent pair, Cu/Ru, is larger than that of the host Co ion as long as both Cu and Ru ions are incorporated in the perovskite framework (see also below).The rhombohedral distortion, i.e. the deviation from 60" of the rhombohedral angle, aR,increases with increasing y value (see Fig. 3).The rhombohedral distortion parameter, Y (in %), is defined as: y= loo{ 1-(~H/UH,/~)} (1) where uH and cHare the lattice parameters of the hexagonal unit cell. The Y value is zero for perovskites with cubic symmetry but is >0 for distorted perovskite framework struc- tures.As shown in Fig. 3, the Y values display the same tendency to increase with increasing y value as the aR values do. The tolerance factor, t, is commonly used as a measure of the distortion or deformation of the perovskite structure and is defined as: t= {rav(A)+r(O)}/,/2 x {rav(B)+r(O)I (2) where r(0) is the ionic radius of oxygen and rav(A) and rav(B) 60.8 1 60.7 g1 2 A 0Q,zv b F 1 60.6 1.o 60.5 0.00 0.05 0.10 0.15 0.20 Y in La0.8Sr0.2Col-2fiufiuf13 Fig. 3 The rhombohedral angle, a,, and distortion parameter, Y, plotted as functions of y in Lao.8Sro.zCo, -,,Cu,Ru,O, (Y and aR are defined in the text) are the average ionic radii of the A-site and B-site cations, respectively. For the ideal cubic structure t=1, decreasing as the crystal lattice becomes more di~torted,'~ and as a rule the tolerance factor is 0.8dtd1.0. In the present oxide system, r(0) and rav(A) are constant, and rav(B) increases with increas- ing y value, implying that the tvalue decreases simultaneously. In this case it is difficult to obtain realistic tvalues, however, as the true oxidation states of the B ions are not known.The variation of the cell parameters and degree of lattice distortion with composition can be satisfactorily explained by the replace- ment of smaller host Co ions by larger Cu/Ru pairs. Oxygen desorption behaviour Oxygen desorption from Co-based perovskites samples has been investigated e~tensively,'~-~~ and it has been revealed that the oxygen desorption process yields information about the valence states of metal cations incorporated in the perov- skite structure and about the catalytic properties of the com- pounds.The TPD chromatograms for oxygen release from the samples Lao,8Sro,,Co, -zyCuyRuy03( y =0, 0.05 and 0.1) are given in Fig. 4, where the recorder response, which is pro- portional to the rate of oxygen desorption, is plotted us. the temperature. The oxygen desorption from CCRO started at around 150"C and after passing a maximum at around 300 "C the desorption rate attained a constant value up to 800°C. Above 800°C the rate of oxygen desorption increased again. This oxygen desorption behaviour of CCRO is very similar to that reported for Lao.8Sro.2Co03 in ref. 18. The oxygen released below 800 "C emanates from weakly bonded lattice oxygens, and this release is accompanied with a reduction by Co4+ to Co3+ (see below) and is closely related to the catalytic proper- ties of these The onset temperature for the oxygen desorption increased and the amount of oxygen released decreased with increasing y value, as seen in Fig.4. By the partial substitution of Cu/Ru for Co, the amount of weakly bonded oxygens is decreased, as is the redox capacity of the oxide lattice. J. Muter. Chern., 1996, 6(l), 97-102 99 'i p, 0.8 0 200 400 600 800 T 1°C Fig.4 TPD chromatograms for oxygen release from La,,8Sr,~,Col~2,Cu,Ru,03.(a) y=O (CCRO), (b) y=O.O5 (CCR5) and (c) y=O.lO (CCR10). When Sr2+ ions substitute for La3+ ions in LaCoO,, the charge compensation is achieved either by the forma- tion of oxide ion vacancies, according to the formula L~,-,S~,COO,-~,,~,or by the formation of Co4+ ions as described by the formula La,-,Sr,(Co, -,)"(CO,)~+~, (or more likely electron holes), or a combination of these two possibilities according to the formula La, -xSrx- ~~~1-,~3+~~~,~4+~3-~~x,2~+~2,2~11.It has been shown that when these types of oxides are prepared in air, oxygen vacancies and nominal Co4+ ions are simultaneously formed, and that oxygen released below and above 800°C corresponds to the reduction of Co4+ to Co3+ and of Co3+ to Co2+, respectively.18 The decrease in the amount of oxygen released below 800 "C with increasing Cu/Ru content (see Fig.4) can thus be interpreted in terms of the Cu/Ru substitution restraining the formation of Co4+ ions. Cu occurs as divalent or, in a formal sense, trivalent ions in perovskite-related structures, and it has been shown that the oxygen desorption which starts around 200 "C for both Lao~6Sr,~4Coo.8Cuo.20321the high-T, superconductor and YB~,CU,O~-,~~and is centred at 300 and 550 "C, respectively, is due to the reduction of Cu3+ to Cu2+. Thus the reduction of oxygen desorption capacity with increasing Cu/Ru substi- tution in our Lao.8Sro.2C01 2,Cu,Ruy0, samples, especially -that which occurs below 400"C, suggests that the Cu ions in our samples are predominatively divalent. It has been reported that the Ru ions are tetravalent in SrRuO,, C~RUO,,~and La2MRU06 (M=Mg, Co, Ni, Zn),25 while Ru in BaLaMRuO, (M=Mg, Co, Ni, Zn),25 S~L~CURUO,,~~~~~Ba2( LaRuo.5Sbo.5)06 and Ba2( TaRuo.,- Na0.5)0628 is pentavalent.Asmentioned above, the Cu ions seem to be divalent in La,.,Sr,.,Co, -,,Cu,Ru,03 and thus the substitution of Cu/Ru for Co gives rise to a reduction of the concentration of Co4+ ions in the present system. This can be explained by the occurrence of either tetravalent or pentavalent Ru ions, if the variation of the concentration of oxide ion vacancies is taken into account. In order to confirm the exact valence state of each metal cation as well as to give a more complete explanation of the oxygen desorption behaviour, further studies are necessary. Catalytic activity In the oxidation of CO over Lao~8Sro~2Col-2,CuyRu,03with y=O, 0.05 and 0.1, C02 was the sole reaction product, and the temperature dependence of the conversion of CO to C02 is given in Fig.5. Among three oxides tested, CCRO catalysed the reaction at the lowest temperatures and is accordingly regarded as the most active one, followed by CCRS and CCR10. The degree of conversion of CO over CCRO and CCRS 100 h 80 g0 .-E 60 8 40 0-g 20 8 0 0 200 400 600 T 1°C Fig. 5Catalytic oxidation of CO over La,,,Sr,,,Co, ~,,Cu,Ru,O,; 0, CCRO (y=O); 0,CCRS (y=0.05); A, CCRlO (y=O.lO) increased sharply with increasing temperature and reached 100% conversion within very narrow temperature ranges. Over CCR 10 the increase of CO conversion with increasing tempera- ture was lower and tended to level off at higher conversions.These results show that the CO oxidation activity of La,.,Sr,~,Co, -2,Cu,Ru,0, decreases progressively with the substitution of Cu/Ru for Co. In the CO-NO reaction, formation of N20, N2 and CO, was observed and is assumed to proceed according to eqn. (3) and (4). 2N0 +CO+N20 +CO, (3) 2N0 +2CO-+N2+2C02 (4) As shown in Fig. 6, N20 is predominantely formed below 350 "C and N2 above this temperature. The formation of N2 is preferable to that of N20, as the latter is a greenhouse gas, i.e. the selectivity to N2 formation should be as high as possible. In this regard, the substitution of Cu/Ru for Co played a positive role because the conversion to N,O decreased with increasing y value.The reason why no formation of N20 was observed over CCRlO seems to be related to the fact that this compound did not show any activity for NO conversion below 400 "C, where N20is formed in larger amounts than N2.From a comparision of the conversion curves of CCRO and CCR5, however, it is obvious that the conversion to N,O decreases with increasing y value. AsCO, is a common product of both 100 h85z" 80 -0 (d 60 0 c .-c e 40 .c0 c .-0 t! 20 a>c 8 0 0 200 400 600 800 T1°C Fig. 6 Temperature dependence of the degree of conversion of NO to N, and N20 in CO-NO reaction over La,,,Sr,,,Co, ~2,Cu,Ru,03; 0 and 0,CCRO (y=O); and H,CCRS (y=0.05); A and A,CCRlO (y =0.10).Open and filled symbols represent conversions to N, and N,O, respectively. 100 J. Muter. Chem., 1996, 6(l), 97-102 Table 3 Lattice parameters and composition of the prepared K,NiF,-type oxides of overall composition (Lao.8Sro.2)z(Coo.5Cuo,5-ZRuz)04 ~ ~ ~~ z value alA c/A 0.05 0.10 3.81 16(3) 3.8142( 1) 12.864( 2) 12.838( 1) reactions, the degree of conversion of CO in the CO-NO reaction is plotted us. temperature in Fig. 7. As seen in this figure, CCRlO was far less active than CCRO and CCR5, which in turn showed comparable activities although CCRO was more active than CCR5 in the lower temperature region, i.e. where the degree of conversion was less than 50%. These results indicates that the activity of CO oxidation in the CO-NO reactions decreased with the substitution of Cu/Ru for Co, though the substitution had a positive effect on the selectivity for N, formation.Over Co-based perovskite-type oxides, oxidation reactions of CO and hydrocarbons have been classified as being intra- facial and involve redox-type reactions in which lattice oxygens, especially weakly bonded species, takes part.16,19920 The CO-NO reaction has also been reported to proceed with an intrafacial mechanism., Accordingly, it is reasonably concluded that the decrease in the amount of weakly bonded lattice oxygens with the increasing extent of substitution of Cu/Ru for Co is primarily responsible for the decrease in the catalytic activity for both CO oxidation and CO-NO reactions.Synthesis of K,NiF,-type oxides As described above, a K2NiF,-type phase was formed in minor amounts in connection with the preparation of the perovskite- type oxides in the La-Sr-Co-Cu-Ru-0 system. The TEM-EDS studies of crystal fragments present in the sample with the overall composition Lao.8Sro.2Coo.6CUo.2RU0.203 revealed the presence of grains with an electron diffraction pattern different from that of the perovskite phase. These grains had a La :Sr :Co :Cu :Ru ratio of 52 :15 :17: 15: 1 yield- ing a (La +Sr):(Co +Cu +Ru) ratio close to 2. Samples of the composition (Lao.8Sro,,)2(Coo.sCuo,s-ZRu,)04 with z =0.05 and 0.10 were accordingly prepared. The XRD patterns of these two compositions revealed the presence of a predomi- nantly tetragonal K-type phase, minor amounts of a perovskite phase and some unreacted oxides.The K-type phase in the samples with z=O.O5 and 0.10 had quite similar average metal composition and lattice parameters, as seen in Table 3. If the formula (Lao.8Sro.2),(CoCu)l -ZRuz)04 is used to express the solubility limit of Ru in this phase one obtains a z value of 0.03. This low solubility value can be understood in terms 200 400 600 800 T1°C Fig. 7 Temperature dependence of the degree of conversion of CO to C02 in CO-NO reaction over Lao,8Sro,2Co, -,,Cu,Ru,03; 0, CCRO (y=O); 0,CCR5 (y=0.05); A, CCRlO (y=O.lO). VIA composition suggested by EDS 186.91 (La0.82Sr0. 18)(c046cu0.5 IRU0.03 )O4 186.79 (La0.83Sr0.17 )(C0~6CU0.~1RU0.0~)04 of the B-site in La-based K-type oxides being normally occu- pied by divalent ions, implying that the amount of tetravalent Ru ions that can enter this position should be small. The TEM-EDS studies of the grains with the perovskite structure showed that they had an average metal composition of La :Sr :Co :Cu :Ru =37.9 :12.6 : 19.0 :7.8 :22.7, which yields a composition of La~~~~Sr~~~~CO~~~~CU~~~~~U~~~~~~.This is in accordance with the observation that the La-based perovskite framework is more capable of incorporating tetravalent cations than the corresponding K-type framework. Conclusions Monophasic samples of the composition Lao~,Sro,,Col -2y-Cu,Ru,03 with 0 <y <0.2 possessing the rhombohedral (hexagonal) perovskite structure have been prepared.The lattice parameters and rhombohedral distortion increased monotonically with increasing y value, due to the replace- ment of smaller Co ions by larger Cu/Ru pairs. This re-placement brought about a decrease in the amount of weakly bonded lattice oxygens, which are known to be the most active oxygen species in connection with intrafacial CO oxidation and CO-NO reactions. Accordingly the catalytic activity for CO oxidation and CO-NO reactions was decreased by the replacement of Co ions by Cu/Ru pairs. The Cu/Ru substi- tution did, however, selectively increase the N, formation and decreased that of N20 in the CO-NO reactions. The studies of the (Lao.8Sro~2)2(Coo~sCuo~5-.Ru,)O, system showed that the solubility of Ru ions in this K-type framework was very low.This work was supported by the Swedish National Research Council, the Swedish Board for Industrial and Technical Development and by The Ministry of Education, Science and Culture of Japan. References 1 F. S. Galasso, Structure, Properties and Preparation of Perovskite-type Compounds, Pergamon Press, Oxford, 1969. 2 R. J. H. Voorhoeve, in Advanced Materials in Catalysis, ed. J. J. Burton and R. L. Garten, Academic Press, New York, 1977, p. 129. 3 L. G. Tejuca, J. L. G. Fierro and J. M. D. Tascon, Adv. Catal., 1989, 36, 237. 4 Perovskites, ed. M. Misono and E. A. Lombard0 (Special issue of Catalysis Today, vol. 8, no. 2), Elsevier, Amsterdam, 1990.5 Perouskite Oxides, ed. L. G. Tejuca and J. L. G. Fierro, (Special issue of Catalysis Reviews, vol. 34, no. 4), Marcel Dekker, New York, 1992. 6 H. M. Zhang, Y. Shimizu, Y. Teraoka, N. Miura and N. Yamazoe, J. Catal., 1990, 121,432. 7 D. Klvana, J. Vaillancourt, J. Kirchnerova and J. Chaouki, Appl. Catal. A, 1994,109,181. 8 M. Skoglundh, L. Lowendahl, K. Jansson, L. Dahl and M. Nygren, Appl. Catal. B, 1994,3,259. 9 K-E. Johansson, T. Palm and P-E. Werner, J. Phys. E, 1980, 13, 1289. 10 JCPDS 25-1060. 11 JCPDS 28-1229, 36-1392: H. Obayashi, T. Kudo and T. Gejo, Jpn. J. Appl. Phys., 1974, 13, 1. 12 R. D. Shannon and C. T. Prewitt, Acta Crystallogr., Sect. B, 1969, 25, 925. 13 N. Ramadass, Mater. Sci. Eng., 1978,36,231. 14 N.Yamazoe, Y. Teraoka and T. Seiyama, Chem. Lett., 1981,1767. 15 N. Yamazoe, S. Furukawa, Y. Teraoka and T. Seiyama, Chem. Lett., 1982,2019. J. Mater. Chem., 1996, 6(l), 97-102 101 16 17 T. Nakamura, M. Misono and Y. Yoneda, Bull. Chem. Soc. Jpn., 1982,55,39 T. Nakamura, M. Misono and Y. Yoneda, J. Catal., 1983,83,151. 23 24 N. Miura, H. Suzuta, Y. Teraoka and N. Yamazoe, Jpn. J. Appl. Phys., 1988, 27, L337. A. Callaghan, C. W. Moeller and R. Ward, Inorg. Chem., 1966, 18 Y. Teraoka, M. Yoshimatsu, N. Yamazoe and T. Seiyama, Chem. Lett., 1984, 893. 25 5, 1572. I. Fernandez, R. Greatrex and N. N. Greenwood, J. Solid State 19 20 21 Y. Teraoka, S. Furukawa, N. Yaamazoe and T. Seiyama, Nippon Kagaku Kaishi, 1985, 1529. H-M. Zhang, Y. Teraoka and N. Yamazoe, J. Su$. Sci. Soc. Jpn. (Hyomen Kagaku), 1987,8,23. H-M. Zhang, N. Yamazoe and Y. Teraoka, J. Muter. Sci. Lett., 1989, 8, 995. 26 27 28 Chem., 1980,32,97. M. P. Attfield, P. D. Battle, S. K. Bollen, S. H. Kim, A. V. Powell and M. Workman, J. Solid State Chem., 1992,96, 344. S. H. Kim and P. D. Battle, J. Magn. Magn. Muter., 1993, 123,273. S. A. Almaer, P. D. Battle, P. Lightfoot, R. S. Mellen and A. V. Powell, J. Solid State Chem., 1993, 102, 375. 22 H-M. Zhang, Y. Shimizu, Y. Teraoka, N. Miura and N. Yamazoe, J. Catal., 1990,121,432. Paper 5/05365J; Receiued 10th August, 1995. 102 J. Muter. Chem., 1996, 6(l), 97-102

ISSN:0959-9428    

DOI:10.1039/JM9960600097

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Incorporation of poly (acrylic acid), poly (vinylsulfonate) and poly (styrenesulfonate) within layered double hydroxides Christopher 0.Oriakhi, Isaac V. Farr and Michael M. Lerner* Department of Chemistry and Center for Advanced Materials Research, Oregon State University, Corvallis, Oregon 97331, USA Poly(acry1ic acid), poly(vinylsu1fonate) and poly(styrenesu1fonate) have been incorporated between the positively charged sheets of layered double hydroxides (LDHs) M1 -xA1,(OH)2f (M =Mg, Ca, Co) and Zn,-,M',(OH),+ (M'= Al, Cr) to form layered nanocomposites. The resulting nanocomposites contained the LDH sheet structures separated by 7.6-16.0 A,which is sufficient to accommodate polymer bilayers between the LDH sheets. Preparations were carried out in deaerated aqueous base by a template reaction, involving the formation and precipitation of nanocomposites from metal nitrate-salt precursors in the presence of the dissolved polymer.Structural and compositional details were provided by X-ray diffraction (XRD), FTIR spectroscopy, elemental analysis, differential scanning calorimetry (DSC) and thermogravimetry (TG). Scanning electron microscopy (SEM) indicates that the nanocomposition of LDHs with ionomers significantly alters the particle microstructure from that of the LDH carbonates derived from aqueous precipitation. The layered double hydroxides (LDHs), otherwise called 'anionic clays', and their intercalation compounds have received considerable attention in recent years in view of their potential technological importance as catalysts, ion exchangers, optical hosts, ceramic precursors and antacids.'-'' The LDH structure consists of brucite-like M(OH), sheets, where partial substitution of trivalent for divalent cations results in a positive sheet charge compensated by anions within interlayer galleries. LDHs are represented by the general formula [M111-xM111'x(OH)2]Xt*[(An"-),,,,, -nH,O] with representative examples M =Mg, Zn or Ca; M' =Al, Cr or Fe; An"-=C032-, C1-, OH-, NO3-or SO4'-, and x taking values between 0.16 and 0.45.'' Thus, the mineral hydrotalcite can be abbreviated as Mg-A1-C03-LDH with x=O.25, and has the approximate formula unit M&jA12(OH)16CO3 4H2O.The two most important methods of incorporation (or nanocomposition) of polymers within the smectite group of clay minerals and other layered inorganic hosts, such as metal dichalcogenides, are in situ polymerization, i.e.the intragallery polymerization of intercalated monomer,12-16 and exfoliation/ adsorption, i.e. the interaction of the separated sheets of an exfoliated host with a solution containing the desired poly- mer.17-20 In the resulting nanocomposites, the polymers are included as monolayers or bilayers between anionic sheets, and are either themselves positively charged or associated with other cations within the nanodimensional polymeric layers. Unlike the smectite clays and some layered inorganic solids, the LDHs do not exfoliate into single-sheet colloids in aqueous or other polar solutions, consequently, and there have been no reports of preformed polymers incorporated uia the exfoli- ationladsorption method.The incorporation of organic anions into LDHs has been accomplished by the following methods: (1) an ion-exchange reaction with an LDH containing monovalent anions such as itra rate;^'-^' (2) the reconstitution of calcined, amorphous LDH in the presence of an organic and (3) a 'template' synthesis where the LDH sheets are grown in the presence of by the slow diffusion associated with macromolecules. Attempts to produce nanocomposites by the second (reconstitution) method have thus far afforded only disordered or nearly amorphous prod~cts.~' Recently, Messersmith and Stupp rep~rted~'~~~ the prep- aration of well ordered layered nanocomposites derived from Ca-A1-LDH and poly(viny1 alcohol) by a template reaction similar to method (3) where the polymer is included within the host during the original sol-gel processing to form the LDH sheets.It appears likely that this method will prove to be generally applicable towards the introduction of preformed anionic polymers within LDHs. There are relatively few other reports describing the incorporation of polymers between LDH A poly(ani1ine)-LDH nanocomposite has been prepared by in situ oxidative polymerization of aniline in an organic-pillared hydr~talcite,~~and a similar method is described in the prep- aration of poly(acrilonitrile)-hydrotalcite.25 The in situ poly-merization routes employed thus far have some disadvantages, namely that the neutral monomers are incorporated into an LDH containing organic anions, limiting the polymer content of the prod~ct,~~,~~ and the factors controlling polymerization and the nature of a polymer formed within the interlayer space are not easily understood.In this study, we describe the synthesis and characterization of several new nanocomposites derived using the template method from Mg-Al-LDH, Ca-Al-LDH, Zn-Al-LDH, Zn-Cr- LDH, and Co-A1-LDH sheet structures and three vinylic polymers; poly(acry1ic acid) (PAA), poly(vinylsu1fonate) (PVS) and poly(styrenesu1fonate) (PSS). Experimental Materials The inorganic starting materials used for all preparations described in this work were analytical reagent grade and used without further purification.Poly(sodium styrene-6sulfonate) (M, =70 x lo3); poly(sodium acrylate) (M, =30 x lo3, 40% the desired organic anion with exclusion of carb~nate.~~?~' w/w in water); and poly(sodium vinylsulfonate) [M, = Since the ion-exchange reaction proceeds via a topotactic (40-60) x lo3, 25% w/w in water], were used as obtained from mechanism within the LDH layers, this method of incorpor- Aldrich Chemical Co. Distilled and deionized water was used ation of preformed anionic polymers will be kinetically limited in all preparations. J. Muter. Chem., 1996,6(l), 103-107 103 Preparation of LDHs M-A1-C0,-LDHs (M =Mg, Ca or Zn) were prepared following a standard aqueous precipitation and thermal crystallization method described by Reichle., In a typical synthesis, a solu- tion containing Mg(N03)2 *6H,O (2.88 g, 0.011 mol) and A1(N03), *9H20 (1.41 g, 0.0038 mol) in 70 ml water was added over 1 h to a vigorously stirred 100ml solution containing NaOH (3.00 g, 0.075 mol) and Na2C0, (2.34 g, 0.022 mol).The gel obtained was aged at 65 "C for 24 h and, on cooling, was filtered and washed with a copious amount of water. The resulting white solid was dried in air for 24 h. Mg-Al-CO,- LDH was prepared with starting Mg:A1 mole ratios of 2: 1 (x=0.33) and 3: 1 (x=0.25), and XRD powder data refined on a hexagonal cell provided lattice parameters of a =3.04 and 3.06 A, respectively. A correlation between the a parameter and x in the formula unit has been e~tablished,,~ and provides values of xxO.33 for the former and ~~0.25 for the latter.The formula units were therefore similar to the starting metal ratios, and were taken to be Mg4A1,(OH),,CO3 *nH20 and Mg,A12(OH)16C03 nH,O. Both samples gave similar diffraction results (Table l), and the water content for Mg4A12(OH)12C03-nH20 is estimated at nz3 from TG data. The Ca-A1-C0,-LDH and Zn-A1-C0,-LDHs were prepared from the corresponding nitrate precursors in a similar manner. Preparation of polymer-LDH nanocomposites Polymer-containing LDHs were synthesized by reacting mixed aqueous salt solutions with a basic solution containing dis- solved polymer. Nanocomposites were prepared with the fol- lowing metal ratios: x =0.33 for MI -xA1,(OH),+ (M =Mg, Ca), and x=0.25 for Zn,-,M',(OH),+ (M'=Al, Cr) and Col-,A1,(OH),+.It was necessary to carry out the prep- arations under an N2 atmosphere (using Schlenk procedures) in order to exclude air, as the carbonate form of these LDHs form preferentially in the presence of CO,. A typical synthesis, of the PAA-Mg-Al-LDH nanocomposite, is described. The polymer (2.50 g) was dissolved in 100 ml of deaerated water in a three-necked flask and solutions of Mg(NO,), *6H,O (3.16 g, 0.012 mol), Al(NO3), -9H20 (2.34 g, 0.0062 mol) and NaOH (3.00 g, 0.075 mol) were added simultaneously under vigorous stirring. PAA is generally prepared under alkaline conditions, and is stable in aqueous base.35 The resulting precipitate was aged at 65 "C for 24 h and then filtered, washed several times with hot water to remove excess polymer, and dried under vacuum for 24 h.The nanocomposite is stable in air once formed. Other nanocomposites were prepared by a similar route from nitrate precursors. The Zn-Cr-PVS-LDH nanocomposite was prepared by dis- persing ZnO (3.03 g, 0.038 mol) in 100 ml of a deaerated aqueous solution containing 1.000 g of dissolved polymer. A solution of CrC1, (4.27 g, 0.019 mol in 50 ml water) was added slowly, and the precipitate was isolated and dried as described above. Characterization X-Ray powder diffraction (XRD) data were collected on a Siemens D5000 powder diffractometer, using Cu-Ka radiation, at 28 =0.02" s-'between 2 and 70". IR spectra were recorded on samples pressed into KBr disks using a Nicolet 510P FTIR spectrometer (resolution =2 ern- ', 100 scans averaged).A spectrum of pure KBr was collected for background correction. The morphology and microstructure of the samples were examined using an AMRAY lO00A scanning electron micro- scope. Thermal analyses of powdered samples (10-20 mg) were carried out using Shimadzu TGA-50 and DSC-50 instruments at 10°C min-' in flowing air or N, (50ml min-'). Carbon, hydrogen and nitrogen elemental analyses were carried out by Desert Analytics (Tuscon, AZ). Sulfur analyses were found to be unreliable due to the formation of incombustible sulfate salts. Results and Discussion Air must be carefully excluded (and solutions deaerated) prior to nanocomposite synthesis in order to avoid incorporation of carbonate ions into the product.The preferential accommo- dation of carbonate is readily explained as a result of the favourable lattice stabilization enthalpy associated with the small, highly charged C03,- anions, and is well known to cause difficulties in preparing LDHs with singly charged anions such as hydroxide and nitrate. Attempted syntheses of the polymer-LDH nanocomposites in air always resulted in the carbonate form with no evidence of polymer incorporation. Once prepared under N,, however, the polymer-containing nanocomposites are air stable. As a test, some nanocomposites were stirred in an aqueous solution of sodium carbonate for 1 week, with no evidence of exchange of polymeric anion for carbonate.The kinetic stability of the nanocomposites should derive from the slow diffusion of macromolecules within the galleries once formed. XRD patterns for Mg4A12(OH)12C03*nH,O and some nanocomposites are shown in Fig. 1 and 2, and the XRD data obtained are summarized in Table 1. The products described in Table 1 all exhibit only a single phase in the XRD patterns. The formation of nanocomposite products is demonstrated principally by the lack of peaks associated with the carbonate phase, and the appearance instead of a phase with an increased basal-plane repeat distance. The c repeat distance for Table 1 XRD data for LDH carbonates and nanocomposites including anionic polymers product M2' /M3' c repeat distance/A Ad/A" domain size/A Mg6A12(OH),&03 .nHzO 3 7.63 2.83 890 Mg4A12(OH),,C0, * nH20 2 7.63 2.83 940 Mg- A1-PSS-LDH 2 20.8 16.0 120 Mg- A1-PVS-LDH 2 13.1 8.3 120 Mg- A1-PAA-LDH 2 12.0 7.2 90 Ca-A1-C0,-LDH 2 7.62 2.82 - Ca- A1-PSS-LDH 2 19.6 14.8 160 Ca- A1-PVS-LDH 2 13.2 8.4 200 Ca- A1-PAA-LDH 2 12.4 7.6 290 Zn-A1-C03-LDH 3 7.65 2.85 3 70 Zn-Al-PSS-LDH 3 21.6 16.8 110 Zn-A1-PVS-LDH 3 13.3 8.5 100 Zn- A1-PAA-LDH 3 12.4 8.6 90 Zn- A1-PVS-LDH 3 13.0 8.2 130 CO- A1-PVS-LDH 3 13.3 8.5 130 a Ad is the gallery height taken as Ad =c repeat distance- brucite layer thickness (4.80 A).104 J. Muter. Chem., 1996, 6(l), 103-107 20.4 A Fig.1 Powder XRD patterns for Mg-Al-LDH containing (a) carbon-ate, (b) poly(acrylate), (c) poly(vinylsu1fonate) and (d) poly(styrene-sulfonate) ,7----I 4 6 8 io iz 14 16 is' '20 izT 28/degrees Fig. 2 Powder XRD patterns for (a) Mg-Al-PSS-LDH, (b) Ca-A1-PSS-LDH and (c) Zn-Al-PSS-LDH Mg,Al,(OH),,CO, *nH,O is 7.63 A,and the products in Fig. 1 have repeat distances of 12.0 (Mg-Al-PAA-LDH), 13.1 (Mg-Al-PVS-LDH) an! 20.8 A (Mg-Al-PSS-LDH). Allowing for a thickness of 4.8 A for the brucite-like LDH sheets, these distances correspond to galleries with polymer layer dimen-sions along the c axis of 7.2, 8.3 and 16.0A for the PAA-, PVS-and PSS-containing nanocomposite!, respectively. These can be compared with a distance of 2.8 A for the carbonate-containing phases.The dimensions obtained are consistent with those expected for the incorporation of bilayers of anionic polymers between LDH sheets. The c repeat distances for PAA and pOSS can be compared with those reportedofor acrylate (13.8 A),, and toluene-p-sulfonate (17.5-17.8 A)37q38 inter-calated into LDHs. Although details of the polymer confor-mation cannot be determined from the XRD data, a model for the arrangement of PVS within the intersheet galleries is shown in Fig. 3. The arrangement represented has anionic substituents oriented towards the LDH layers to maximize electrostatic LDH A+ + + + 13.1A + + + LDH Fig. 3 Schematic model for a bilayer packing of poly(vinylsu1fonate) in the LDH interlayer space attractions, and allows for the hydrophobic/hydrophilic bilayer separation favourable in many organic-pillared LDHs and other organoclays.This model may be generalized for the other polymer-LDH nanocomposites obtained. Reactions where the limiting reagent is PSS do not appear to form a related monolayer structure, instead mixtures of the bilayer phase and the LDH hydroxide are observed in XRD patterns of the products. The XRD pattern for Mg,A12(0H)12C03*nH,O exhibits a relatively sharp set of (001) reflections, indicative of a long-range ordering in the stacking dimension. The analogous peaks for the nanocomposites are broader, indicating a less organized stacking arrangement. The extent of peak broadening is similar 110 100 90 I T ' ' I " ' ' I " " 3000 2500 2000 1500 1000 50( wavenumberkm-' Fig.4 FTIR spectra for (a) M&A12(0H),,C03 *nH,O, (b) Mg-A1-PAA-LDH, (c) Mg-A1-PSS-LDH and (d) Mg-Al-PVS-LDH x (6) 3 100 10.1 5 0 E k --.o.oj0.080 .-0.1 4 60 -0.2 -0.3 40 Icci-0-3 J-0.4 0 500 1000 T/"C Fig. 5 TG-DrTG profiles for (a) Mg,Al,(OH),,CO, .nH,O and (b) Mg-A1-PSS-LDH J. Muter. Chem., 1996, 6(l), 103-107 105 with Mg, Ca and Zn-A1-LDH nanocomposites (Fig. 2). Domain sizes are obtained (Table 1) using the Scherrer relationship, d =0.9A/(D cos O), where d is the coherence length for crystal ordering, A is the X-ray wavelength, D is the peak width (in radians) at half height, and 8 is the diffraction angle.IR spectra of Mg4A12(OH),,C0, .nH,O and related nano- composites are provided in Fig.4. The strong peak at 1373 cm-l [Fig. 4(a)] is attributed to the v,(asym) stretching mode of the carbonate anion.39 The absence of both v3 peak splitting and a v1 peak near 1064cm-' also indicate that the C032- resides in a high symmetry (D3h)site. Peaks at 660 and 428 cm-l are associated with M-0 stretching modes in the LDH sheets. The nanocomposites [Fig. 4( b)-(d)] exhibit IR peaks characteristic of both the polymers and the LDH sheets. For example, the IR spectra of PVS-Mg-Al-LDH and PSS-Mg-Al-LDH both contain sharp peaks at 1040 cm-l and broader peaks at 1196 cm-' which are characteristic of S-0 vibrations in RSO, .40 The PAA-Mg-Al-LDH nanocomposite shows peaks characteristic of RC0,-at 1455 and 1566 cm-1.41 The absence or small size of carbonate-related peaks in the nanocomposite spectra provides evidence that the predominant anionic species incorporated between LDH sheets are the CHN elemental analyses and derived polymer contents for the Mg-Al-polyanion nanocomposites are provided in Table 2.The polymer content is established by assuming that carbon arises only from nanocomposited polymer, i.e. no residual carbonate content, which will tend to produce an overestimate of the polymer content. In all analyses, the N content was 0.03-0.06 mass(%), placing the maximum nitrate content at less than 1mol% of the anions incorporated. TG and derivative TG (DrTG) traces for Mg4A1,(OH),,C03 -nH,O and the Mg-Al-PSS-LDH nano-composite are shown in Fig.5. Mg4Al,(OH)12C03 *nH,O shows two mass loss events. The first, corresponding to 15% loss between 50 and 210 "C, has been attributed to elimination of both surface adsorbed water and interlayer water molecules. The second event, approximately 28% loss from 300-450 "C, derives from both C02 loss from the decomposition of carbon- ate and water loss by dehydroxylation of the brucite layers.' The total mass loss, 43.5%, provides a water content of n=2.8 based on complete conversion to the metal oxides, which agrees well with previous estimates of nw2-4 for air- dried samp1es.l In contrast, three events are observed in the thermal trace obtained on the Mg-A1-PSS-LDH nanocomposites.The first polymers. nancomposite Mg- A1-PSS-LDH Mg-Al-PVS-LDH Mg-A1-PAA-LDH Table 2 Elemental analyses and derived stoichiometries for the Mg-AI-LDH series mass% C H N polymer 24.63 4.64 0.03 47.0 7.43 4.24 0.06 33.1 14.12 4.61 0.05 27.9 empirical formula Mg2A1(OH),[CH2CH(C6H4S03)] * 1.6 H20 Mg,Al(OH), [CH2CHS03] * 2.2 H20 Mg,AI(OH),[CH,CHCO,] -0.4H20 Fig. 6 SEM images (3000x magnification) for (a) M&A12(OH),,C03 -nH,O, (b) Mg-Al-PAA-LDH, (c) Mg-A1-PVS-LDH and (d) Mg-Al-PSS-LDH 106 J. Muter. Chem., 1996, 6(l), 103-107 event, a 15% loss completed by ca. 100 "C, should correspond to the elimination of adsorbed water at the surface and between LDH layers. A second loss from 300 to 500°C (23%) is ascribed to the dehydroxylation of the LDH layers and partial decomposition of the polymer.After heating to 500"C, the sample colour changes from white to black, indicating that the polymer has begun to carbonize. The material obtained by heating to 500°C is also amorphous in the XRD analysis, which suggests a disruption of the brucite-like sheet structure due to intrasheet dehydroxylation. The final mass loss is observed at approx. 800 "C (22%) and is ascribed to complete oxidative elimination of the carbonaceous residue derived from the initial polymer degradation. The mixed oxide recovered after heating above 900 "C is white, confirming the low carbon content in the final product, and exhibits XRD reflections due to both MgA120, and MgO.Quantitative conversion of the starting material to metal oxides can be written as: where n is taken as ca. 4 to provide an H20 content near 15 mass% for the starting composition, and 'Mg3A103.5' reflects the overall stoichiometry of a mixture of spinel (MgAl,O,) and MgO. The calculated total mass loss of 68% is greater than that observed (6O%), which allows that a small quantity of sulfate is formed during thermolysis. A previous study on poly(viny1 al~ohol)-Ca-Al-LDH~~ pro-vides interesting observations on the morphology of products obtained by thermolysis and indicates an enhanced thermal stability associated with the nanocomposite over the parent phases. The morphology of catalysts derived from pyrolysis of LDH structures is also of considerable interest.' We have examined in some detail the thermal characteristics of the nanocomposites described above and evaluated the microstruc- tures of the thermolysis products, and will report these results separately.,, SEM images indicating the microstructures of Mg,Al,(OH),,C03 *nH20 and the three Mg-Al-LDH nano- composite powders are shown in Fig.6. The more crystalline carbonate phase [Fig. 6(u)] shows an aggregate of submicro- metre-sized particles that are estimated to be 0.1-0.3 pm in diameter under higher magnification. Some of these particles have aggregated into a larger platey mass, and a similar fine microstructure is evident on the surface of these aggregates. In contrast, all the nanocomposites [Fig.6(b)-(d)] show only large (>5 pm) platey aggregates with no observable structural details of submicrometre dimension. Conclusions We have described the synthesis and characterization of several new nanostructural materials based on Mg,A1,(OH)12C03 nHzO and related LDH sheet structures and the vinyl ionomers poly(acry1ic acid), poly(vinylsu1fonate) and poly(styrenesu1fonate). The products are obtained under air-free conditions by a template reaction where the LDH sheets are grown in a solution containing the desired polymer. The nanocomposi!es contain the LDH sheet structures separ- ated by 7.2-16.0 A, which is consistent with polymer bilayers between the sheets. These results broaden the application of the template method, indicating that the method is generally suitable for the preparation of nanocomposites between LDH and anionic polymers.The authors gratefully acknowledge supporting grant DMR- 9322071 from the National Science Foundation. References 1 F. Cavani, F. Trifiro and A. Vaccari, Catal. Today, 1992, 11, 173 and references therein. 2 W. T. Reichle, Chemtech, 1986,16, 58. 3 W. T. Reichle, J. Catal., 1985,94, 547. 4 E. Suzuki, M. Okamoto and Y. Ono, Chem. Lett., 1989,1487. 5 E. Suzuki and Y. Ono, Bull. Chem. SOC. Jpn., 1988,61,1008. 6 J. Twu and P. Dutta, J. Phys. Chem., 1989,93,7863. 7 S. Miyata, Clays Clay Miner., 1980,28, 50. 8 A. Corma, V. Fornes and F. Rey, J. Catal., 1994,148,205. 9 A. Corma, V. Fornes, F. Fernando, A. Cervilla, E.Llopis and A. Ribera, J. Catal., 1995,152,237. 10 V. R. L. Constantino and T. J. Pinnavaia, Inorg. Chem., 1995, 34, 883. 11 K. Allmann, Chimia, 1970,24,99. 12 J. Wu and M. Lerner, Chem. Muter., 1993,5,835. 13 J. P. Lemmon and M. M. Lerner, Solid State Commun., 1995, 94, 533. 14 J. P. Lemmon and M. M. Lerner, Chem. Muter., 1994,6,207. 15 P. Aranda and E. Ruiz-Hitzky, Chem. Muter., 1993,4, 1395. 16 M. G. Kanatzidis, R. Bissessur, D. C. DeGroot, J. L. Schindler and C. R. Kannewurf, Chem. Muter., 1993,5,595. 17 V. Mehrotra and E. P. Giannelis, Solid State Commun., 1991, 77, 155. 18 P. B. Messersmith and E. P. Giannelis, J. Polym. Sci. Part A: Polym. Chem., 1995,33,1047. 19 P. B. Messersmith and E. P. Giannelis, Chem. Muter., 1993,5,1064.20 K. J. Chao, T. C. Chang and S. Y. Ho, J. Muter. Chem., 1993,3,427. 21 H. Kopka, K. Beneke and G. Lagaly, J. Colloid Interface Sci., 1988, 123,427. 22 M. Meyn, K. Beneke and G.Lagaly, Inorg. Chem., 1990,29,5201. 23 K. A. Carrado, J. E. Forman, R. E. Botto and R. E. Winans, Chem. Muter., 1993,5,472. 24 P. K. Dutta and D. S. Robins, Langmuir, 1994, 10, 1851. 25 Y. Sugahara, N. Yokoyama, K. Kuroda and C. Kato, Ceram. Znt., 1988,14, 163. 26 K. Chibwe and W. Jones, J. Chem. SOC., Chem. Commun., 1989,926. 27 H. Tagaya, S. Sato, H. Morioka, J. Kodakawa, M. Karasu and K. Chiba, Chem. Muter., 1993,5, 1431. 28 I. Y. Park, K. Kuroda and C. Kato, J. Chem. SOC., Dalton Trans., 1990,3071. 29 L. Raki, D. G. Rancorut and C. Detellier, Chem. Muter., 1995, 7, 221. 30 C. 0.Oriakhi and M. M. Lerner, unpublished results. 31 P. B. Messersmith and S. I. Stupp, J. Muter. Res., 1992,7,2599. 32 P. B. Messersmith and S. I. Stupp, Chem. Muter., 1995,7,454. 33 T. Challier and R. T. C. Slade, J. Muter. Chem., 1994,4,367. 34 I. Pausch, H. Lohse, K. Schurmann and R. Allmann, Clays Clay Miner., 1986,34, 507. 35 For example, see V. Kabanov, D. Topchiev and T. Karaputadze, J. Polym. Sci., Symp. 42, 1973, 173. 36 M. Tanaka, I. Y. Park, K. Kuroda and C. Kato, Bull. Chem. SOC. Jpn., 1989,62, 3442. 37 E. D. Demotakis and T. J. Pinnavaia, Inorg. Chem., 1990,29,2393. 38 T. Kuwahara, 0. Onitsuka, H. Tagaya, J. Kadokawa and K. Chiba, J. Inclusion Phenom. Mol. Recognit. Chem., 1994, 18, 59. 39 M. Del Arco, C. Martin, I. Martin, V. Rives and R. Trujillano, Spectrochim. Acta, Part A, 1993,49, 1575. 40 D. H. Williams and I. Fleming, in Spectroscopic Methods in Organic Chemistry, 3rd edn., McGraw-Hill, London, 1980, p. 64. 41 W. R. Feairheller Jr. and J. E. Katon, Spectrochim. Acta, Part A, 1967,23,2225. 42 C. 0. Oriakhi, I. V. Farr and M. M. Lerner, J, Muter. Chem., submitted. Paper 5/04853B;Received 24th July, 1995 J. Muter. Chem., 1996, 6(l), 103-107 107

ISSN:0959-9428    

DOI:10.1039/JM9960600103

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Synthesis and ion-exchange properties of Na-4-mica Kevin R. Franklin* and Elizabeth Lee Unilever Research Port Sunlight Laboratory, Quarry Road East, Bebington, Wirral, UK L63 3JW The preparation of the high-charge-density sodium fluorophlogopite mica Na-4-mica has been investigated with respect to developing a simplified procedure, and controlling crystal size. Na-4-mica can be readily prepared using solid state synthesis procedures with reaction temperatures between 700 and 1100 "C. A simple 'dry mix and calcine' method (all-in-one method) has been found to yield essentially the same product as that obtained by the more complex multi-step procedure described previously in the literature. The level of potassium which can be exchanged into the Na-4-mica was found not to be critically dependent on the ion exchange capacity of the clay, as determined by alkylammonium ion exchange, but rather on the initial hydration state of the Na-4-mica.Studies on the ion exchange, adsorption and catalytic proper- ties of clay minerals have for many years been restricted to low-charge-density materials, such as the smectite and kaolinite types, and their synthetic analogues, e.g. laponite.lv2 These materials typically have ion exchange capacities of <100 milli- equivalents (meq) per 100g of dry clay. In recent years, however, synthetic clays with ion-exchange capacities in the order of 200 to 250 meq (100 g)-l have been prepared and are now the subject of considerable interest, particularly with respect to their catalytic and adsorption In many cases solid state synthesis procedures have been used to prepare these materials, and fluoride ions have been employed to aid minerali~ation.~-' The preparation in the last few years of Na-4-mica, a very highly charged sodium fluorophlogopite mica of chemical composition Na4Mg,A14Si4020F4 xH20 and theoretical ion-exchange capacity of 468 meq (100 g)-' of anhydrous clay, has extended the range of interesting clay minerals still further.1°-12 Large crystals (0.1-2 mm) of Na-4- mica were first prepared as one of a heterogeneous mixture of crystalline phases by Gregorkiewitz and Rausell-Colomlo by heating a mixture of ground augite, sodium fluoride and magnesium fluoride at 1080°C in an open furnace and then cooling slowly to 850 "C.Paulus et ul." later showed that pure phase materials of much smaller crystal size, could be prepared using an elaborate multi-step process. They further showed that the materials so produced had high ion-exchange selec- tivities for many divalent transition-metal ions and for stron- tium and barium, but not for other alkali-metal ions, magnesium and calcium." The exchange of strontium ions was accompanied by the collapse of the clays interlayer region leading to trapping of the strontium ions, a process considered highly desirable for the removal of radioactive "Sr from contaminated solutions." We now report the development of a much simplified process for the preparation of pure phase Na-4-mica, and show how the crystal size may be readily controlled.The effect of synthesis conditions on the hydration state of the Na-4-mica, and its subsequent effect on the ion exchange properties of the clay, are also discussed. Experimental All reagents used in this study were analytical grade and were purchased from Fluka, B. D. H. and J. T. Baker. Na-4-mica was prepared by a number of methods developed sequentially from that described previously by Paulus et uL1' Following the original procedure, 82.05 g magnesium nitrate hexahydrate, and 80 g aluminium nitrate nonahydrate were dissolved in 342 cm3 ethanol in a 500 cm3 glass bottle. After stirring for 1 h, 44.43 g tetraethylorthosilicate was added and the mixture stirred for a further 3 h. The bottle was then capped and heated at 60°C for 3 days to form a gel.The bottle was then uncapped and the gel heated at 100°C until dry. The dry gel was broken up, transferred to a silica crucible and then calcined at 475°C for 18 h. The resulting solid was ground in a ball mill until it passed through a 325 mesh sieve. 80g of the powder produced by the above method (about 2 batches worth) was mixed thoroughly with 80 g of sieved (325 mesh) sodium fluoride. The mixture was transferred to a 500 cm3 platinum dish and calcined for 18 h at a temperature between 700 and 1100°C. The resulting solid was ground up, dispersed in water and then thoroughly washed with saturated boric acid solution to remove fluoride impurities. The product was washed with water and either freeze dried or air dried at room temperature.The Na-4-mica was finally stored in a desiccator over saturated sodium chloride solution (~~0.75) for 2 days. To establish the importance of some of the steps in this rather long process, a series of modifications were carried out. (i) The same basic procedure was employed but without any sieving of the precursor gel or the sodium fluoride. (ii) The procedure was modified as (i), but the sol-gel method for preparing the precursor gel was replaced; the magnesium nitrate and aluminium nitrate were dissolved in a small amount of water and then 12.80g of fumed silica (CAB-0-SIL M-5) was stirred in to make a thick paste. This was dried and then ground up before calcining at 475°C.(iii) As (ii), but the fumed silica was dry-mixed with the magnesium nitrate and aluminium nitrate. (iv) The magnesium nitrate, aluminium nitrate, fumed silica and sodium fluoride were ground together for 5 min using a pestle and mortar, and then calcined directly at between 700 and 1100 "C. The ion exchange of sodium for potassium was investigated at room temperature and at 90°C under forcing conditions. 5 g of Na-4-mica were weighed into a 1 dm3 plastic bottle. 25.3 g of potassium nitrate dissolved in 250 cm3 water were then added and the dispersion thoroughly shaken. After 4 days the solution phase was filtered off and replaced by a fresh potassium nitrate solution of the same strength. After a further 3 days the mica was filtered off, washed with water and then freeze dried.The composition of the solid was subsequently determined after the mica had been contacted with water vapour as described above. Power X-ray diffraction (XRD) was carried out to determine the basal spacings of the Na-4-mica and to monitor crystallinity using a Siemens D5000 diffractometer fitted with variable J. Muter. Chem., 1996, 6(l), 109-115 109 divergence and anti-scatter slits. Crystal shape and size were determined with a Cambridge Instruments Stereoscan 360 SEM apparatus. Sodium, potassium, magnesium, aluminium and silicon contents were determined from X-ray fluorescence (XRF) analysis using a Philips XRFS PW1404 instrument. Fluoride was determined by ion-selective electrode analysis.Water contents were determined by thermal analysis using a Perkin-Elmer TGA7 thermogravimetric analyser. Maximal ion-exchange capacities were determined by contacting the Na-4-mica with a large excess of octadecylammonium chloride solution at 60 "C overnight. The resulting alkylammonium- exchanged mica was filtered off, washed with large quantities of ethanol and then freeze-dried. The amount of alkyl-ammonium present in the mica was obtained from CHN microanalysis. Results and Discussion Synthetic procedures Employing the procedure of Paulus et u2.l' and a crystallisation temperature of between 700 and 1100°C gave crystalline materials with XRD patterns which were consistent with natural mica-type structures13 and with the limited powder XRD data reported previously for Na-4-mica." Specific characteristic features were the strong reflections at 1.54, 2.63, I-* 2 I 3.25 and 4.20A.Peaks were also observed at about 9.8, 12.1 and 13.2A which correspond to the basal spacing of the anhydrous material, and the first and second interlayer hydrates, respectively." The relative peak intensities and the peak definition were, however, found to vary with the crystallis- ation temperature as illustrated in Fig. 1. This is probably a function of changes in crystal shape and size, and the crystal- linity of the materials. The change in crystal shape and size is clearly seen in Fig. 2; as the final calcination temperature was increased the crystals became larger, more 'book-like', and had more clearly distinguishable edges (hexagonal in shape).Repeating the experiments at 900 and 1000°C, but without sieving the calcined Mg-Al-Si precursor gel and the NaF before combining them, had no adverse effect on the Na-4- mica produced. When the silica source was changed from tetraethylorthosilicate to fumed silica and the ethanolic sol-gel process replaced by a procedure in which the silica was simply combined with a solution of magnesium nitrate and aluminium nitrate, the products again remained essentially unaltered. Furthermore, when the water was removed and the silica, aluminium nitrate and magnesium nitrate were simply dry- mixed and calcined at 475 "C to prepare the precursor gel, the final product was again found to be Na-4-mica.In a final permutation, the sodium fluoride was dry-mixed with the other solid reagents at the start and the mixture calcined directly at I I I I I I 4 I I 5 10 15 20 25 30 35 40 45 50 55 60 I I ---T+ 5 2e /degrees Fig. 1 XRD patterns for samples of Na-4-mica prepared by the method of Paulus et al." Samples prepared at (a) 800 and (b) 1000°C. &Spacings are given above the major peaks. 110 J. Muter. Chem., 1996, 6(1), 109-115 Fig. 2 Electron micrographs of Na-Cmica prepared by the method of Paulus et al." Samples prepared at (a) 700, (b) 800, (c) 900, (d) 1O00, and (e) 1100"C.Size bars =1pm. the final crystallisation temperature. Again, all the products obtained at between 700 and 1100°C were found to be Na-4-mica (Fig.3). Comparison of the XRD patterns for the products obtained by this 'all-in-one' procedure and by the method reported by Paulus et al." (Fig. 1 and 3) shows very good agreement for the higher temperature products. With the lower reaction temperatures agreement is less good; the 'all-in-one' materials exhibit more XRD peaks and in general show greater similarity with the patterns obtained for the higher temperature products. SEM images of the products from the 'all-in-one' method (Fig. 4) show the same effects of crystallisation temperature as observed with the procedure of Paulus et al." It is, however, also noticeable that crystals produced by the 'all-in-one' method show a wider size distribution. The effect of crystallisation time on the literature route and the 'all-in-one' route products was also investigated.With both methods, powder XRD showed that Na-4-mica was produced in all reactions carried out with crystallisation temperatures of 900 and lOOO"C, and crystallisation times of between 2 h and 5 days. SEM images showed that crystal size increased with time. At times >18 h, however, the images showed clear degradation of some of the crystals. This was particularly noticeable with the high temperature reactions (Fig. 5). Chemical composition and ion exchange capacity The molar chemical compositions, standardised with respect to the silicon content, of the Na-4-micas prepared by the literature route of Paulus et al." and by the 'all-in-one' method, are given in Table 1.The compositions are broadly in line with the proposed ideal composition for Na-4-mica (see introduc- tion), although the Mg: Si ratios were generally higher. The sodium contents were also consistently higher than expected with the highest temperature preparations. The hydration water contents of the Na-4-micas were all lower than the 0.5 water molecules per :odium reported previously by Paulus et ul." for the 12.2 A hydrated material, and the 2.7 water molecules per sodium suggested by Gregorkiewitz and Rausell- Colom'' for the same phase. This is, however, in line with the XRD patterns obtained for these materials,o which in many cases showed the presence of anhydrous 9.8 A phase material.The effective ion-exchange capacity (IEC) of a range of the Na-4-micas, as determined by ion exchange with octadecyl- ammonium ions, are given in Table 2. Clays have very high selectivities for such alkylammonium ions and their replace- ment of exchangeable sodium ions can generally be carried out quantitatively.' In all cases the 900 "C products were found to have the highest IECs, and to be close to the theoretical J. Muter. Chem., 1996,6(1), 109-115 111 .-0 5 10 15 20 25 30 35 40 45 50 55 60 I 5 10 15 20 25 30 35 40 45 50 55 60 Fig. 3 XRD patterns for samples of Na-4-mica prepared by the 'all-in-one' method. Samples prepared at (a)800 and (b) 1000"C. d-spacings are given above the major peaks. IEC based on the ideal chemical composition [468 meq (lOOg)-l], while those obtained for the 700 and 1100°C products were generally the lowest.While there are some quite large individual variations between materials prepared at the same temperature, in general all methods appear to give very similar products in terms of IEC. The measured IECs were found to vary with reaction time as well as reaction temperature as shown in Table 3. In general the IECs increased with reaction time up to about 18 h and then declined again at longer reaction times. This decline was most pronounced with the higher temperature reactions where crystal degradation was observed (see earlier). Loss of IEC could also have resulted from glass formation at the extended reaction times, however, no evidence was obtained to suggest that glass formation occurred.Characterisation of octadecylammonium exchanged Na-Cmica XRD patterns obtained for the octadecylammonium-exchanged fluoromica! showed the expansion of the mica basal spacing to about 48 A. This was the case irrespective of the method and temperature employed to prepare the Na-4-mica. Strong diffraction peaks corresponding to d/l to d/5 were visible (Fig. 6), indicating a high degree of ordering. From simple molecular-size calculations, it is probable that the alkylammonium ions are packed in bilayers between the clay plates, with the alkyl chains slightly tilted from the perpendicu- lar. Peaks corresponding to the basal spacing of the original Na-4-mica, whether due to anhydrous or hydrated interlayers, were absent in all cases from the alkylammonium clays, suggesting that the hydration state of the sodium clay had no effect on the ability of the Na-4-mica to undergo exchange.The interlayer expansion which accompanied octadecyl- ammonium-ion exchange of the fluoromica had a visible effect on the crystal shape; the crystals became far more block-shaped rather than plate-like. The layer striations also became more evident (Fig. 7). Effect of hydration state on potassium-ion exchange While alkylammonium-ion exchange in clays occurs readily, the exchange of sodium ions with other inorganic cations is often more difficult (especially with highly charged clays). Previous work suggests that Na-4-mica shows little selectivity for potassium ions.12 It was therefore thought that the Na-K system may provide a useful method of probing the effects of hydration and interlayer swelling on the ion-exchange reaction.The levels of potassium exchange achieved at room tempera- ture and at 90 "C for 7 days are given in Table 4. The first point of interest is the relatively low levels of potassium exchange in all cases, despite the use of highly forcing con- ditions. There also seems little correlation between the levels of potassium exchange and the IECs measured by alkyl- ammonium exchange (Table 2). Indeed, the material with the 112 J. Muter. Chem., 1996, 6(1), 109-115 Fig. 4 Electron micrographs of Na-4-mica prepared by the 'all-in-one' method.Samples prepared at (a) 700 and (b) 900 "C. Size bars = 1 pm. Fig. 5 Electron micrograph of Na-4-mica prepared by the 'all-in-one' method at 1000 "C for 5 days. Size bar = 1 pm. Table 1 Chemical composition of some Na-4-mica samples Preparative method", temperaturePC compositionb A, 700 Na4.1 Mg6.4 si4 A14.1 O20.20 F4.8 * 1.20H20 A, 800 Na3.8 Mg6.3 si4A14.0 020.40 F3.6 *0.58H20 A, 900 Na3.8 Mg6.3 si4 A14.0 020.25 F3.9*0-37HzO A, lo00 Na5.2 Mg6.6 si4 A14.3 O21.05 F5.2 1.28H20 A, 1100 Na5.0 Mg6.5 si4 A14.2 O20.90 F4.8 1.07HZ0 B, 700 Na5.4 Mg7.0 si4 A14.3 021.05 '6.2 1.40H20 B, 800 Na4.1 Mg6.9 si4A14.4 021.45 F4.2 * 1*88H20 B, 900 Na4.0 Mg7.2 si4 A14.6 O22.15 F3.9. 1-43H~O B, lo00 Na5.3 Mg7.z si4 Al4.7 022.65 F4.5 * 1.49H2O B, 1100 Na4.5 Mg6.7 si4 A14.3 021.95 F2.9 '1.40H20 a A =ref. 11; B =all-in-one.The oxygen value was not determined, but was rather inferred from electrical neutrality considerations. highest IEC (Paulus et al." at 900 "C)gave the lowest level of potassium exchange. In fact, it appears that the level of potassium exchange is controlled by the water content (hydration state) of the sodium form starting material. The effect was most clear with the 'Paulus route' (ref. 11) 800 and Table 2 Ion-exchange capacities" of Na-4-micas as determined by octadecylammonium ion exchange reaction literature method aqueous gel two-step dry mix all-in-one tempPC (ref. 11) methodb method' method 700 192 356 282 243 800 381 415 266 289 900 470 433 45 1 439 lo00 413 341 321 310 1100 213 275 262 318 a In meq (100 g)-' of dry sodium clay.Aqueous solution of Mg and A1 nitrates combined with fumed silica, calcined at 450 "C, mixed with NaF, and calcined at reaction temperature given in the table. 'As for aqueous gel method, but Mg and A1 nitrates dry-mixed with the silica. Table3 Effect of reaction time on the ion exchange capacities of Na-Cmica prepared by the all-in-one method. ion-exchange capacity reaction time/h at 900 "C at 1o00"C 2 323 264 6 404 178 18 439 310 48 302 222 300 303 146 900°C materials which had very low water contents an! showed low levels of potassium exchange. Major 9.8 A anhydrous sodium-phase basal-spacing peaks were present in the XRD patterns of the starting sodium forms and in the potassium-exchanged forms of these materials.Conversely, the 'all-in-one' route 800 "C material, which gave much greater potassium exchang?, had a much higher water content and a predominant 12.1 A hydrated-phase basal-spacing peck in its XRD pattern. Upon potassium exchange the 12.1 A phas? peak completely disappeared and was replaced by a 12.8 A peak, which is consistent with a hydrated potassium phase," and !n unassigned (possible anhydrous potassium phase) 10.4 A peak. It is concluded, therefore, that formation of the hydrated sodium-ion phase is a prerequisite for potassium-ion exchange to occur. The reduction in the total cation content (Na+K) of the exchanged materials relative to the starting material (Table 4) is most probably due to the dissolution of impurities during the ion exchange process.Since hydration of the interlayer region of Na-4-mica appears to be so important for effective potassium exchange, attempts were made to hydrate the Na-4-mica prior to exchange. Storing the materials in a humid atmosphere for 6 weeks had virtually no effect on the hydration state. Further contacting the mate- rials with water at room temperature and at 90°C for periods up to 1 month followed by air-drying, resulted in no change in the hydration state. Potassium exchange carried out sub-sequently likewise showed that the material treated in this way had no greater tendency to undergo exchange than the starting materials. Thus it appears that, contrary to previous once dehydrated, Na-4-mica is not readily reh ydrated.In all of the studies reported above, the materials used had been freeze-dried after synthesis. To investigate whether this extreme form of drying was the cause of irreversible dehy- dration, some materials were prepared as previously, but were air-dried. All materials so prepared, however, had very similar levels of hydration to those prepared using freeze-drying. Once again, the 'Paulus et al. (ref. 11) route at 900°C' material had a particularly low level of hydration. Likewise, potassium exchange of the air-dried materials occurred to essentially the same level as in the corresponding freeze-dried materials. J. Mater. Chem., 1996,6(l), 109-115 113 I I I I I I I I mvi nIL10 15 10 15 20 25 30 35 40 45 50 55 60 Fig 6 XRD patterns of octadecylammonium exchanged Na-4-mica. Na-4-mica prepared at 800°C using (a) the method of Paulus et al.” and (b)the ‘all-in-one’method.D-spacings are given above the major peaks. Fig 7 Electron micrograph of octadecylammonium-exchanged Na-4-mica. Size bar = 1 lm. The possibility that the difficulty in hydrating the Na-4-mica was due to the presence of an impurity of a lower-charge-density non-swelling clay similar to natural micas’ was also considered. To be responsible for the observed hydration effects, it would require a large amount of such a phase to be present. The presence of a large amount of a lower-charge-density phase is, however, inconsistent with the measured chemical compositions and IECs, and this possibility can therefore be discounted.Thus, it is still unclear why some 114 J. Muter. Chem., 1996, 6(1), 109-115 Table 4 Composition of potassium-exchanged micas room starting temperature exchanged 90 “C exchanged preparative method: material - product product temperature/”C Na:Si K:Si Na:Si K:Si Na:Si A, 700 1.03 0.19 0.69 0.55 0.28 A, 800 0.95 0.08 0.84 0.38 0.55 A, 900 0.95 0.07 0.88 0.13 0.78 A, 1000 1.30 0.35 0.76 0.53 0.33 A, 1100 1.25 0.46 0.67 0.63 0.30 B, 700 1.35 0.18 0.89 0.38 0.43 B, 800 1.03 0.33 0.58 0.50 0.28 B, 900 1.00 0.38 0.53 0.45 0.35 B, 1000 1.33 0.35 0.88 0.43 0.45 B, 1100 1.13 0.38 0.60 0.43 0.45 A =ref.11; B =all-in-one. samples of Na-4-mica appear to undergo irreversible dehydration. Conclusions Na-4-mica may be prepared readily using solid-state synthesis procedures with reaction temperatures between 700 and 1100"C.The multi-step procedure used to combine the reac- tants described previously by Paulus et a!." is, however, complex, and a much simpler 'dry mix and calcine' method ('all-in-one' method) has been shown to yield essentially the same product. The crystallisation temperature is very import- ant in controlling the crystallinity and crystal size of Na-4- fluoromica; both increase with temperature. Reaction time is also important, particularly with high temperature (21000 "C), as with long reactions (>18 h) crystal degradation can occur.The ion-exchange capacity (IEC), as measured by alkylam- monium-ion exchange, varies with reaction temperature and time; the highest values [close to the theoretical maximum, 468 meq (100g)-'] were obtained with 900 "C reactions. The level of potassium-ion exchange that could be achieved under forcing conditions was not, however, a function of the IEC, but was rather controlled by the hydration state of the interlayer region of the Na-4-mica. Ion exchange of sodium ions out of anhydrous layers proved very difficult. Despite earlier reports to the contrary, hydration of anhydrous phase Na-4-mica did not occur readily, and thus, avoiding the formation of this phase during the synthesis of the material is highly important if it is to be ion exchanged with inorganic cations.The authors would like to thank the following people for their contribution to the characterisation of the Na-4-micas; Ms. Sharon Evans and Mr. Ian Tucker (XRD), Mr. Jaz Lechy, Ms. Helen Bills and Ms. Jane Munro-Brown (SEM), and Ms. Di Savage (XRF, fluoride analysis). All are employees of Unilever Research Port Sunlight Laboratory. References 1 H. van Olphen, An Introduction to Clay Colloid Chemistry, 2nd edn., Wiley, New York, 1977. 2 B. K. G. Theng, The Chemistry of Clay Organic Reactions, Wiley, New York, 1974. 3 Y. Morikawa, T. Goto, Y. Mora-Oka and T. Ikawa, Chem. Lett., 1982,1667. 4 H. Sakurai, K. Urabe and Y. Izumi, J. Chem. SOC.,Chem. Commun., 1988,1519. 5 J. W. Johnson, J. F. Brody, R. M. Alexander, L. N. Yacullo and C. F. Klein, Chem. Muter., 1993,5, 36. 6 H. Tateyama, K. Tsunematsu, H. Hirosue, K. Kirura, T. Furusawa and Y. Ishida, Proc. 9th Znt. Clay Conf, Strasbourg, 1989, ed. V. C. Farmer and Y. Tardy, Sci. Giol. Mim., 1990,86,43. 7 H. Tateyama, S. Nishimura, K. Tsunematsu, K. Jinnai, Y. Adachi and M. Kimura, Clays Clay Miner., 1992,40, 180. 8 K. Kitajima, F. Koyama and N. Takusagawa, Bull. Chem. SOC. Jpn., 1985,58, 1325. 9 F. D. Duldulao and J. M. Burlitch, Chem. Muter., 1991,3, 772. 10 M. Gregorkiewitz and J. A. Rausell-Colom, Am. Mineral., 1987, 72, 515. 11 W. J. Paulus, S. Komarneni and R. Roy, Nature, 1992,357,571. 12 S. Komarneni, W. J. Paulus and R. Roy, in New Developments in Ion Exchange; Proc. Int. Con$ Ion Exchange, 1991,p. 51. 13 R. E. Grim, Clay Mineralogy, McGraw-Hill, New York, 1953, pp. 93-95. Paper 5103989D; Received 20th June, 1994 J. Muter. Chem., 1996, 6(1), 109-115 115

ISSN:0959-9428    

DOI:10.1039/JM9960600109

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

MATERIALS CHEMISTRY COMMUNICATIONS Electrochemical doping of a-ethyl-disubstituted oligothiophenes and electrical conductivities of the resulting radical-cation salts Naoki Noma, Kazuhiro Kawaguchi, Ichiro Imae, Hideyuki Nakano and Yasuhiko Shirota” Department of Applied Chemistry, Faculty of Engineering, Osaka University, Suita, Osaka 565, Japan Electrochemical doping of a-ethyl-disubstituted oligothiophenes and the electrical conductivity of the resulting radical-cation salts have been studied. 5,5”‘-Diethyl- 2,2’:5’,2”:5”,2’”-quaterthiophene(DEt4T) and 5,5””-diethyl- 2,2’:5””:5”,2”’:5”’,2’”’-quinquethiophene(DEt5T) are found to contrast with each other in the stoichiometry and electrical properties of the resulting solid radical-cation salts. The DEt4T and DEt5T radical-cation salts with a doping extent of ca. 100 and 50%, respectively, produced by electrochemical doping exhibited room-temperature conductivities of 9 x 10-8 and 5 x S cm-’ with activation energies of 0.54 and 0.11 eV, respectively.Oligothiophenes linked at the a-position of the thiophene ring have recently received attention not only as model compounds for electrically conducting polythiophenes but also as a new class of organic 7c-electron systems. Unsubstituted and subs tit uted oligo thiophenes have been synthesized,1-3 and their 0ptica1,~ electro~hemical,~-~ electrical8-” and photoelec- trical’’v’2 properties have been studied. With regard to the electrical conductivities of oligothio- phenes, both charge-transfer complexes and chemically doped materials have been studied.Charge-transfer complexes of the unsubstituted thiophene trimer or tetramer with TCNQ have been reported to exhibit room-temperature conductivities of 10-10-10-9 S cm-’ .8 Electrical conductivities of the unsubsti- tuted thiophene tetramer and pentamer doped with iodine have also been reported to be ca. 10’ S cm-’; however, unsub- stituted oligothiophenes undergo doping-induced coupling reaction to produce higher oligomers.’ Single crystals of the a-methyl-disubstituted thiophene tetramer doped with I,, NOBF, or NOPF6 have been reported to exhibit conductivities of 10-2-10-1 S cr”;’’ however, the stoichiometry of these complexes is not clear. Electrochemical doping is of interest in that definitely ident- ified radical-ion salts with dopants derived from support- ing electrolytes can be prepared.We report here the first ex- ample of electrochemical doping of oligothiophenes and the electrical conductivities of the resulting solid radical-cation salts. a-E th yl-dis ubs ti t uted oligo thiophenes, 5,5”’-die thyl- 2,2’:5’,2”:5”,2”’-quaterthiophene(DEt4T) and 5,5””-diethyl-2,2’:5’,2“:5”,2”’:5”’,2’”’-quinquethiophene(DEtST), where the coupling reaction of the resulting radical cations was expected to be prevented due to the presence of the two a-ethyl groups, were investigated and were found to contrast with each other in the second anodic oxidation process, and in the stoichi- ometry and electrical properties of the resulting solid radical- cation ~a1ts.l~ DEt4T DEt5T DEt4T and DEt5T were synthesized in diethyl ether by Grignard coupling reactions of 2-bromo-5-ethylthiophene with 5,5’-dibromo-2,2’-bithiophene, and 2-bromo-5-ethylthiophene with 5,5”-dibromo-2,2’:5’,2”-terthiophene,respectively.DEt4T: mp 183°C. Elemental analysis: Found: C, 61.99; H, 4.59; s, 33.22%. Calc. for C20H18S4: c, 62.13; H, 4.69; s, 33.18%. DEt5T: mp 246°C. Elemental analysis: Found: C, 61.26; H, 4.31; S, 34.19%. Calc. for C24H20S5: C, 61.50; H, 4.30; S, 34.20%. Fig. 1 shows cyclic voltammograms for the anodic oxidations of DEt4T and DEt5T in dichloromethane. Whereas the anodic oxidation processes of the unsubstituted thiophene tetramer and pentamer are irreversible, those of DEt4T and DEt5T are reversible owing to the presence of the ethyl group at the a-position of each terminal thiophene ring.The oxidation poten- tials of DEt4T and DEt5T were determined to be 0.58 and 0.52 V us. Ag/Ag+ (0.01 mol dmP3), respectively. DEt4T and DEt5T were found to contrast with each other in the 0 +0.5 +1.0 EN(vs. AS/Ag+, 0.01mol dm4) Fig. 1 Cyclic voltammograms of a, DEt4T (1.0x mol dm-j) and b, DEt5T (3.0 x lop4mol dm-3) in dichloromethane in the presence of Bu4NC104 (sweep rate: 100 mV s-’) J. Muter. Chem., 1996, 6(1), 117-118 117 was ca. 2: 1, as determined from elemental analysis. That is, the anodically oxidized species of DEt4T and DEt5T are regarded as a simple salt with a doping extent of ca.loo%, ,.:.*.a9 .*' 'b.,-' 400 500 600 700 wavelengthtnm Fig.2 Electronic absorption spectra of a, neutral DEt4T; b, DEt4T radical-cation salt; c, neutral DEt 5T; and d, DEt5T radical-cation salt in the solid state second anodic oxidation process. Whereas the second oxi- dation process of DEt4T was irreversible, that of DEt5T was reversible; a second anodic wave and a corresponding cathodic wave were observed at 0.82 (Epa)and 0.75 V (Epc).This result can be compared with the literature data for the second oxidation process of a thiophene tetramer capped at both the CI-and /?-positions of the terminal thiophene rings which was reversible, although the second oxidation process of a trimer was irreversible.6 Based on the information obtained by cyclic voltammetry, electrochemical doping of DEt4T and DEt5T was carried out by controlled-potential anodic oxidation at 0.65 and 0.55 V us.Ag/Ag+ (0.01 mol dm-3) for dichloromethane solutions of DEt4T and DEtST, respectively, for 24 h. In the case of DEt4T, the solution turned dark green when the electrolysis started, and black powders were deposited onto the surface of the working electrode. In the case of DEtST, however, there was no deposition onto the working electrode, but black powders were precipitated in the solution. The black powders of the anodically oxidized DEt4T and DEt5T were washed with fresh dichloromethane. They were identified as radical-cation salts with a perchlorate anion as a dopant, as characterized by various spectroscopies and elemen- tal analysis.The IR absorption spectra of the electrochemically oxidized DEt4T and DEt5T show strong bands at ca. 1100 cm-' due to the dopant perchlorate anion. As Fig. 2 shows, the electro- chemically oxidized DEt4T and DEt5T show new broad electronic absorption bands assignable to the DEt4T and DEt5T radical cations in the wavelength region from 650 to 750nm (A-ca. 700nm) and from 700 to 850nm (Amax ca. 790 nm), respectively. The presence of unpaired electrons in the solid radical- cation salts of DEt4T and DEt5T was also confirmed by electron paramagnetic resonance spectroscopy. Sharp signals with g values of 2.004 and 2.003, and peak-to-peak linewidths of 4.4 and 2.3 G were observed for electrochemically oxidized DEt4T and DEtST, respectively.A striking difference in the stoichiometry of the resulting radical-cation salts was observed for DEt4T and DEt5T. While the mole ratio of oligothiophene : dopant for the DEt4T radical- cation salt was ca. 1 : 1, that for the DEt5T radical-cation salt and a complex salt with a doping extent of ca. So%, respect- ively. The electronic absorption spectral data are also in support of this; i.e. whereas the n-n* absorption band in the wavelength region from 420 to 540nm, which is due to the neutral thiophene tetramer, becomes significantly smaller in intensity for the electrochemically oxidized DEt4T relative to the neutral DEt4T, the n-n* absorption band in the wavelength region from 420 to 580nm, which is due to the neutral thiophene pentamer, still remains in the case of the electro- chemically oxidized DEt5T.The difference in the stoichiometry between DEt4T and DEt5T is thought to be caused probably by the difference in the mode of the deposition of the oxidized species, because coulometry in a thin cell exhibited one-electron oxidation for both DEt4T and DEt5T. Electrical conductivity of radical-cation salts of DEt4T and DEt5T was measured for pellet samples by a two-probe dc method in the temperature range 20-70 "C.The radical-cation salts of DEt4T with a doping extent of ca. 100% and DEt5T with a doping extent of ca. 50% exhibited room-temperature conductivities of (7-9) x S cm-' and (2-5) x S cm-', with activation energies of 0.54 and 0.11 eV, respect- ively. The four orders of magnitude difference in electrical conductivity between DEt4T and DEt5T radical-cation salts is attributed to the difference in the extent of doping; i.e., the on-site coulombic repulsion in the complex salt of DEt5T is much lower than that in the simple salt of DEt4T.The present study presents the first example of electrochemi- cal doping of oligothiophenes, showing a striking difference in the electrical conductivity of the resulting radical-cation salts between DEt4T and DEt5T. It will be of interest to investigate further the correlation between molecular structure and electri- cal properties of electrochemically doped oligothiophenes with different n-conjugation lengths.This work was supported in part by Grants-in-Aid for scientific research, nos. 06226244 and 06453155, from the Ministry of Education, Science, and Culture of Japan. References 1 J. Kagan and S. K. Arora, J. Org. Chem., 1983,48,4317. 2 J. Nakayama, T. Konishi and M. Hoshino, Heterocycles, 1988, 27, 1731. 3 W. ten Hoeve and H. Wynberg, J. Am. Chem. SOC., 1991,113,5887. 4 D. Fichou, G. Horowitz, B. Xu and F. Garnier, Synth. Met., 1990, 39,243. 5 Z. Xu, D. Fichou, G. Horowitz and F. Garnier, J. Electroanal. Chem., 1989,267,339. 6 P. Bauerle, U. Segelbacher, A. Maier and M. Mehring, J. Am. Chem. SOC.,1993,115,10217. 7 J. Guay, P. Kasai, A. Diaz, R. Wu, J. M. Tour and L. H. Dao, Chem. Muter., 1992,4, 1097. 8 S. Hotta and K. Waragai, Synth. Met., 1989,32, 395. 9 E. E. Havinga, I. Rotte, E. W. Meijer, W. ten Hoeve and H. Wynberg, Synth. Met., 1991,4143,473. 10 S. Hotta and K. Waragai, J. Muter. Chem., 1991,1, 835. 11 Y. Kuwabara, K. Miyawaki, K. Nawa, N. Noma and Y. Shirota, Nippon Kagakukaishi, 1992,1168. 12 N. Noma, T. Tsuzuki and Y. Shirota, Ado. Muter., 1995,7,647. 13 K. Kawaguchi, I. Imae, H. Nakano, N. Noma and Y. Shirota, Polym. Prepr., Jpn., 1993,42,2842. Communication 5/05331E; Received 9th August, 1995 118 J. Muter. Chem., 1996, 6(1), 117-118

ISSN:0959-9428    

DOI:10.1039/JM9960600117

出版商:RSC    

年代:1996

数据来源: RSC