摘要:

XPS and XRD study of photoconductive CdS films obtained by a chemical bath deposition process Mitko Stoev" and Atanes Katerski Institute of General and Inorganic Chemistry, Bulgarian Academy of Sciences, 1113 Sofia, Bulgaria Cadmium sulfide layers have been obtained from CdCl,, CS(NH,),, NH,Cl, NH, and H,O solutions using the chemical bath deposition technique at 50 "C. After dipping in a solution of 0.7 mass% CdC1, in CH,OH and thermal treatment under Ar at 400 "C for 30 min, the CdS layers obtained acquire a photosensitivity amounting to a ratio, between photocurrent values in the light and in the dark, of four orders of magnitude. The chemical changes of the layers occurring, as evidenced by X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS), during their preparation in solution and after thermal treatment are discussed.The interest in CdS goes back to 1954 when the photovoltaic effect in CdS/Cu2S rectifiers was discovered.' Thin CdS films have been applied to the production of n-type window layers for heterojunction solar cells such as n-Cds/~-CdTe~,~ and n-CdS/p-CuInSe, .4 The cheapest and easiest preparation method of polycrystalline CdS films is via chemical bath deposition (CBD). The earliest investigations on the prep- aration of CdS thin films by chemical precipitation in solutions are associated with the use of cadmium salts, thiocarbamide and sodium hydr~xide.~.~ The optimum concentrations of the reaction mixtures and the conditions of chemical pFecipitation for the preparation of CdS films of up to 1OOOA thickness have been established. To enhance the CdS film thickness by a single dip, NH40H is used instead of NaOH.7 The kinetics and mechanism of CdS formation with the participation of ammonium thiocarbamide complexes' of cadmium has been elucidated and the properties'-'' of the resulting films have been determined.Since the CdS preparation process is difficult to control, a technique using the buffer effect of ammonium salts has been devel~ped.'~.~~ The application of a buffer solution technique14 to the preparation of CdS films by CBD has been discussed. It has been found that in the case of buffer solutions of ammonium salts the growth of CdS films is easier to control because the ammonium salt slows down the release of cadmium ions into the solution to form CdS ion-by-ion on the substrate.With the bath deposition method, the cadmium is in a complexed form. In addition to NH,OH, N( CH,CH,OH), (triethan~larnine)'~-'~has also been used as a complexing agent for the preparation of high-quality CdS thin films. When Na3C6H507 2H20 (sodium itr rate)'^'^^ was used, a citrate complex of cadmium was formed in solution, which yielded CdS thin films. Films of CdS have also been obtained using KCN as a complexing agent in which the [Cd(CN),I2- complex was formed in solution.21 The effect of CdCl, on the electronic properties of solar cells of the type CdS/CdTe has been inve~tigated.~~?~~ Treatment with a satu- rated solution of CdC1,-CH,OH leads to better properties of the CdS/CdTe interface, induced CdTe grain growth and significantly improved overall efficiency.XRD and XPS studies of the conversion of chemically deposited photosensitive CdS thin films to n-type materials by air annealing and ion- exchange reactions have been reported.,' In these cases, the conversion of the films to n-type materials was achieved by immersing the film in a 0.01 mol dm-, HgC1, solution for 15 min followed by air annealing for 1h at 200°C. The chemistry of formation of photosensitive CdS films obtained by the buffer solution technique and treatment with CdCl,-CH,OH solutions has not yet been elucidated, nor has any interpretation of the changes in contour of the XP spectra during the formation of photosensitive films been provided.The present study is a continuation of investigations on the preparation of CdS thin films by chemical method^.^^,^^ It is aimed at: (i) deposition of CdS films suitable for solar cells on glass substrates using the buffer solution technique; (ii) thermal treatment of CdS films in order to obtain photosensitive films, and (iii) investigation by XPS and XRD of the chemical changes occurring on the surface of CdS layers as a result of thermal treatment . Experimental Preparation of CdS thin films by chemical bath deposition The CdS thin films were obtained from solutions of 0.02 mol dm-, CdCl,, 0.065 mol dmP3 NH,C1, 0.15 mol dmP3 CS(NH,), and 0.4mol dm-, NH, at pH 10.3- 1 l.l.I3 The films were deposited on previously cleaned glass substrates with dimensions of 28 x 48 x 1.0 mm3 by dipping, four times, in a reaction bath at 50 "C for 30 min intervals.The thermal treatment of the films, which was aimed at achieving photosensitivity, was performed after dipping them into a solution of 0.7 mass% CdCl, in absolute CH30H. When the glass substrate, with a CdS film deposited on it at 25"C, was removed from the saturated methanol solution of CdCl,, an evaporation process began on the substrate surface. Methanol is an appropriate solvent since it is volatile and because cadmium chloride is only poorly soluble in it (a saturated methanol solution of CdC1, attained only 1.16 mass% CdCl,). A thin film of the crystalline solvate CdC1, -2CH,OH was formed on the CdS film surface.26 As the crystalline solvate is hygroscopic and absorbs moisture from the air, desolvation and hydration to the crystalline hydrates CdCl, * H20 and CdC1, 2.5H20 occurred.The thermal treat- ment was carried out in a quartz furnace under Ar at 400°C for 30 min. Under these conditions, the crystalline solvate and possibly formed crystalline hydrates were desolvated. As a result, anhydrous CdC1, appeared on the film surface, and this led to recrystallization of CdS. The substances used for the experiments were: CdC12.2.5H20 (A. R., Merck), CS(NH,), (A. R., Merck), NH4C1 (A. R.), aqueous NH, (A. R.), CH30H (A. R., Merck) and HzO (double distilled). The anhydrous CdCl, was obtained by thermal dehydrati~n~~ of CdC1, .2.5H2O, and the CH,OH used was dehydrated by a standard method.28 XPS and XRD measurements The XPS measurements were performed in the ultra-high vacuum chamber of an ESCALAB-Mk I1 (VG Scientific) J.Mater. Chem., 1996, 6(3), 377-380 377 electron spectrometer with a base pressure of 1x Pa. The spectra were taken at normal emission with A1-Ka excitation (hv= 1486.6 eV) and an analyser pass energy of 50 eV. The total instrumental resolution was 1.2eV as measured by the full width at half maximum (FWHM) of the Ag 3ds,, photoelec- tron peak. The energy scale was calibrated using C 1s (at 285 eV) of adventitious carbon as a reference. The CdS stan- dard was prepared by pelleting CdS powder (A. R., Merck, 99.999%) in air.The XP spectra were recorded numerically with an Apple I1 plus 8 bites personal computer and treated with original software provided by VG 1000-Scientific. The spectra were additionally transferred to an AT 486 DX2 instrument and processed with Fourier deconvolution (FD) Spectra TOO IS,^^^^* PHI-MATLAB V4 and Microcal Origin v.3.54 software. The atomic concentrations were calculated using normalized photoelectron intensities I/. where 0 is the photoelectron cross-~ection.~~ The X-ray analysis was carried out with a DRON-2 X-ray diffractometer using a cobalt anode, Ka radiation and a nickel filter for a radiation. Contacts of In, Au, Ag and C were deposited on the CdS layers by solder, vacuum evaporation and sputtering methods as well as in the form of paste.The layers were illuminated with a lamp having a heatable tungsten wire, at 100 mW crnp2. Results and Discussion XPS characteristics of CdS films The results from the XPS study are given in Fig. 1-7. Fig. 1 presents the XP spectrum of the CdS standard within the range 5-1005eV. Owing to the high purity of the CdS used for tabletting, lines for Cd and S are observed on the tablet surfaces, the C impurity being due to the oils of the vacuum pumps. The oxygen on the surface originates mainly from the atmosphere under which the CdS standard sample was pre- pared. The Cd: S : 0 ratio on the sample surface in atom% is 45.7 : 42.5 : 11.8. The oxygen on the surface at 532 eV is most probably chemisorbed oxygen.32 The peak positions (Cd 3d5,, at 405.3 eV and S 2p3,, at 161.7 eV) correspond to the com- pound CdS.33,34 Fig.2 presents the XP spectra of Cd 3d of standard (l),as-deposited (2) and heat-treated CdS (3). Both peaks (Cd 3d,,, and Cd 3d3,,) of the standard and as-deposited samples have FWHM = 1.7 eV and positions of the maxima at 405.4 and 412.1 eV, respectively, which correspond to Cd in the form of CdS. For sample 3 the FWHM of the Cd 3dS,, and Cd 3d3,, peaks is increased to 3 eV and the maxima are observed at 405.5 and 412.3 eV, respectively. Fig. 3 shows the results from a fit procedure with PHI-MATLAB V4 of a XP spectrum of Cd 3d5,, of heat-treated CdS. Here, 2.9 atom% Cd are bound as CdCl,, and 28.2 atom% Cd are present as CdS, CdO and Cd(OH),.Fig. 4 presents the S 2p XP spectra of as-deposited (1) and heat-treated (2) CdS. The CdS layer I I 1 I ri Cd 3dY2 200 400 Mx) 800 1000 binding energylev Fig. 1 XP spectrum of standard CdS I . I . 1 . 1 . 400 405 410 415 420 binding energylev Fig. 2 Cd 3d XP spectra of CdS films: 1, standard; 2, as-deposited; 3, heat-treated CdS- 16oooo -140000 -120000 -I ;loo000 -c. t3 8 8oooo-1 .-a UJ c.s--.-C 40000 -2oooo -0--2oooo' I I ' I400 ' 402 ' 404 ' 406 ' 408 ' 410 binding energyleV Fig.3 Cd 3d,,, XP spectrum of heat-treated CdS. 0, Experimental spectrum; -, fit spectrum; . . . , .., synthesized spectra; 0, residue (experimental spectrum minus fit spectrum). obtained by CBD in solution contains S in the form of SO4,-, the amount of S being 2.4 atom%.After thermal treatment under Ar, the sulfates disappear from the surface of the CdS film. Fig. 5 demonstrates the results after the application of the FD procedure to the S 2p spectrum of as-deposited CdS. The peak at 168.7eV corresponds to S present as SO4,-, probably as CdSO,. Fig. 6 illustrates the results from XP spectra of C1 2p of as-deposited (1) and heat-treated (2) CdS. The chlorine concentration in as-deposited CdS is 0.2 atom% and is due to the reagents CdC12*2.5H20 and NH,C1 used for the preparation of the CdS layer by CBD. On the surface of the heat-treated CdS, a concentration of 16.5 atom% C1 as CdC1, is established. Fig.7 shows the XP spectra of 0 1s of as-deposited (1) and heat-treated (2) CdS. The XP spectral contour of 0 1s for heat-treated CdS is difficult to interpret because on both sides of the main peak, which can be deconvol- uted into two peaks at 532 and 533.7eV, there are two new peaks. The surface oxygen is probably chemisorbed and in the 378 J. Muter. Chem., 1996, 6(3), 377-380 160 165 170 175 binding energy/eV Fig. 4 S 2p XP spectra of CdS films. 1, As-deposited CdS; 2, heat- treated CdS. -'-IY Ii! .-c3 I 1. 1 1 I 160 165 176 binding energy/eV Fig.5 S 2p XP spectra of an as-deposited CdS film (no. 4, K = 100-200). +, Experimental spectrum; . . . . .., deconvoluted spectrum. form of CdS04, CdO and Cd(OH)2. The larger oxygen content on the CdS surface after thermal treatment is due to the hygroscopicity of the CdC1, formed as a result of decompo- sition of the crystalline solvate CdC1, 2CH30H.The con- centrations of the elements in atom% are determined on the surfaces of the CdS layers, as follows: as-deposited CdS, Cd :S : C1: 0=48.9 : 25 :0.2 :25.9; heat-treated CdS, Cd:S:Cl:0=31.1: 10.6:16.5:41.8. XRD characteristics of CdS films Fig. 8 shows the results from X-ray analysis of a heat-treated CdS film in comparison with the standards35 a-and P-CdS. The as-deposited CdS layer obtained after precipitation in the reaction bath is amorphous. After dipping of the amorphous CdS layer in a 0.7 mass% CdCl, solution, and thermal treatment under Ar at 400 "C for 30 min, recrystallization sets in and a hexagonal CdS film is formed.195 200 205 210 binding energylev Fig. 6 C1 2p XP spectra of CdS films. 1, As-deposited CdS film (no. 4, K = 100-200); 2, heat-treated CdS film (no. 4,K = lo4). 525 530 535 540 binding energy/eV Fig. 7 0 1s XP spectra of CdS films. 1, As-deposited (no. 4, K = 100-200); 2, heat-treated (no. 4,K = lo4). Electrical characteristics of as-deposited and heat-treated CdS Table 1 summarizes the data on the electrical characteristics of as-deposited and heat-treated CdS films. All layers are deposited at the same concentration of substances participating in the reaction bath at 50 "C. The best photoconductivity ratio (K = lo4) of a CdS layer suitable for solar cells was exhibited by sample no.4 which was dipped in a solution of 0.7 mass% CdC1, in CH30H and heat treated under Ar at 400°C for 30 min. Conclusions Thermal treatment at 400°C under Ar of as-deposited amor- phous CdS films leads to their crystallization into hexagonal CdS. A combination of thermal treatment, pre-treatment of the CdS surface with a methanol solution of CdC1, and exposure of the layer to air results in enhancement of the J. Mater. Chem., 1996,6(3), 377-380 379 Table 1 Resistance characteristics of as-deposited and heat-treated CdS films pD (in !2 cm)=dark resistivity, pL (52 cm) =light resistivity, K (photoconductivity ratio) =pD/pL,no 1 two bath immersions, pH 11 3, 45 min, no 2 4 bath immersions, pH 11 14, 10 3, 10 3, 10 3, 25 min, no , 3 four bath immersions, pH 10 3, 20 min, no 4 4 bath immersions, pH 11 18, 10 3, 10 3, 10 3, 30 min heat-treated CdS as-deposited CdS 200°C, 30 min, Ar no PD PL PD PL 1 107 (3-6) x lo6 Kz3 3 x lo6 K=15 2 x lo6 2 6 7 x lo6 3 x lo6 5 x lo6 25x106 K=2 K=2 3 3 x lo6 12x 105 15x107 3 x lo6 Kz20 K=5 4 2 x los (1-2) x 103 K = 100-200 2 x 104 K = 3-10 2 x 103 6 7 8 9 10 11 12 13 14 15 16 17 dlA 18 Fig. 8 X-ray analysis of (a)a heat-treated CdS sample (no 4) obtained 19by the chemical bath deposition technique, compared with (b)a-CdS and (c)P-CdS standards 20 21 amount of oxygen on the film surface due to the hygroscopicity of anhydrous CdCl, Treatment of as-deposited CdS prelimin- 22 arily dipped in a methanol solution of 0 7 mass% CdCI, and 23 heat treated under Ar at 400°C for 30min makes the film photosensitive, the ratio between the photocurrents in the light 24 and in the dark being of four orders of magnitude 25 26 This work was supported by a grant from EC Brussels under 27 contract ERB JOU2-CT92-0241 28 29 References 30 1 D C Reynolds and G M Leies, Electr Eng, 1954,73,734 31 2 D Bonnet,M Carter, M Burgelman,N Romeo,A W Brinkman, 32 M Carter, E Ozsan, H Hogg and J Vedel, Comm Eur Communities, [Rep] EUR 1993, (EUR 15098) 59, Chem Abstr , 33 120 11645m 3 R W Buckley, E Kuetz, D Valik and F Steueter, Comm Eur 34 Communities [Rep] EUR 1993, (EUR 15098), 173, Chem Abstr , 119277036n 35 4 H W Schock, Sol Energy Mater Sol Cells 1-4,1994,34, 19 5 G A Kitaev, S G Mokrushin and A A Uriskaya, Kolloidn Zh (Russ ), 1965,28,51 380 J Muter Chem , 1996,6(3), 377-380 400"C, 30 min, Ar, 4OO0C, 30 min, Ar CdCl, PD PL PD PL 103 7 x 10, 14x 10' 4 x 104 K<2 K =300 5 x 103 103 4 x lo8 1O6 1o6 K=5 105 1os K =400 lo6 K=lO 14 x 107 105-1o6 K = 103 2 5 x 107 2 x 103 K 10-100 K = 104 G A Kitaev, A A Uriskaya and S G Mokrushin, Zh Kolloidn Khim (Russ ), 1965,34,2065 T G Leonova, T V Kramareva and V M Shulman, Zh Kolloidn Khim (Russ), 1968,30,61 T G Leonova and V I Kozbanov, Izv Sib Otd Akad Nauk SSSR Ser Khim Nauk (Russ ), 1977,7,108 N R Pavaskar, C A Menezes and A P B Sinha, J Electrochem SOC,1977,124,743 J M Doiia and J Herrera, J Electrochem SOC , 1992,139,2810 K I Grancharova, J G Bistrev, L J Bedikjan and G B Spasova, J Mater Sci Lett, 1993,12,852 R W Buckley, Phys Educ , 1992,27,323 R W Buckley, Proceeding of the 11th EC Photovoltaic Solar Energy Conference, Montreux, Switzerland, 1992, p 962 N G Dhere, D L Waterhouse, K B Sundaram, 0 Melendez, N R Parikh and B Patnaik, J Mater Sci Mater Electr, 1995, 6, 52 A Mondal, T K Chaudhuri and P Pramanik, Sol Energy Mater , 1983,7,431 P K Nair and M T 5' Nair, Sol Cells, 1987,22, 103 P K Nair, J Campos and M T S Nair, Semicond Sci Technol, 1988,3,134 J A Ugai, E M Averbah, A S Skuratov and T V Gavnkova, Zh Neorg Khim, 1990,352192 M T Nair, P K Nair, R A Zingaro and E A Meyers, J Appl Phys , 1993,74,1879 M T S Nair, P K Nair, R A Zingaro and E A Meyers, J Appl Phys, 1994,75,1557 R L Call, N K Jaber, K Seshan and J R Whyte, Jr , Sol Energy Mater , 1980,2,373 S A Ringel, A W Smith, M H MacDougal and A Rohatgi, J Appl Phys, 1991,70,881 A Rohatgi, R Suharsanan, S A Ringel and M H MacDougal, Sol Cells, 1991,30, 109 M Stoev and S Ruseva, Monatsh Chem , 1994,125,599 M Stoev, S Ruseva and B Keremidchieva, Monatsh Chem , 1994, 125,1215 M Stoev and I Zlateva, Z Anorg Allg Chem, 1988,558,223 C Dural, Anal Chim Acta, 1959,20,263 Purlfieation of Laboratory Chemicals, ed D D Perrin, W L F Armarego and D R Perrin, Pergamon, Oxford, 1980, p 252 T Kalkanjiev, V Petrov and J Nikolov, Appl Spectrosc, 1989, 43,44 V H Astinov and M Stoev, J Raman Spectrosc , 1994,25,381 J H Scofield, J Electron Spectrosc , 1977,8, 129 C A Estrada, P K Nair, M T S Nair, R A Zingaro and E A Meyers, J Electrochem SOC , 1994,141,801 S W Gaarenstroom and N Winograd, J Chem Phys, 1977, 67, 3 500 V G Bhide, S Salkalachen, A C Rastogi, C N Rao and M S Hegde, J Phys D, 1981,14,1647 Powder Diflraction File, International Center for Diffraction Data, Pennsylvania, 1979,6, 314, 1979, 10,454 Paper 5/05302A, Received 8th August, 1995

ISSN:0959-9428    

DOI:10.1039/JM9960600377

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

The solid solution BaLi, -,Cu,PO, (x <0.5): an example of Cu+ single-ion luminescence in oxide insulators P. Boutinaud,"C. Parent,""G. Le Flem," B. Moineb and C. Pedrinib "lnstitut de Chimie de la Mati2re Condensie de Bordeaux, Chateau Brivazac, Avenue du Dr A. Schweitzer, 33608 Pessac Cidex, France bLaboratoire de Physico-Chimie des Matkriaux Luminescents, Universiti Lyon I, Bdtiment 205, 43, boulevard du 11 novembre 191 8, 69622 Villeurbanne Cidex, France The crystallographic and spectroscopic properties of the solid solution BaLi, -$u,PO, (0<xG0.5) have been investigated as a function of temperature and monovalent copper concentration. A structural description of the compounds is proposed, on the basis of an isotypy with the hexagonal form of p-SiOz trydimite.This description involves the existence of a 1:1ordering between the cross-linked LiO, and PO, tetrahedra, in a way such that the replacement of lithium by copper cannot involve any direct Cu+-Cu+ interaction in the network. This argument is supported by the spectroscopic results. The materials exhibit a unique blue emission under UV excitation, whatever the temperature and the composition. This luminescence possesses a time-constant which is independent of the copper concentration. It is attributed unambiguously to Cu single ions. + Several studies have been devoted to the understanding of the luminescence properties of Cu+ ions in insulators. Many of the works concerned copper-doped alkali-metal halide crystals and, therefore, the description of the fluorescence mechanisms in these systems is well documented.lP6 In particular, the emission of the Cu' single ions occurs in these hosts in the near-UV range and arises from forced 3d94s-+ 3d1' transitions.In contrast to the case of the halide compounds, very few reports concern the fluorescence of copper in oxide insulators. Recent investigations have been performed on the monovalent copper-rich materials CuZr,( P04)37and CuLaO,,' on the +copper-doped Na p"-alurnina,' and on phosphate glasses belonging to the BaLiPO,-P,OS system." All these materials exhibited several luminescence signals under UV excitation, which were partially related to the strong tendency of Cu' to cluster in pairs into large sites. The attributions of these signals, either to single or to associated ions, were often deduced from the spectroscopic data obtained from halide crystals, and have never been experimentally proved.To clarify this point, the need of an oxide insulator contain- ing only one type of luminescent centre is obvious. For instance, the spectroscopic investigation of unique Cu' needs host structures in which this ion can unambiguously occupy a small site, without any aggregation effects. This condition is fulfilled by replacing lithium by copper in the monophosphate BaLiPO,. Materials The elaboration of the solid solution BaLi, -,Cu,PO, (0<x<0.5) required two steps. Mixtures of BaCO, (Aldrich, 99.999%), (1-x)/2 LizCO, (Aldrich, 99.997%), x/2 CuO (Merck, and (NH,),HPO, (Aldrich, 98%) were heated for 15 h under an Nz flow at 200, 400 and 600 "C successively, in order to remove the CO,, H,O and NH, gases.Then, the copper was introduced as a fine metallic powder (Merck, 99.7%) and the final mixture was fired at 850°C for 15 h in a sealed silica tube, in the presence of metallic copper. BaLiPO, was prepared by using a similar process, but the solid-state reaction was performed in the air, with a final thermal treatment at 900 "C. No divalent copper was detected by EPR spectroscopy up to x=0.2. Above this value, the Cu2+ quantity increases slightly, but remains very low (60.5%). The chemical analyses of the two samples BaLiPO, and B~L~,~,,CU~~~~PO,were performed at the Service Central d'Analyses du CNRS of Vernaison (France).The results, reported in Table 1, reveal a very small barium deficit in both compounds. Structural Features Copper-free phosphate BaLiPO, An X-ray powder diffraction pattern of the copper-free phos- phate BaLiPO, was recorded between room temperature and 600°C, using the Guinier-Simon technique, and is shown in Fig. 1. A phase transition occurring between 150 and 220 "C is clearly visible. The low-temperature (LT)form was obtained after slow cooling from 600 "C to room temperature (RT), at a rate of 0.1 "C rnin-', whereas the high-temperature (HT) form could be isolated after strong quenching in air from 900"C, in a Pt/Rh crucible. Fig. 2 compares the X-ray patterns of these two phases to that published by Paques-Ledent." In agreement with this author, the LT form of BaLiPO, [Fig.2(b)] was indexed to Table 1 Calculated and experimental amounts (in mass%) of Ba, Li, P and Cu in BaLiPO, and BaLio~,oCuo~SOPO, BaLiPO, calc.expt. (fO.l) Ba 57.4 55.9 Li 2.9 2.9 P 12.9 13.1 cu - - BaLi,,,Cu0,,PO, calc. expt. (kO.1) 51.3 50.3 1.3 1.3 11.6 11.6 11.9 12.0 Fig. 1 Guinier-Simon diagram of BaLiPO, J. Mater. Chern., 1996, 6(3), 381-384 381 10 40 30 20 2Oldegrees Fig. 2 XRD patterns of BaLiPO, at 300 K (a) After ref 11, (b) LT form, (c)HT form the orthorhombic system, whereas the HT form was indexed using a hexagonal unit cell, on the basis of an isotypy with the p-SiOz trydimite l2 Solid-solution BaLi, -,Cu,P04 (0 <x 60.5) The introduction of copper stabilized only the hexagonal form, even for the smallest x values (Fig 3) The unit-cell parameters, calculated after least-squares refinement of the X-ray data, of both varieties of BaLiPO, and of the compounds x =0 01,O 10 and 0 50, are collected in Table 2 Experimental densities agree with the values calculated, assuming four formula units per orthorhombic unit cell and two per hexagonal unit cell The evolution of the a and c parameters us the copper concentration, x, follows Vegard's law (Fig 4) In these phosphates, the copper ions are assumed to occupy the lithium site, in a tetrahedral coordination This environ- ment is not usual for Cu' ,but it has been reported previously, e g in molybdates l3 l4 Luminescence Properties Experimental techniques The emission and excitation spectral distributions were ana- lysed using a computer-controlled Jobin-Yvon HR3 mono- chromator, equipped with a liquid helium cryostat The signals were detected with an R928 Hamamatsu PMT instrument 50 40 30 20 28/degrees Fig.3 XRD patterns of (a) BaLi, 99cuo 01P04,(b)BaLio 90Cuo 1oP04 and (c)BaLi, 50Cuo sop04 5 70 $5 652 Fig.4 Evolution of the a and c parameters of the BaLi, ,Cu,P04 hexagonal unit cell, as a function of x The continuous emission spectrum of a xenon lamp was used as the excitation source Care was taken to correct appropriately the profiles of the luminescence spectra The fluorescence decays were obtained between L/HT and RT using the 280nm pulsed excitation of a Q-switched Nd3+-YAG pumped dye laser beam, doubled and frequency mixed with the 1064pm fundamental The emission was dispersed by a Hilger computer-scannable monochromator and detected with an R1477 Hamamatsu PMT instrument, followed by an amplifier/discriminator and a photon counter Table 2 Symmetry, unit-cell parameters and densities (expt and calc ) for BaLiPO, (LT and HT), BaLi, 99cuo 01P04, BaLi, 90Cuo loPo, and BaLIO5OCu0 5OP04 symmetry unit-cell parameters/A volurne/A3 density (expt )/g cm density (calc)/g cm-3 BaLiPO, (LT) BaLiP04 (HT) BaLi, ,9Cu0 01P04 BaLiO 9OCu0 lop0, BaLiO 50cu0 5OP04 orthorhombic hexagonal hexagonal hexagonal hexagonal a =8 640(4) ~=5 131(2) ~=5 179(2)127(2) ~=5 138(2) ~=5 b=5 204(2) c=8 664(2) c =8 666( 2) c=8 653(2) c =8 627(4) c=8 749(4) 393 37 197 23 197 58 197 82 200 39 3 98(4) 3 97(4) 4 W4) 4 lO(4) 443(4)4 039 4028 4030 4 110 4 430 382 J Muter Chem, 1996, 6(3), 381-384 The decays were recorded on a Canberra multichannel ana- lyser, operating with a minimum dwell-time of 200 ns per channel.Experimental results The luminescence spectra were analysed at 10 and 300K, under 300 nm continuous excitation, for the compositions x = 0.01, 0.10 and 0.50. Whatever the temperature and the Cu+ concentration, a unique emission band was detected (Fig. 5). The main characteristics of this blue luminescence are collected in Table 3. When the temperature is increased, the emission band broadens and the position of its maximum shifts slightly towards the low-energy side.The broadening is greater when the copper concentration is increased. The excitation spectra consist of a broad band peaking in the vicinity of 290 nm, with a shoulder at 260 nm. The spectral position of this band is not altered by changing the temperature or the copper concentration. The temperature dependence of the blue emission time- constant, z, is shown in Fig. 6, for the compositions x=O.Ol, 0.10 and 0.20. All the decays were purely exponential. z decreases markedly upon warming from 4.2 to 30K, then excitation emission I 1 200 300 400 500 GOO AInm Fig. 5 Excitation and emission spectral distribution at 10 K (continu-ous line) and 300 K (broken line), for (a) BaLio.,,Cuo.olP04, (b) BaLi,,,,Cu, 10P04and (c) BaLi, soCuo.soP04.I,, =460, A,,, = 300 nm. Table 3 Emission and excitation spectral characteristics at 10 and 300 K of BaLi, _,Cu,P04 (x =0.01, 0.10 and 0.50) emission excitation peak position peak position x value T/K /nm /cm-' FWMH/cm-' /nm /cm-' 0.01 10 300 438 471 22830 21230 2260 3265 286 288 34900 34710 0.10 10 300 439 466 22765 21460 2335 3450 286 289 34880 34540 0.50 10 300 447 464 22344 21550 2595 3873 285 291 35050 34370 55 50 v)4 45 40 0 I0 20 30 40 50 60 70 80 90 100 T/K Fig. 6 Temperature dependence of the blue luminescence time-constant z, for BaLi,,,,Cu, 01PO4 0, BaLi,,,,Cu,,loP04 (x) and BaLi,,,,Cu,,,,P04 (0).The continuous and broken lines represent the best fits to the experimental data, using eqn.(1) and the parameters given in Table 4. more gradually up to 80 K. Between this temperature and 300 K, z remains constant. Below 7 K, the onset of a plateau is observed. This peculiar thermal evolution of z is typical for the UV copper(1) luminescence previously observed in several rock- salt systems like NaCl,, NaF4 or LiC1.6 Similar features were also reported for one of the fluorescence signals detected in andCUZ~,(PO~)~~in Cu' -doped borate and phosphate glas~es.'~,'~This behaviour is typical of a three-level system, which is described by the rate equation: z-'= [A31+A21 exp(-~/kT)]/[l+ exp(-~/kT)], where A,, and A31 are the radiative transition probabilities between the excited states (upper level 2 and lower level 3) and the ground state and E represents the energy mismatch between the two excited states.Table 4 collects the best parameters obtained after a fit of the experimental data, using the rate equation, for the compounds BaLio~,,Cuo~,,P04 and BaLi,~,,CuO~,,PO,. In opposition to these strong thermal effects, z remains independent of the copper concentration. Discussion The copper-free phosphate BaLiPO, exhibits no luminescence, even at low temperature, indicating that the observed blue emission is due only to the presence of monovalent copper. This luminescence shows the same characteristics as those reported by Wanmaker and Spier in B~L~PO,:CU'.~~ In particular, it is clear from Table 3 that this emission arises from the 3d94s-,3d1' vibronic transition.More specifically, an increase of copper concentration does not generate any additional emission due to Cd-Cu' pairs, as was observed in several oxide insulator^.^.^'^^ The absence Table 4 Transition probabilities and energy mismatch between the emitting levels involved in the luminescence process BaLi0.99Cu0 01p04 17 700 35 000 13 BaLi0.90Cu0. 10p04 18 350 34 500 10 J. Muter. Chem., 1996, 6(3), 381-384 383 value of the blue luminescence decay-time with respect to the Cu+ concentration evidences the absence of any Cu+-Cu+ LIO, coupling at short or long range Conclusion o Ba ie Fig.7 Representation of the structures of (a) BaLiPO, (LT) and the (b)p-SiO, trydimite (from ref 11) of such centres is consistent with the structure of the materials According to Paques-Ledent, the structure of the orthorhombic form of BaLiPO, is made up of LiO, and PO4 tetrahedra, sharing common corners, giving rise to a three-dimensional skeleton which contains large cavities available for the Ba2+ ions The LiO, and PO4 tetrahedra are assumed to be 1 1 ordered, as illustrated in Fig 7(a) For each ring perpendicular to the c axis, three adjacent tetrahedra are arranged 'peak up' and the three others 'peak down' This arrangement is close to that of the J3-SiO, trydimite, in which the Si04 tetrahedra peak alternately up and down [Fig 7(b)] Thus, the phase transition observed for BaLiPO, consists of a cooperative upside-down permutation of some tetrahedra but has no influence on the Li(Cu) P ordering, even for the highest copper concentrations Consequently, this distribution excludes direct interactions between the Cu' ions In addition, the constant The investigated barium lithium copper phosphates exhibit a unique blue luminescence under UV excitation, attributed unambiguously to Cu' single centres This conclusion is consistent with the 1 1 Li(Cu) P ordered distribution in the tetrahedral sites of the crystal structure and also with the independence of the luminescence time-constant with respect to copper concentration Therefore, the spectroscopic proper- ties of these compounds can be considered as reference data to interpret the spectroscopic behaviour of copper(r) ions in more complex oxide insulators, which are investigated in the context of the research of tunable solid-state lasers emitting in the visible range References 1 J Simonetti and D S McClure, Phys Rev B, 1977,16,3887 2 C Pedrini and B Jacquier, J Phys C Solid State Phys 1980, 13,4791 3 H Chermette and C Pedrini, J Chem Phys , 1981,75, 1869 4 B Moine and C Pedrini, Phys Rev B, 1984,30,992 5 S A Payne, A B Goldberg and D S McClure, J Chem Phys, 1983,78,3668 6 C Pedrini, Phys Status Solzdi, 1978,87,273 7 P Boutinaud, C Parent, G Le Flem, B Moine and C Pednni, J Phys Condens Mater , 1992,43031 8 J P Doumerc, C Parent, J C Zhang, G Le Flem and A Ammar, C R Acad Sci Paris, 1982,306,1431 9 J D Barrie, B Dunn, G Hollingworth and J I Zink, J Phys Chem, 1989,93,3958 10 P Boutinaud, E Duloisy, C Pedrini, B Moine, C Parent and G Le Flem, J Solid State Chem , 1991,94,236 11 M T Paques-Ledent, J Solid State Chem , 1978,23,147 12 R E Gibbs, Proc R Soc London A, 1926,113,351 13 B 0 Marinder, P E Werner, E Wahlstrom and G Malmros, Acta Chem Scand A, 1980,34,51 14 E M McCarron I11 and J C Calabrese, J Solid State Chern , 1986, 65,215 15 S A Payne, L L Chase and L A Boatner, J Luminescence, 1986, 35,171 16 W L Wanmaker and H L Spier, J Electrochem Soc, 1962, 109, 109 Paper 5/04363H, Received 5th July, 1995 384 J Mater Chem, 1996, 6(3), 381-384

ISSN:0959-9428    

DOI:10.1039/JM9960600381

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Synthesis and structural investigation of the Eu, _,Bi,VO4 scheelite phase: X-ray diffraction, Raman scattering and Eu3 luminescence+ Jean Luc Blin," Annick Lorriaux-Rubbens,"" Francis Wallart" and Jean Pierre Wignacourt' "Laboratoire de Spectrochimie Infrarouge et Raman, UPR CNRS A2631 L, Universitk des Sciences et Technologies de Lille, Bdtiment C5, 59655 Villeneuve d'Ascq Cedex, France 'Laboratoire de Cristallochimie et de Physicochimie du Solide, URA CNRS 452, Universitk des Sciences et Technologies de Lille, Ecole Supkrieure de Chimie de Lille, Boite Postale 108, 59652 Villeneuve d 'Ascq Cedex, France A previous investigation of EuV0,-BiVO, binary system has indicated the existence of a solid-solution of Eu, -,Bi,V04 up to x =0.60, with a zircon-type structure.In this work we have reinvestigated the synthesis conditions, thus extending the upper limit to x =0.74. X-Ray diffraction techniques associated with Raman and luminescence spectroscopies have led to the unambiguous identification of the structural type as a scheelite model. Rare-earth-metal ions are introduced as dopants in many systems because of their potential applications in communi- cations or lasers technologies.lp2 Partial rare-earth-metal sub- stitutions have already been investigated in vanadate systems such as Ln,-,Bi,VO, (Ln=Eu, Gd), mainly for their ferro- electric proper tie^,^,^ and also for Eu3 l~minescence.~+ In studies aimed at developing new materials, we have studied the binary diagram EuV04-BiV04 in a limited com- position domain corresponding to a 35-90% bismuth for europium substitution. This particular limitation results from a previous in~estigation:',~ a maximum substitution ratio of 0.65 was then proposed; and near this limit, an interesting enhancement of the Eu3+ luminescence was noted, but the synthesis problems were not completely clarified.On the basis of the structural results of the starting compounds BiV04 and EuVO,,'** we have studied the reproducibility of the synthesis conditions, found the upper limit of the solid-solution domain, and have investigated the structure of the corresponding composition. Experimental Synthesis conditions The studied compounds were obtained from the solid-state reaction of the appropriate proportions of decarbonated and dehydrated bismuth, europium and vanadium oxides, according to reaction (1).( 1-u)Eu,03 +xBi,O, +V205+2Eu, -,Bi,V04 ( 1) The starting materials were ground and mixed in an agate mortar; the resulting mixture was then calcined at 850°C in an aluminium crucible for several hours and then air quenched. After several intermediate regrindings, the completeness of the reaction was checked by powder X-ray diffraction (XRD). Powder diffraction Room-temperature diffraction patterns were obtained from a Guinier de Wolff camera, using Cu-Ka radiation (0.154178 nm). NH4Br was introduced as an internal standard, and allowed the indexation of the pattern, the refinement of the lattice parameters and the identification of the solid-solution limit.The Rietveld refinement of the structure was made from a powder sample; the data was taken on a Siemens D5000 diffractometer equipped with a Cu anode, a back monochroma- tor and a rotating sample holder (2mm depth) in the range 8 <28 <120°, with a scanning step of 0.02". Molecular spectroscopy Raman scattering and luminescence spectra were recorded on a computerized RT30 Dilor spectrometer, consisting of three dispersive stages (800 mm focal length) in an additive optical scheme, each stage having a plane holographic grating with 1800 lines permm. This apparatus is equipped with two different laser sources, i.e. ionised argon and krypton (2025 Spectra Physics), thus giving access to a set of exciting lines.The detector is a photomultiplier (Hamatsu R943 02) with an AsGa photocathode cooled by the Peltier effect.' In order to investigate polycrystalline compounds, we chose the 647.1 nm exciting line of the Kr' laser with 25 mW power at the sample. The spectrum obtained at 0.7 cm-I resolution is recorded in the spectral range 6-1500 cm-l. Results X-Ray diffraction analysis We have investigated samples within a composition range 0.35<x<0.90, using 0.05 steps for x, and identified two different behaviours: first, a tetragonal solid-solution is unam- biguously observed for 0.35 <x <0.70 (domain 1);then, in the range 0.75 <x <0.90, a biphasic mixture of the previous tetra- gonal phase and a monoclinic one, is noted (domain 2).X-Ray patterns of domain 2 compositions were the key to this identification:" we used mixtures of constant NH,Br mass (q),as an internal standard, and a constant mass m2 of a sample of a given composition x;then a non-saturated NH4Br diffraction line is used as a reference, and the evolution of a selected reflection of the monoclinic phase is plotted versus x. The desired limit point, x =0.74, is obtained when I(mono- clinic)/l(standard)=0, i.e. when the tetragonal phase is pure (Fig. 1).This was confirmed by a specific sample preparation, and the corresponding compound EUO$%O 74vo4 has tke following lattice parameters: a =7.282(2) A and c =6.430( 2) A. This new limit, which is quite different from the literat~re,~ results explicitly from the longer synthesis time: 1% weeks at 850 "C, thus leading to complete reaction.Several attempts at preparing single crystals of Euo 26Bi0 74v0, have failed; thus we decided to investigate the structure from a powder sample, using the Rietveld method. In order to avoid any composition problems, we selected a sample (x=0.72) just below the limit of the solid-solution. According to the literature, EuVO, presents a tetragonal structure of the ~ircon-type,~ space group 14,/arnd(D,,19), with J. Muter. Chem., 1996, 6(3), 385-389 385 cl 0 70 075 0 80 0 85 090 subshtution ratio, x Fig. 1 Determination of the solid-solution limit from the X-ray patterns (a) I(sample/I, (NH,Br), (b) I(sample)/I,( NH,Br), (c), I(sample/I,(NH,Br) [I,(NH,Br) =most intense line of NH,Br, 12(NH4Br) =less intense line of NH,Br, Z,(NH,Br) = Il(NH4Br)+ 12(NH,Br)] 4 1 I I(a) 28/degrees Fig.2 Computed Rietveld diffraction profiles compared with exper- imental data for the Euo2,Bio7zV04 composition (a) in the 14,la space group, (b) in the 14, famd space group lattice parameters a=7 2373(2) A,c=6 3661(3) A and 2=4 As for BiV04, at room temperature, the structure is monq- ~linic,~space 6roup 12/b (C2,6),ofergusonite type, a =5 196(1)A, b= 11 704(2) A, c=5 093(2) A, ,!?=90 38(2)", Z=4 Both phases present a high-temperature variety, the scheelite-type structure,'' l2 tetragonal, in the space group 14,/a (C4h6) The data refinement was performed for the two tetragonal possibil- ities, Fig 2(a) and (b),and a good fit between the experimental and simulated data is noded in both cases, the ~$11parameters are then a =7 28296(9) A and c =6 43407( 10)A The corres- ponding structural features are given in Table 1 The only difference results from the oxygen position, in a special (025) or a general position, but even in the latter case, the deviation compensates for the difference The reliability factors (Table 2) are very similar in both models, and no clear conclusion about the space group can be made at this stage Structural information from molecular spectroscopy In the composition range O<x<O74, Raman spectra of Eul-,Bi,V04 samples (Fig 3) show in the frequency range 6-1000 cm-l some spectral modifications in the low-frequency region (lattice modes) as well as in the typical internal modes such as the V-0 stretching motions in the frequency range 700-900 cm-' Such a change may be due to a space group modification directly linked to the composition When x is close to the maximum value of 074, the Raman spectra can be interpreted in the C4h6factor group, this is confirmed by the luminescence spectra of the corresponding samples irradiated with the 568 2 nm excitation line (Fig 4) where the transition 5DO+7F0 is noted at 398 cm-' As a matter of fact, this band is absent for EuVO,, in good agreement with the selection rules, when Eu is located in the site 4b (D2d),space group 14,lamd But it is observed when the bismuth content is increased, which implies a symmetry change for the Eu/Bi sites, related to a space group modification The location of the mixed cation in the site 4b (S,) of the 14,/a space group, verifies the selection rules and fits the other electronic transitions, such as 5DO-+7F4 which is noted between 1100 and 1300 cm-' under the 647 1 nm exciting line These preliminary results obtained from Raman and lumi- nescence spectroscopies, allow the selection of a space group compatible with the scheelite form, and corresponding to a stabilisation of the high-temperature forms of BiVO, and EuVO,, which both have this structural type 'I l2 Table 2 Final reliability factors for the refinement using the Rietveld method I4,lamd space group 14,/U space group 14 6 14 6 18 5 18 5 13 46 13 45 188 188 3 82 3 84 2 52 2 61 Table 1 Atomic fractional coordinates and thermal coefficients for the two possible tetragonal cells I4,lamd 14, la Xla Ylb z/c B/A2 xla Ylb ZIC B/A2 72% Bi 00 0 75 0 625 0 39(2) 72% Bi 00 0 25 0 625 0 39( 2) 28% ELI 28% Eu 4b 4b v 00 0 75 0 125 032(6) V 00 0 25 0 125 0 31(7) 4a 4a 0 0 1779(7) 025 0 0428 (8) 0 6(1) 0 0 1778(7) 0743(6) 00429(8) 05(2) 16h 16f 386 J Mater Chem , 1996,6(3), 385-389 X = 0.40 I X= 0.60 r X= 0.65 I 200 400 600 800 lo00 1200 1400 200 400 600 800 1000 1200 1400 relative wavenumberkm-I Fig.3 Raman and luminescence spectra of the Eu,-,Bi,VO, compounds in the 6-1500 cm-' frequency range (A,,,=647.1 nm) Eu,.,Bi,,,VO, structure Thus, the crystal structure of Euo.2sBio.72V04 can be described as having tetragonal symmetry, space group 14,/a, with four formula units (Fig.5). The atomic positions in the unit cell are given in Table 1, and the cation position has been refined with 72% and 28% imposed occupancy factors for Bi3+ and Eu3+, Table 3 Interatomic distances and angles in Bi, 72E~0 28V04 interaction number distance/A angleldegrees 4 2 4 2 1.687( 5) 2.59( 5) 2.83(5) 4 4 4 2 4 4 4 4 4 4 2 2 2.406( 5) 2.499( 5) 2.59(5) 2.87(2) 3.07( 5) 3.18(5) 3.48(6) 4.5 1(4) 4.65(3) 4.69( 5) 4 4 4 4 4 4 2 respectively; the thermal coefficients are given in Table 1. The crystal structure is formed from tetrahedral vo43-anions, + +and mixed Bi3 /Eu3 cations; the orthovanadate ions are described by four identical bond lengths of 1.687( 5) A, and six bond angles, four at 114(3)" and two at 100( 3)".In comparison, the geometry of the orthovanadate ion in the highly symmetri- cal form of EuV0413 is clper to an ideal tetrahedron, with V-0 distances of 1.66(7) A and angles of 109(8)Oand llO(9)O; in the BiVO, scheelite form, obtained either by heat treatment', or by high pres?ure" synthesis, the V-0 bond lengths are similar, 1.72(1)A, with four angles at 105.9(4)' and two others at 116.9(8)". Thus our results show an angular deformation of the orthovanadate ion, but they are still compatible with the literature data. A closer analysis of the data of Sleight et a1.,12 David and Glazer14 or Mariathasan et ~1.~'~links the ferro- elastic-paraelastic transition (fergusonite-scheelite) in BiV0, to a deformation of the Bi3+ coordination polyhedron, with a minor influence of the VOd3-anions.In all cases, Bi3+ is surrounded by 8 oxyge; atoms, with bond lengths14 of 2.343, 2.379, 2.50? and 2.640A in the monoclinic phase, and 2.448 and 2.497 A in the tetragonal phase. In this investigation, the mixed trivalent cation coordination is described in Fig. 6: six different vo43-tetrahedra are involved; four of the oxygen atoms are closer than the other four, which are provided by two Vo43- edges. The M"'0, polyhedrons are linked either by a tetrahedral anion, or by two oxygen atoms delimiting a parallelogram with the two corresponding trivalent cations.These results are featured in Table 3. Molecular spectrum assignment Raman scattering. The Raman spectra of these compounds are located below 900cm-' for the Stokes domain, which is J. Muter. Chem., 1996,6(3), 385-389 387 200 400 600 800 11 10 relative wavenumber/cm-1 Fig. 4 Raman and luminescence spectra of the Eul-,Bi,V04 com-pounds in the 6-1000 cm-' frequency range (Ibe,,=568 2 nm) 8 cation M (M = 61,Eu) Fig. 5 Tetragonal cell scheme of the Eu, 28B10 72vo4 scheelite type structure the most intense side (excitation of the electrons of the funda- mental vibrational level) The anti-Stokes side is symmetrical with the Stokes side, with the excitation line as a reference, but the band intensities quickly decrease because only the electrons of the excited vibrational level contribute Thus, it is easy to distinguish the Raman effect from other phenomena such as luminescence in our case For our compounds, we obtained the internal vibrational modes of the V043-anion between 700 and 900 cm-' for the symmetrical and antisymmetrical stretching motions, the most intense band being assigned to the v,(A,) frequency, and the others to v3(F,)vibration, and between 350 and 480 cm-' for 388 J Muter Chem, 1996, 6(3), 385-389 vanadium 8 catDon M (M = Bi.Eu)111 o oxygen Fig. 6 Surroundings of the (a) M3+ cations by the oxygen atoms and their arrangement, (b)vanadium atoms I 5D0-7F0 1 -500 -1000 -1500 wavenumber/cm 1 Fig.7 (a) Raman' Stokes and luminescence spectrum in the 6-1500 cm frequency range, and (b) Raman antistokes and lumi- nescence spectrum in the -2000--200 cm ' frequency range of the Eu, 26Bi0 74vo4 sample (Ae,, =647 1nm) the symmetrical and antisymmetrical bending motions, corre- sponding to the vz( E) and v4( F,) frequencies In the very low- frequency domain, the obtained lines are due to lattice modes On the basis of our previous work,' and using the correlation tables between molecular, site and factor groups, we can assign the different lines observed by Raman scattering as indicated in Table 4 Luminescence spectroscopy.Using 647 1 nm radiation as the exciting line, two bands are clearly obtained in the 1080-1300cm-1 frequency range, and some others in the -1800 to -400 cm-' anti-Stokes domain, which are not due to vibrational motions [Fig 7(u) and (b)] We have already explained the Eu3+ luminescence observed under experimental conditions by a phonon-assistance mechanism l6 These elec- tronic transitions involve the first emitting level, 'DO,located in our compound at 17212cm-l, and the 7FJlevel of this cation Those transitions ascribed to 5Do+7F3transitions are expected in the exciting line range Because of the strong intensity of this exciting line, the presence of the lattice vibrations and the weakness of these electronic transitions.it Table 4 Internal vibrational mode assignment and lattice mode localisation of the Bio,74Euo,,,V04 Raman spectrum molecular group site group factor group Td s4 C4h frequency/cm~ A 8 59 A 368 B B 738 E 767 B 448 E lattice modes 3T(M3-): B+E 250 3T(V043-): 208 B+E 119 76 59 3L(V043-): A+E Table 5 Observed emission wavelength of the Eu3+ luminescence in the Bi0,74E~0,26V04 scheelite structure and its expected numbers, using the 647.1 nm excitation line of the Kr' laser" number of electronic expected observed relative emission transition transition types transition types frequency/cm-' wavelength/nm 5D, j7F0 1 md 5~o+7~, 1 md 1 (ed, md) 'Do +7F2 1 md 2 ed 1 (md, ed) 'Do j7F3 1 md 2 ed 2 (md, ed) 5D0+7F4 3 md 2 ed 2 (md, ed) "md =magnetic dipole; ed =electric dipole.would be difficult to detect them under such experimental conditions. The expected 5D,+7Fj transitions with the Eu3+ cation in an S4 symmetry site of the scheelite cell are presented in Table 5, where the observed relative wavenumbers, taking into account the exciting line reference, and their corresponding emission wavelengths are summarized. Conclusion An optimization of the synthesis conditions has led to a new definition of the tetragonal solid-solution limits. The refinement -1758 581.0 -1385 593.9 -1397 593.4 -998 607.8 -906 61 1.3 -805 615.1 -707 6 18.8 162 653.9 1118 697.6 1233 703.2 3 J. W. Hur, H. C. Lee, M. S. Jang, D. H. Yo0 and H. K. Kim, Ferroelectrics, 1990, 109, 197. 4 M. S. Jang, M.S. Lee, H. C. Lee, J. W. Hur and H. K. Kim, Ferroelectrics, 1990, 109, 185. 5 J. Ghamri, PhD Thesis, Universite de Lille I, 1990. 6 A. Lorriaux-Rubbens, J. Corset, J. Ghamri and H. Baussart, Adu. Mater. Res., 1994, 1-2,433. 7 A. T. Aldred, Acta Crystallogr., Sect. B, 1984,40,569. 8 J. D. Bierlein and A. W. Sleight, Solid State Commun., 1975, 16, 69. 9 G. Walker, Cryocoolers, Plenum Press, New York, 1983. 10 P. Conflant, These de 3" Cycle, Lille I, 1975. 11 I. H. Ismailzade, R. N. Iskenderov, A. I. Alekberov, R. M. Ismailov, A. M. Habibov and F. M. Salayev, Ferroelectrics, 1981,31,45. of this composition are developed. References 1 F. Auzel, Opto. 65, Echo des Recherches, 1992,143,24. 2 C. Hsu and R. C. Powell, J. Luminescence, 1995,10,273. 12 A. W. Sleight, H. Y. Chen and A. Ferretti, Nut. Res. Bull., 1979,of the structure of a composition Eu~.~~B~~.~~V~~,using the 14, 1571. Rietveld method on a polycrystalline sample near the solid- 13 W. 0.Milligan and L. W. Vernon, J. Phys. Chem., 1952,56, 145. solution limit, associated with Raman scattering and lumi- 14 W. I. F. David and A. N. Glazer, Phase Transitions, 1979, 1, 155. nescence spectroscopies, describes the crystal structure in a 15 J. W. E. Nariathasan, R. N. Hazen and Z. W. Finger, Phase scheelite model. The structural and molecular characterisations Transitions, 1986, 6, 165. 16 A. Lorriaux-Rubbens, J. L. Blin, L. Rghioui, F. Wallart, J. P. Wignacourt, A. Mizrahi, M. Drache, P. Conflant, Proceedings of the 14th International Conference on Raman Spectroscopy, ed. N. T. Yu and X. Y. Li, Wiley, New York, 1994, pp. 566-567. Paper 51048736; Received 24th July, 1995 J. Mater. Chem., 1996, 6(3), 385-389 389

ISSN:0959-9428    

DOI:10.1039/JM9960600385

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

~~~~~~ ~~ Microwave-assisted selective deoxygenation of layer- and chain-containing oxides Balasubramanium Vaidhyanathan, Munia Ganguli and Kalya J. Rao* Solid State and Structural Chemistry Unit, Indian Institute of Science, Bangalore 560 012, India Very rapid (within 5 min), selective, single-step deoxygenation of layer- and chain-containing oxides, MOO,, CrO,, V205, a-VOPO4-2H,O and Ag,MoloO,, has been accomplished using graphitic carbon in a microwave-assisted reaction. The products were found to be MOO,, Cr203, VO,, VPO, and a mixture of (Ag +MOO,), respectively. Products were characterised by X-ray diffraction (XRD), differential scanning calorimetry (DSC), IR and electron paramagnetic resonance (EPR) spectroscopies. Although conventional methods of preparing these materials are tedious, the present method is simple, fast and yields very homogeneous products of good crystallinity.Our results reveal that while layer- and chain-containing oxides undergo rapid microwave-assisted carbothermal reduction, the non-layered materials do not. The high structural selectivity of these reactions is suggestive of the topochemical nature of the fast reduction process. Microwave heating offers several advantages for the synthesis of inorganic materials compared to conventional methods. The greatest advantages appear to be the very short timescales involved in the preparation and the selectivity in energy transfer from the microwave field.',, A number of oxides (e.g. V205, WO,, Cr203 and MnO,) and a few halides (e.g.AgI and CuI) absorb energy from microwaves very efficiently and can be heated to 1000 K or more.,-, Carbon in the form of a powder is also known to be a good microwave susceptor and is heated It has been realized that microwave radiation can be used to accelerate a variety of chemical reactions, and products of good structural uniformity have been obtained.*yg In this paper we report a novel microwave- assisted selective single-step deoxygenation of simple and binary layered oxides using graphitic carbon as a reductant. Our results reveal that while layer- and chain-containing oxides are prone to microwave-assisted carbothermal reduction, non-layered materials are not. Experimental Analar grade MOO,, CrO,, V205 and high purity graphitic carbon were used as starting materials for the preparation of simple lower-valent oxides.a-VOPO4-2H,O was prepared by refluxing V,O, in a solution of distilled water and concentrated H3P04 for 16 h in air." The product was characterised using X-ray powder diffraction (XRD), IR spectroscopy and electron paramagnetic resonance (EPR)spectroscopy. The layered silver decamolybdate (Ag,Mo,,O,,) was prepared by heating a mixture of Ag,MoO, and MOO, in air'' and was identified by XRD. A microwave oven (Batliboy, Eddy) operating at 2450 MHz with variable (integral) power levels up to a maximum of 980 W was used in our preparations. Powders of the layered oxides MOO,, V205, a-VOP0,-2H20 and Ag,Mol0O,, and the chain oxide CrO, were thoroughly mixed with graphitic carbon in a 3: 1 mass ratio (4 g batch) and the mixtures in silica crucibles were placed on a firebrick inside the microwave oven.Each mixture was exposed to microwaves for 4-5min. The charge begins to warm within 45 s and the material becomes red-hot in 2 min. The products were allowed to cool inside the oven. Carbon was found to be almost completely burnt out. However, any traces of excess carbon were washed out using water (carbon particles are very fine and simply float to the surface). In the case of the CrO, +C mixture the excess carbon can be burnt out. The products were identified by XRD as MOO,, Cr203, VO, and VPO, and were further characterised using differential scanning calorimetry (DSC), EPR and IR spectroscopies. Longer exposure to microwaves of the MoO,+C mixture leads to the reoxidation of MOO, to MOO, (white needle-like crystals) at the surface.In order to avoid this, the carbothermal reduction of MOO, was carried out in an NH, atmosphere. NH, was also created in situ by decomposing a mixture of ammonium metavanadate and V,05 in a separate crucible alongside in the microwave oven. Even a 10 min exposure did not cause reoxidation. We did not observe any evidence of melting of the oxides during the reduction process, indicating that these fast reactions occur entirely in the solid state. Carbothermal reduction of a number of non-layered oxides (WO,, TiO,, SnO,, Nb205 and ZrO,) were also examined under similar conditions.Independent experiments revealed that of the oxides used, only V205,WO,, ZrO, and TiO, are good microwave susceptors. In other cases, the graphitic carbon alone is responsible for the microwave absorption. The temperatures of the various reaction mixtures were monitored as a function of microwave exposure time. The temperatures were measured by interrupting very briefly the microwave exposure and quickly inserting a chromel-alumel thermocouple into the crucible. Measurement of the tempera- ture in microwave oven is very difficult. The actual tempera- tures can be up to 30-40°C higher than recorded here (differences are high at higher temperatures). Results Fig. 1 gives the X-ray diffraction patterns of the products, namely MOO,, Cr203 and VO,, obtained by the microwave- assisted reduction of the simple oxides MOO,, CrO, and V,O,, respectively. The good crystallinity of the products is evident.The lattice parameters of the products obtained agree well with the reported data (see Table 1j. In the case of VO,, two additional peaks (intensity not more than 10%)corresponding to the high temperature modification of VO, were seen.', This may be due to retention of the high-temperature phase since the present method of preparation involves rapid cooling., The DSC trace of microwave-prepared VO, exhibited a clear endothermic transition at 340 K. This corresponds to the semiconductor-metal transition of VO,, as reported by earlier workers.', The EPR spectrum of the prepared Cr203 and that of the commercially available Cr203 were compared, and we found that the resonances of the microwave-prepared sample are J.Mater. Chem., 1996, 6(3), 391-394 391 I I I30 40 50 € 2Ndegrees Fig. 1 X-Ray powder patterns of (a) MOO,, (b) Cr,O, and (c) VO,, * denotes the peaks corresponding to the high-temperature phase of VO, Table 1 Lattice parameters of the microwave-prepared lower-valent oxides lattice parameters compound calc reported ref ~~ ~ MOO, a=5 615 A b=4 859 .$ 5 607 A 4 860 4 19 c=5 533 A 5 537 A p = 119"32 u =4 962 A, C= 13 5820A a=57424 b=4 519 .$ 119"350 13 5940A 5 743 4 4 517 4 4 959 AD 25 20 c=5 377 A 5 375 A VPO, p= 122"5$ a=5 220 4 b=7 770 .$ 122'6q 5 245 4 7 795 4 22 c=6266A 6 285 A sharper We attribute this feature to the better crystallinity of the microwave-synthesized sample This simple deoxygenation procedure was extended to more complex oxides also a-VOPO, 2H,O, which is a layered hydrate, was found to undergo rapid carbothermal reduction, giving rise to VPO, Fig 2 gives the X-ray diffractogram of the product a-VOPO, 2H,O and VPO, were further charac- terised by EPR spectroscopy At room temperature, the EPR signal of V4+ in a-VOPO, 2H20 (V4+ is always found to be present in this compound in low concentrations because the presence of phosphorus in a vanadium-oxygen lattice improves the reducibility of vanadium) appeared anisotropic with hyper- fine structure, comparing well with the earlier reports l4 The product VPO, showed a simple EPR signal with a g-value of 1 97 This is in good agreement with the EPR signal emanating from a V3+ ion (g= 1 97) in V203 From Fig 2 it can also be seen that in the case of silver decamolybdate, which is also layered, the reduced product shows peaks corresponding to both metallic Ag and MOO, 392 J Muter Chem, 1996, 6(3), 391-394 1 I I I 1 1 10 20 30 LO 50 60 28ldegrees Fig.2 X-Ray diffractograms of the products (a) VPO, and (b) the (Ag+ MOO,) mixture, see Fig 1 for indices of reflections of MOO, 0 1 2 3 4 tlmin Fig. 3 Time-temperature profiles of the various mixtures under micro- wave exposure 0,MOO,+ C, *, V205+ C, 0,W03+ C, A,T10, + C It is important to note that though MOO, and WO, are chemically similar, only the layered MOO, undergoes micro- wave-assisted carbothermal reduction, while WO, does not This clearly demonstrates that the structural chemistry of these materials plays an important role in these fast reactions This is further proved by the fact that a number of non-layered oxides such as TiO,, SnO,, Nb205 and ZrO, do not undergo carbothermal reduction when they are subjected to similar microwave irradiation (even under longer microwave exposures) The time-temperature profiles of the various reaction mix- tures are shown in Fig 3 In all the cases the temperature shows an initial abrupt increase followed by a levelling-off region It is clear from the figure that the maximum tempera- tures reached in the microwave-assisted reaction are different for different mixtures for the same exposure time Discussion Lower-valent metal oxides exhibit rich structural chemistry and have important technological applications MOO,, which has a distorted rutile structure, is chemically very interesting because of the presence of metal-metal bonding l6 Cr203 and vanadium phosphates have received a wide amount of interest owing to their important catalytic properties l7 VO, exhibits interesting semiconductor-metal transition l3 Note that the conventional method of preparation of MOO, requires 90-100 h of heating in the absence of air l9 Cr,O, is normally prepared by thermal decomposition of Cr6+ oxide for several hours.20 Since the direct reduction of V205will not give a well defined product, VO, is usually prepared by the reaction of V203 and V205 in evacuated quartz tubes for 40-60 h at 1023-1073 K.The conventional method of prep- aration of VPO, either involves the reaction of (NH,),HPO, and NH,V03 at 1223 K in an argon atmosphere for several hours,'l or the reduction of a-VOPO, in a controlled high- purity hydrogen atmosphere.,, Hence there is a necessity for developing new methods of preparation of these materials. The present microwave-assisted carbothermal reduction route is much simpler, extremely fast and is a single-step process. Also, microwave-assisted carbothermal reduction takes place at lower temperatures than required for conventional methods. Indeed, similar behaviour was observed for various other microwave When the same sample mixture is kept in a preheated electric furnace at a temperature slightly higher than that reached in microwave reaction, no reduction is observed for similar time durations.This clearly demon- strates that the reaction mechanism is quite different for the conventional and the microwave routes. A close look at the time-temperature profile (Fig. 3) reveals that higher levelling-off temperatures are attained if both the constituents are microwave susceptors (WO, and V20, are good microwave susceptors at room temperature while TiO, absorbs microwaves at higher temperature). Though the reac- tion temperatures reached in the case of non-layered oxides (WO, and Ti0,) are much higher than those of the layered oxides, reduction is not found to occur.Thus the deoxygenation procedure is found to be highly structure-selective in the sense that only layered and chain structure compounds are reduced in the microwave process, implying a strong topochemical propensity for the reaction. Graphitic carbon is a good micro- wave absorber and we have previously suggested7 that the microwave susceptibility may originate from the activation of interlayer (weak) graphitic bonds (in amorphous carbon, how- ever, its lossy character may result in the absorption of microwave energy,). microwave-activated graphitic carbon P MO 0Oxygen 0 Fig. 4 (a) Layered MOO, structure (two layer blocks are shown). The oxygen atoms between the layers are removed by reaction with carbon, as marked.(b) The Soordination of Mo atoms in MOO, (longest Mo-0 bond =2.33 A). 0 0dL VO P0,*2H20 VOPO4 c = 7.41 A c = 4.434 A Fig.5 Schematic representation of the layer separation of the dihydrated and dehydrated a-vanadyl phosphate MOO, has a unique layered structure [Fig. 4(u)]. Of the six oxygens of the MOO, octahedron, three are common to three octahedra, two are shared between two octahedra and the sixth one is ~nshared.'~ The Mo-0 bond distances in MOO, octahedron are given in Fig. 4(b).,, The unshared oxygen is connected to Mooweakly through the unusually long Mo-0 distance of 2.33 A. We suggest that the microwave-activated graphitic carbon may slip into the interlayer region of MOO, and readily react with the weakly connected oxygen, causing the reduction.A similar mechanism may be operative in layered v20514715 CrO, also. However, and in the ~hain-type'~ further confirmation of the suggested mechanism is necessary. a-VOP0,.2H20 is an attractive potential host for a variety of coordination intercalation reactions." In the crystal structure of a-VOP04.2H,0 the two water molecules are inserted between [VOPO,], layers. Even after dehydration, a-VOPO, retains the layered structure.14 A schematic represen- tation of the layer separation of a-VOP04-2H20is given in Fig. 5. During the carbothermal reduction of a-VOP0,-2H20, water is removed as indicated by the disappearance of the IR absorption bands associated with water (3300and 1600 cm-I).The dehydrated a-VOPO, layered structure can be thought of as having V05 square pyramids. The layers are held together only by a weakly coordinated apical oxygen of one square pyramid to the open square face of a vanadium pyramid in the adjacent layer.'' Here, also, we consider that rapid micro- wave-assisted deoxygenation can occur through the removal of the weakly connected interlayer oxygen, resulting in V3+ phosphate, VPO,. It is possible that in the case of which is also layered,,, both the interlayer silver and the weakly connected oxygens of the molybdate units simul-taneously react with activated carbon giving rise to metallic Ag and MOO, (Fig. 2). Further experimental studies using XRD and high resolution electron microscopy (HREM) are in progress to obtain a clear insight about the reaction mechanism. It is interesting to note that a variety of non-layered oxides such as WO, (distorted form of cubic ReO, structure26), TiO, (anatase), SnO, (rutile-type structure), Nb205 and monoclinic ZrO,, when subjected to microwave exposure with graphitic carbon, do not undergo any reduction.Even longer microwave exposures (10min) do not give reduced products in these cases. WO, .2H20, although layered, does not undergo reduction by microwave-activated carbon under similar conditions. This may be due to the fact that unlike a-VOP0,.2H20,WO3.2H,O not only loses water (probed by IR spectroscopy) but also its layered structure, and no reduction occurs.Thus only layered oxides are found to be reduced in microwave-assisted reduction by graphitic carbon. This is also emphasized by the fact that none of the non-layered lower-valent product oxides exhibit any tendency to undergo further reduction. Conclusions A number of simple and binary layered metal oxides have been easily reduced to lower-valent oxides in a very short time by microwave-assisted selective deoxygenation, and the meas- ured physical properties of the microwave-prepared materials J. Muter. Chem., 1996, 6(3), 391-394 393 agree well with those of the conventionally prepared ones A possible topochemical mechanism has been proposed for these fast reactions It is suggested that any layered or one-dimen- sional (chain) oxide with weakly bonded interlayer or interchain oxygen, can be selectively deoxygenated by the 8 9 10 B Vaidhyanathan, M Ganguli and K J Rao, Muter Res Bull, 1995,30,1173 D R Baghurst, A M Chippindale and D M P Mingos, Nature, 1988,332,311 J W Johnson, A J Jacobson, F Brody and S M Rich, Inorg Chem, 1982,21,3820 microwave-activated carbon The present method appears to be quite general, simple, rapid and exhibits structural selectivity 11 12 13 14 C Rosner and G Lagaly, J Solid State Chem , 1984,53,92 F Theobald, J Less-common Metals, 1977,53,55 F J Morin, Phys Rev Lett, 1959,3,34 D Ballutaud, E Bordes and P Courtine, Muter Res Bull, 1982, 17,519 The authors thank Professor C N R Rao, FRS, for encourage- ment and the referees for useful suggestions One of the authors (M G ) thanks CSIR, India for financial support 15 16 17 A F Wells, Structural Inorganic Chemistry, Clarendon Press, Oxford, 1975, p 473 F A Cotton and G Wilkinson, Advanced Inorganic Chemistry, Wiley Eastern Limited, India, 1979, p 947 C P Poole Jr and D S MacIver, Adv Catal, 1967,17,223 18 J T Vaughey, W T A Harrison, A J Jacobson, D P Goshorn References 19 and J W Johnson, Inorg Chem , 1994,33,2481 Nut Bur Stand (US)Monogr ,1981,25( 18), 44 1 A G Whittaker and D M P Mingos, J Chem Soc, Dalton Trans, 1992,2751 2 C C Landry and A R Barron, Science, 1993,260,1653 3 D M P Mingos and D R Baghurst, Chem SOCRev, 1991,20,1 4 D R Baghurst and D M P Mingos, J Chem Soc, Chem Commun , 1988,829 20 21 22 23 24 H McMurdie, Powder Diffraction Journul, 1987,2,45 N Kinomura, F Muto and M Koizumi, J Solid State Chem, 1982,45,252 J Tudo and D Carton, C R Acad Sci C Paris, 1979,289,219 F Levy, Structural Chemistry of Layer- Tvpe Phases, Reidel, Dordrecht, 1976, p 172 B M Gatehouse and P Leverett, Chem Commun , 1969,1093 5 B Vaidhyanathan, M Ganguli and K J Rao, J Solid State Chem , 25 Anderson, Acta Chem Scand, 1954,8, 1599 1994,113,448 26 A R West, Solid State Chemistry and its Applications, Wiley, 6 H K Worner and D H Bradhurst, Fuel, 1992,72,685 Chichester, 1989, p 32 7 P D Ramesh, B Vaidhyanathan, M Ganguli and K J Rao, J Muter Res ,1994,12, 3057 Paper 5/04607F, Received 13th July, 1995 394 J Muter Chem, 1996, 6(3), 391-394

ISSN:0959-9428    

DOI:10.1039/JM9960600391

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Structural and magnetic properties of Sr -,M,IrO (M =Ca Zn Cd Li Na) Nanu Segal," Jaap F. Vente," Timothy S. Bush"*band Peter D. Battle," "Inorganic Chemistry Laboratory University of Oxford South Parks Road Oxford UK OX1 3QR bDavy-Faraday Laboratory The Royal Institution of Great Britain 21 Albemarle Street London UK W1X 4BS The compounds Sr4-,CaxIr06 (x=0.5 1,2,3,4) and Sr,MIrO (M =Li Na Zn Cd) have been synthesized and studied by X-ray powder diffraction and SQUID magnetometry. They all adopt the K,CdCI structure with the cations M and Ca occupying only the trigonal prismatic sites in the [Ool] chains unless x > 1. Those compounds which contain only IrIV order antiferromagnetically at temperatures between 14 and 22 K and those which contain only IrV show temperature-independent paramagnetism.The development of a self-consistent interatomic potential for the IrIV-O interaction is described. The lattice parameters predicted for the K,CdCl,-like phases using this new potential agree with the experimental values to within ca. 2%. Recently the level of interest in ternary and quaternary oxides of the platinum metals has been increasing rapidly with the K,NiF,-like compounds Sr2M04 (M=Ru Rh and Ir) attracting particular attenti011.l~~ Oxides of Ir Ru Rh and Pt which are isostructural with K4CdCl6 are also being studied exten~ively~-~and we describe below the synthesis and charac- terisation of some iridium-containing members of this struc- tural family. Sr,IrO was prepared originally by Randall and Ward,' and subsequently by Rodig and McDaniel and Schneider," but a detailed description of the crystal structure has only been published relatively recently." It is a rhombohedra1 K4CdC1,- like structure (Fig.1) in which IrO octahedra and SrO trigonal prisms share triangular faces to form chains of alternat- ing polyhedra parallel to the three-fold axis. The interchain space is occupied by the remaining alkaline-earth-metal cat- ions. Substitution of Ca for all the Sr leads to the formation of the isomorphous compound Ca,Ir0,.12*13 It has been rep~rted~,'~that the transition-metal cations Ni2+ Cu2+ and Fig. 1 Crystal structure of Sr41r06. IrO octahedra are hatched open circles represent Sr2 ions.+ Zn2+ can be substituted for the strontium cations in trigonal- prismatic coordination.In the case of Sr,CuIrO the structure distorts to give a square-planar coordination around the Cu atom and there is a concomitant lowering of the symmetry to monoclinic (C2/c).14. The magnetic susceptibilities of the com- pounds Sr,MIrO (M=Sr Ni Cu Zn) have been studied in detail. For M =Sr or Zn the compounds are antiferromagnetic below 1215and 20 K5respectively. The success of an alternating Heisenberg chain model in accounting for the magnetic behav- iour of the latter compound has been taken to indicate that a structural distortion as yet undetected in diffraction experi- ments is present in Sr,ZnIrO,. Sr,CuIrO is a weak ferromag- net (T,~40K) with a saturation magnetization of 0.61 pBt per mole while the magnetic susceptibility of Sr,NiIrO shows a complex behaviour with different (anti)ferromagnetic inter- actions dominating in different temperature region^.^ Ca,lrO has been shown to be paramagnetic in the temperature range 77<.T/K<290,l3 but no measurements have been made at lower temperatures. We undertook the present study of cation substitution in Sr,IrO with a number of aims in mind. We hoped to stabilize iridium in a formal oxidation state of five in this structure by substituting one quarter of the strontium with lithium or sodium. The substitution of strontium by calcium zinc and cadmium was carried out to enable us to study the structural and more importantly the magnetic properties of Ir" in a range of isomorphous compounds.Such a study was deemed to be interesting because of the possibility of a transition from antiferromagnetism to ferromagnetism as was observed when Mg was replaced by Zn in La2Mgl-xZn,Ir0,.'5. The final aspect of the work described below is concerned with the use of computer modelling in crystal chemistry. We recently devel- oped' a set of self-consistent potentials for M"+-02-and 02-02-interactions by simultaneously fitting the crystal properties (e.g. unit-cell parameters relative permittivity and elastic constant) of a wide range of binary metal oxides using a single shared O2-O2 -potential. Transition-metal cations with electron configurations other than do high-spin d5 and d" were omitted from that study a deficiency which we began to rectify by developing RuIV-O and RuV-0 pair p0tentia1s.l~ These potentials are valid in relatively complex ternary oxide stru~tures,'~,~~and their value was demonstrated by our recent ab initio determination of the structure of Li,Ru0,.18 We believe it is important to establish the extent to which this type of potential can be used in structural inorganic chemistry and we have therefore developed an IrIV-O potential and t pB (Bohr magneton)=9.27 x J T-I.1.Mater. Chern. 1996 6(3) 395-401 395 tested it against the behaviour observed in the system Sr -,Ca,IrO,. Experimental The synthesis of polycrystalline samples of Sr4-xMxIr06 (M = Li Na Ca Zn and Cd) was initially attempted by firing the appropriate stoichiometric mixtures of dry SrCO Ir Li2C03 Na,CO CaCO ZnO and CdO (all Johnson Matthey Chemicals). The reaction mixtures contained in alumina cru- cibles were heated at 800°C for 1 day in order to decompose the carbonates.Subsequently the temperature was increased in a stepwise manner to the following values Sr,-,Ca,IrO (0.5bx62) and Sr,ZnIrO 1000°C (for up to 2 weeks); Sr,CdIrO 950°C (1 day); Sr,LiIrO 800°C (2 weeks); and Sr,NaIrO 900 "C (4 days). The syntheses were interrupted frequently to facilitate re-grinding and pelletizing of the samples. All the preparations were carried out in air. Attempts to prepare Ca,IrO in this way failed. A mixture of Ca -xIr301219 and CaO was produced initially and the compo- nents of this mixture subsequently reacted only very slowly at temperatures up to 1200°C.We were thus unable to produce a single-phase sample of Ca,IrO by this method. However we were able to suppress the formation of Ca5~xIr3012 and prepare a sample of Ca,IrO by firing a stoichiometric mixture of CaCO and Ir in air at 1100°C for 24 h without prefiring at 800°C. The disadvantage of this method is that it can lead to the loss of iridium owing to the formation of volatile iridium oxides.20 Attempts to prepare Cd,IrO and Sr4.-xMxIr06 (M = K Mg Co and Pd) were unsuccessful as were those to synthesize Ba -,M,IrO,. The progress of the reactions was monitored using X-ray powder diffraction (XRD) and they were deemed to be com- plete when the diffraction pattern did not change on heating the sample further.X-Ray data (Cu-K,a) were recorded on the final products of successful reactions using a Siemens D5000 diffractometer operating at room temperature in Bragg-Brentano geometry. These data were collected over the angular range 15< 28/" d 100 using a 28 step size of 0.02". The results of our powder diffraction experiments were analysed by the Rietveld method2' using the GSAS program package.22. The peak shape was described by a pseudo-Voigt function and the background level was fitted with a shifted Chebyshev function. For each diffraction pattern a scale factor a counter zeropoint three to five peak shape parameters four background param- eters and two unit-cell parameters were refined.The number of atomic parameters (fractional coordinates and thermal parameters) refined varied as will be described below. The magnetic susceptibility of each product was measured using a Quantum Design or a Cryogenic Consultants supercon- ducting quantum interference device (SQUID) magnetometer. Samples were loaded at room temperature and measurements were usually taken in the temperature range 2 < T/K d300 after cooling the sample in zero magnetic field (zero-field cooled ZFC) and also after cooling in the measuring field (field cooled FC). However Sr,CaIrO and Sr,LiIr06 were only measured under FC conditions. The applied field was chosen to lie in the range 1 dH/kG b 10. In the Born model the energy associated with the inter- actions between a pair of ions i and j in an ionic solid can be written as qtqJ = #lJ(rLJ) -(1)rV The second term on the right hand side of eqn.(1) is simply the long range Coulomb interaction. The first term represents the short-range interatomic interactions and is often taken to have the form of the Buckingham potential These short-range interacti9ns are cut off for interionic dis- tances r, greater than 12A. The parameters A, ptJ,c,,are then the variables to be refined when a new ion-pair potential is developed.16. The values which best describe the Ir"-O interaction were calculated using the static lattice computer program GULP,23 which is based on the Born model of the ionic lattice. The parameters were derived from a simultaneous consideration of the structural data on Sr41r06" and La2MgIr06.24 We restricted ourselves to the use of these two model compounds because we believe that the reliability of the potential parameters for the interaction between a heavy metal and oxygen is likely to be enhanced if structures deter- mined from neutron diffraction data are used in the construc- tion of the potential.A detailed description of the fitting methodology has been given previously.16 Potentials describing the two-body interactions between other ion pairs (cg. Sr-0) were taken from Bush et al..',. The newly determined Ir'"-O potential was used to predict the lattice parameters of Sr,-,Ca,IrO (x= 1,4) and Sr,IrO,.,. The number of composi- tions that could be considered for the former system was limited by our inability to deal with compositions having a disordered cation array.Results We were able to prepare samples with the general formula Sr4-,MXIrO6 for M =Ca (x =0.5 1 2 3 and 4) and for M = Zn Cd Li or Na (x=l). The samples of composition Sr,,Ca,,IrO and Sr,Ca,IrO were used only to provide additional data for our study of the composition dependence of the unit-cell parameters; they were not investigated by magnetometry. The refinements of the structures of the majority of the compounds under investigation proceeded smoothly in the hexagonal setting of the space group R3c. However in the case of Ca,IrO we detected the presence of a minor (ca. 1%) face-centred cubic impurity phase having a ~4.8A.We believe this to be CaO and we assume that the enforced omission of the 800 "C firing resulted in a small loss of Ir as II-O. The lattice parameters determined by the profile analyses are presented in Table 1 along with the agreement indices; the lattice parameters determined previously" for Sr41r0 are included for comparison purposes. The fact that the DW-d values2 are smaller than the lower limit of the 90% confidence level indicates that the estimated standard deviations associ- ated with the refined parameters are likely to be underestimated owing to serial correlation effects. It became clear that the dopant cations (M=Li Na Ca Zn and Cd) occupy the 6a (O,O,a)Sr( 1) site in the [OOl] chains rather than the larger 18e (x,O,$) Sr(2) site in the interchain space.In the Ca-doped samples having 1<x <4 the 6a sites are occupied only by calcium and the 18e sites are occupied by a disordered arrangement of calcium and the remaining strontium. The values of the variable coordinates of the latter site and the 36f (x,y z) oxygen site are given in Table 2 along with those of the thermal parameters U,,,. The resulting bond lengths are given in Table 3. The low X-ray scattering factors of Li Na and 0 resulted in poor definition of their structural parameters and in order to ensure physically reasonable values it was sometimes necessary to constrain U, to be the same for all the ions in the structure or to fix the value for a light atom to be equal to the overall temperature factor determined in preliminary refinements.Furthermore in some refinements it was necessary to fix the fractional coordinates of the oxide ions to ensure that chemically reasonable bond lengths were maintained. Such instances are indicated in Table 2. The observed and calculated diffraction patterns for two representa- tive structure determinations Ca41rO6 and Sr,ZnIrO are given in Fig. 2 and 3. All the Ir" compounds studied show a susceptibility maxi- mum in the temperature range 12< T/K<22. The results of our measurements on Ca41r06 are shown in Fig. 4. Qualitatively similar results were obtained for all compositions in the series Sr4-,Ca,Ir06. The data collected on Sr,CdIrO are plotted in Fig. 5. They show a paramagnetic tail at low Fig.3 Observed (dots) calculated (full line) and difference X-ray powder diffraction profiles of Sr,ZnIrO,. Reflection positions are marked. temperatures suggesting that the sample contains a very small amount of an impurity phase. The results of our experiments on Sr,ZnIrO are different (Fig. 6) in that the ZFC and FC susceptibilities do not overlie below ca. 90 K. The data in the temperature range 70 <T/K <296 have been modelled for all these compounds using the Curie-Weiss Law with an additional term a,to allow for a temperature-independent paramagnetic (TIP) contribution to the molar magnetic suscep- tibility (T-6) (3) The predicted value of C for a low-spin d5 ion on a site of cubic symmetry is 0 375 emu K mol-'.The TIP contribution is inversely proportional to the spin-orbit coupling constant [ of the cation which can be evaluated for Ir" as 26 (4) The resulting parameters are listed in Table 4 along with the temperature of the susceptibility maximum It can be seen that the latter changes little in the series Sr4-,Ca,Ir06 but increases in the Zn and Cd compounds. The data on the IrV compound Sr,LiIrO do not show a susceptibility maximum (Fig 7) and eqn (3) was used to fit them over the full measured temperature range. The magnetic susceptibility of a low-spin d4 ion (J =0) is predicted to be temperature-independent and the value of C refined by fitting to eqn (3) can be used to estimate the IrIV content of the sample and hence in the case of a monophasic 26sample the vacancy concentration on the anion sublattice.The derived value of C (Table4) leads to a composition Sr,LiIrOS 99 for our sample Electron repulsion effects render the relationship between the spin-orbit coupling constant of IrV and the TIP27more complex than in the case of Ir" and [ cannot be evaluated analytically in the former case.The results of a preliminary experiment to measure the magnetic susceptibility of Sr,NaIr06 in the temperature range 80<T/K<300 are shown in Fig 8(a). The data are qualitat- ively similar to those measured for Sr,LiIrO and the resulting parameters are included in Table 4 However when the sample was remeasured after being stored in air for some months the results were somewhat different [Fig 8(b)] with the suscepti- bility showing a small but sharp increase on cooling below ca 250 K We take this to indicate that this compound is unstable in air and that ferromagnetic Sr21r0 is one of the decompo- sition products.This phase was undetectable by XRD and the magnitude of the susceptibility change suggests that it is present in a concentration of only ca 500 ppm The potential parameters AZJ,pv and C tor the Ir"-O Jon pair refined to the values 1772 83 eV 0 3039 A and 1 98 eV AP6 Table 4 Magnetic susceptibility parameters for Sr4-xMxIr06 Fig. 7 Inverse molar magnetic susceptibility of Sr,LiIr06 measured in a field of 5 kG. 0,field cooled; - Curie-Weiss with TIP. respectively. Energy minimisation calculations using these par- ameters predicted the unit-cell parameters listed in Table 5.In addition to the model compounds the table includes the compositions Sr,CaIrO and Ca,IrO which are central to this work and as an additional test compound Sr21r04. The experimental values are also tabulated and for ease of compari- son the observed and calculated unit-cell parameters of the compounds Sr -,Ca,IrO are plotted as a function of composi- tion in Fig. 9. The calculated cell dimensions agree with the experimentally determined values to within 2%. Discussion The results described above show that the Sr cations occupying the trigonal prismatic 6a sites in Sr,IrO can be replaced by a T/K Fig. 8 Inverse molar magnetic susceptibility of Sr,NaIrO measured in a field of 1 kG (a)initially and (b)after ca.2 months. A zero-field cooled; 0,field cooled; - Curie-Weiss with TIP. Table 5 Predicted and observed unit-cell parameters for some iridium oxides predicted experimental difference compound parameter value value (Yo) wide range of divalent or monovalent dopants without causing any major disruption to the crystal structure. In the case of a monovalent dopant charge neutrality could be preserved either by the introduction of anion vacancies or by the oxidation of the transition-metal cations. Our magnetic and structural data both suggest that the latter occurs although we have been unable to find a reliable method of chemical analysis by which we can determine the oxygen content of our samples directly.The relatively straightforward synthesis of two more IrV com- pounds Sr,LiIrO and Sr,NaIrO suggests that this can no longer be considered as an unusual oxidation number for Ir in the solid state although the subsequent decomposition of the Na-containing sample must be noted. The fact that the dopant cations all of which are smaller than Sr2+ occupy the 6a sites in preference to the 18e sites is consistent with the lower coordination number and hence smaller size of the former Although we were able to replace more than 25% of the Sr2+ cations with Ca2+ and thus dope the 18e site we were unable to dope beyond 25% with the smaller Cd2+ cation. The unit-cell volume (Table 1 Fig 9) is a linear function of composition in the series Sr4-,Ca,Ir06 but the unit-cell parameters a and c are not.The volume reduction required by the introduction of Ca is initially (x d 1) achieved by a contraction along the c axis with a remaining essentially constant both a and c decrease as the Ca concentration increases beyond x =1. The net effect is that the rhombohedra1 distortion of the unit cell parameterized by c/a (=J1 5 for pseudo-cubic symmetry) is largest when x= 10. The aniso- tropic nature of the dilation enables the Ir cations to retain their optimum coordination geometry over a wide composition range although the Ir-0 bond length (Table 3) is unusually short in the fully substituted compound Ca,Ir06 For all compositions in the series Sr -,Ca,IrO other than Sr41r06 the IrO octahedra are elongated along the three-fold axis despite the overall compression of the unit cell (c/a<Jl 5) The elongation of the octahedra compresses the trigonal prisms and hence facilitates the accommodation of the smaller Ca2+ cations.The composition dependence of the bond lengths M,,-0 and MlSe-0 in Sr3ZnIr06 and Sr,CdIrO as well as in Sr4-xCaxIr06 reflects the fact that the dopant cation does not occupy the 18e site unless x> 1 Our refinement of the structure of Sr,ZnIrO resulted in a lower R-factor than that reported by Nguyen and zur Loye,' and we have no reason to believe as they did that we have overlooked a structural distortion. The unit-cell parameters reported for the two different samples are in good agreement but it is possible that their sample reacted with the Pt used to line the alumina 400 J Muter Chem 1996,6(3) 395-401 crucible,28 thus modifying the composition slightly.The con- straints imposed during the structural refinement of Sr3NaIr06 and Sr,LiIrO make it impossible to discuss the resulting atomic coordinates in any depth but we note that the rhombo- hedral distortion of the unit cell increases as the size difference between the dopant cation and Sr2+ becomes greater. The same effect can be seen by comparing Sr xCaxIrO (x 6 1) and Sr,ZnIrO although in the case of Sr,CdIrO the distor- tion is not as large as might be expected The magnetic susceptibilities of all the Ir" oxides considered in this study were well modelled by eqn (3) in the high- temperature region In each case (Table 4) 0 is small and negative thus indicating that weak antiferromagnetic inter- actions are always present. The data suggest that these inter- actions are stronger in the Zn- and Cd-containing compounds The relatively low values of C (<O 375 emu K mol-') are presumably a consequence of the low symmetry of the crystal field at the Ir (6b) site.The mean value of the calculated spin- orbit coupling constants 2 2 x lo4crn-l is comparable to the values which can be deduced from some previous but higher than those reported elsewhere 26. The low-tempera- ture susceptibility data on Sr -,Ca,IrO suggest that all compositions in this series order antiferromagnetically at low temperatures.The composition-dependence of the Neel tem- perature is consistent with a strengthening of the magnetic interactions as the unit-cell volume decreases and the super- exchange pathways get shorter. The temperature of the suscep- tibility maximum observed for Sr,ZnIrO is very similar to that reported previously,' but the value is somewhat higher than would have been expected on the basis of cell-volume arguments However the most striking feature of our data is the previously unobserved low-temperature hysteresis in the susceptibility. The presence of this effect which was reproduc- ible in Sr,ZnIrO but unobserved in the other compounds suggests that blocked spin clusters form at relatively high temperatures (ca 90 K) in the Zn compound alone It is not clear why this should be the case.The Neel temperature of Sr,CdIrO is the highest observed in these compounds as would have been expected from the magnitude of the Weiss constant 8 Again the value is enhanced to a greater extent than can be explained using simple size arguments and it seems likely that the electronic structure of the Zn2+ and Cd2+ cations which both have a relatively polarizable d10 core is an important factor In contrast to Nguyen and zur Loye' we have not attempted to model our compounds as one-dimen- sional magnets Although it is convenient and usual to describe this crystal structure in terms of [OOl] chains we have shown previously' that the intrachain Ir-Ir distance is not significantly shorter than the interchain distance and the structure should be considered as being three dimensional Darriet and Subramanian3' have recently presented a less misleading description of the structure. The temperature dependence of the magnetic susceptibility of Sr,LiIrO is typical of an IrV oxide.The compound is essentially non-magnetic with a weak Curie tail caused by the presence of a small concentration (ca 1 5%) of reduced Ir" cations. This impurity level is comparable to that observed in other IrV oxides". The limited data available on Sr,NaIr06 show a similar behaviour Our attempts to model the structures of complex Ir oxides have met with limited success. The refined potential parameters model La,MgIrO very well with the observed and calculated unit-cell parameters agreeing to better than 1%.The unit-cell volume is also predicted accurately However the agreement is less satisfactory in the case of the compounds Sr,-,Ca,IrO Although each cell parameter is accurate to better than 2% and their non-linear composition dependence is well repro- duced the calculated values are all systematically greater than the experimental values and the calculated unit-cell volume is thus only accurate to ca 4% In the case of Sr21r0 a layer structure which was used only as a test compound the unit- cell volume was accurately reproduced but through the cancel- lation of errors in the individual cell parameters rather than through a truly accurate calculation We conclude that the parameters derived in this study for the Ir4+-02- potential have limited transferability.The reasons for this might include the different environment of the Ir cations in the various 3 4 5 M K Crawford M A Subramanian R L Harlow J A Fernandez-Baca Z R Wang and D C Johnson Phys Rev B 1994,49,9198. T N Nguyen D M Giaquinta and H C zur Loye Chem Mater 1994,6,1642. T N Nguyen and H C zur Loye J Solid State Chem 1995 117,300 compounds. The IrO octahedra in La,MgIrO are linked to neighbouring MgO octahedra through one common vertex whereas those in Sr,-,Ca,IrO share a face with the neighbour- ing trigonal prismatic group thus markedly reducing the Ir-M2+ distance Our previous study of complex Ru oxides17 also showed that different potentials were needed to describe 6 7 8 9 10 S Frenzen and H Muller-Buschbaum 2 Naturforsch.Tell B 1995,50,581 J F Vente J K Lear and P D Battle J Mater Chem 1995 5,1785 J R Randall and R Ward J Am Chem Soc ,1959,81,2629 F Rodi,. Thesis Universitat. Tubingen 1963 C L McDaniel and S J Schneider J Res Nat Bur Stand A face-and corner-sharing situations Furthermore the Sr2+-02 potential derived previously may not be appropriate when the cation occupies a trigonal-prismatic site. The unsatis- factory unit-cell parameters calculated for SrJrO may reflect the inability of our method to deal with anisotropic cation 11 12 13 1971,75,185 A V Powell P D Battle and J G Gore Acta Crystallogr Sect C 1993,49,189 C L McDaniel and S J Schneider J Solid State Chem 1972 4,275 R F Sarkozy C W Moeller and B L Chamberland J Solid environments but it has been suggested32 that the oxide sublattice in this compound is disordered in which case we would not expect our calculations to be accurate Finally note that the new IrIV-O potential described above is defined by a relatively small number of parameters and that many pub- lished potentials include extra terms which model the polariz- 14 15 16 State Chem ,1974,9,242 M Neubacher and H Muller-Buschbaum 2 Anorg A& Chem 1992,607,124 A V Powell J G Gore and P D Battle J Alloys Comp 1993 201,13.T S Bush J D Gale C R A Catlow and P D Battle J Muter Chem 1994,4,831 ability of the cation It may prove possible to improve the transferability of our potential by the use of these additional parameters but the fit to the model compounds used in this study did not improve significantly when they were included 17 18 19 P D Battle,.T S Bush and C R A Catlow J Am Chem Soc 1995,117,6292. T S Bush,C R A Catlow andP D Battle,J Muter Chem 1995 5,1269 F J J Dijksma J F Vente E Frikkee and D J W IJdo Muter Res Bull 1993,28 1145 Conclusion 20 21 J H Carpenter J Less-Common Met 1989,152,35 H M Rietveld J Appl Crystallogr ,1969,2,65 We have shown that Sr41r0 can tolerate a wide variety of cation substitutions on the Sr sublattices Although all the Ir" compounds order magnetically at low temperatures none of the compounds described above shows the spontaneous mag- 22 23 A C Larson and R B von-Dreele General Structure Analysis System Los Alamos National Laboratories Report LAUR 86- 748,1990 J D Gale General Utility Lattice Program (GULP) Royal Institution of Great Britain London 1994 netisation that IS relatively common in complex oxides of Ir l5 More high-quality structural data are needed if reliable Ir-0 potentials are to be developed and used for structure prediction 24 25 26 J G Gore and P D Battle manuscript in preparation R J Hill and H D Flack J Appl Crystallogr ,1987,20,356 K Hayashi G Demazeau M Pouchard and P Hagenmuller Muter Res Bull 1980 16 1013 We are grateful to EPSRC for financial support to C R A 27 J Darriet G Demazeau and M Pouchard Muter Res Bull 1981 16,1013 Catlow for useful discussions and to S G Carling for exper- imental assistance 28 S J Schneider J L Waring and R E.Tressler J Res Nut Bur Stand A 1965,69,245 29 J F Vente Thesis University of Leiden 1994. 30 R C Byrne and C W Moller J Solid State Chem 1970,2,228 References 31 32 J Darriet and M A Subramanian J Muter Chem ,1995,5,543 Q Huang L Soubeyroux 0 Chmaissem I Natali-Sora 1 Y Maeno H Hashimoto K Yoshida S Nishizaki T Fujita A Santoro R J Cava J J Krajewski and W F Peck Jr J Solid. J G Bednorz and F Lichtenberg Nature 1994,372,532 State Chem 1994,112,355 2 T Shimura M Itoh and T Nakamura J Solid State Chem 1992 98,198 Paper 5/06721I Received 10th October 1995

ISSN:0959-9428    

DOI:10.1039/JM9960600395

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Manganese oxide-zirconium oxide solid solutions. An X-ray diffraction, Raman spectroscopy, thermogravimetry and magnetic study Mario Valigi,*" Delia Gazzoli,' Roberto Dragone,G Alessandra Marucci' and Giorgio Matteib Tentro CNR SACSO, c/o Chemistry Department, University of Rome, 'La Sapienza', P.Ee A. Moro 5, 1-00185 Rome, Italy 'IMAI, CNR-Area della Ricerca di Roma, PO Box 10, I-0001 6 Monterotondo Stazione, Italy Manganese oxide-doped zirconium oxide samples, prepared by heating mixtures of coprecipitated hydroxides at 1073 K in a hydrogen stream (water content 0.2% by volume), were analysed to obtain information on the solid solution formation. The state and the thermal stability of the incorporated species were also investigated. The samples (manganese content up to 14.74 mass%), were studied 'as-prepared' and after subsequent thermal treatments in oxygen up to 753 K.The results of several techniques [X-ray diffraction (XRD), Raman spectroscopy, thermogravimetry (TG) and magnetic susceptibility measurements] show that in the 'as-prepared' samples (1073 K, H2) a high fraction of manganese is incorporated in the zirconia structure, only a small fraction being present as an MnO separate phase. Most of the manganese in solid solution is present in the +2 oxidation state, the remainder as +3 and +4.TG experiments and magnetic susceptibility measurements reveal that the Mn3+ and/or Mn4+ are formed both during the cooling in hydrogen by reaction with water present as an impurity in the gas phase, and during the exposure to the atmosphere.As the amount of manganese in solid solution increases, the volume of the zirconia unit cell slightly decreases. The solid-solution formation favours the tetragonal and the cubic modifications at the expense of the thermodynamically stable monoclinic phase. When the samples are heated up to 753 K in oxygen, the Mn2+ in solid solution is partially oxidized to Mn3+ and/or Mn4+. TG and XRD experiments show that the oxidation starts at low temperature and takes place in solid solution without appreciable manganese oxide segregation. Zirconia solid solutions containing metal ions have been widely investigated because of their scientific and technological inter- est. Usually high temperatures are required for their prep- aration, a procedure that results in highly sintered materials.The preparation of metastable phases by low-temperature processing techniques has been developed, leading to materials with large surface areas and of potential interest as hetero- geneous On the other hand, manganese oxides have long been known as catalytic materials. They have been studied mainly as supported systems, in particular on al~mina,~ while very few papers on manganese oxide dispersed on zirconium oxide have been This system is of great interest, as demonstrated by recent reports on a new class of catalysts, based on superacid sulfated zirconia and containing manganese and iron as promoter^,'^-'^ which are active for hydrocarbon isomerization even at room temperature.Some attention has also been devoted to the manganese oxide- zirconium oxide solid ~olutions.'~-'~ These systems have several points of interest, for example their redox properties since manganese may be present in several oxidation states and the oxygen ions in the zirconia structure are known to have a rather high mobility. Moreover, the detailed investigation of this system may provide information on the effect of the incorporated ions on the host structure, a relevant aspect of zirconia as a material. All these considerations, and specifically the possibility of obtaining information on the conditions of the solid-solution formation, on the extent of dispersion of the incorporated ions and on the features of the oxidic matrix, have prompted us to undertake the present research.The study was carried out mainly by X-ray diffraction (XRD), thermogravimetry (TG), Raman spectroscopy , magnetic susceptibility measurements and chemical analysis. Experimental Sample preparation The samples were prepared by a coprecipitation method by dissolving 9 g of ZrOC12.8H20 (Erba RP) and a given amount of manganese nitrate, Mn(N0,)2-4H20, (Erba RP) in 100 ml of water. Ammonia was added to the stirred solution till the pH was raised to 8.2. The obtained precipitate was separated from the supernatant solution and thoroughly washed with ammonia-containing water to remove C1- ions (negative test in the solid). The solid was then dried at 363 K for 24 h in air. After grinding in an agate mortar, the powder was placed in a platinum vessel and heated at 1073 K for 10 h in a hydrogen stream (from a cylinder, water content cu.0.2% by volume). Such a temperature was high enough to obtain XRD spectra of good quality in the back-reflection region without causing the formation of monoclinic zirconia, the stable phase in the temperature range of preparation. The hydrogen atmosphere was adopted to have manganese in a reduced state. At the end of the thermal treatment the sample was cooled to room temperature, still in a hydrogen atmosphere, and finally exposed to the atmosphere. ZrO, was obtained by a similar procedure except for the addition of manganese nitrate. The samples were designated ZMnx where x stands for the Mn content (in mass%).The samples were studied (i) 'as-prepared', (ii) after rinsing with a HC1 solution and drying under vacuum at room temperature, and (iii) as in (ii) but with subsequent thermal treatments at a given temperature in oxygen for 5 h. The rinsing was carried out by mixing a fraction of each ZMnx sample at room temperature for 12 h with a concentrated HCl solution under stirring. It was checked that such a treatment was able to dissolve MnO, Mn203, Mn304 and MnO,, but did not affect ZrO,. The solid was then separated from the liquid fraction by centrifugation, thoroughly washed in water, and dried at room temperature under vacuum. The liquid fractions were collected and analysed for the manganese content (see later).To clarify the state of manganese in samples heated in hydrogen, the sample ZMnll.50, after rinsing with HC1 and heated at 753 K for 5 h in oxygen, was also studied after a further thermal treatment at 1073 K for 0.5 h in a hydrogen flow. For this treatment a silica reactor connected to a vacuum line and having a lateral tube suitable for the magnetic J. Muter. Chem., 1996, 6(3), 403-408 403 Table 1 Analytical data and ZrOz unit-cell constants for manganese-containing zirconia samples heated at 1073 K for 10 h in a hydrogen stream ZMn4 18 0 10 4 09 5 090 1004 m-ZrO,, t-ZrO, ZMn6 70 0 27 6 50 5 094 1002 m-ZrO,, t-ZrO, ZMn7 79 0 32 7 64 5 093 1000 c-ZrO, ZMnll50 100 1064 5 084 1000 c-ZrO, MnO ZMnl4 74 2 50 12 57 5 085 1000 c-ZrO,, MnO For designation of samples see text Amount of Mn remaining in solid residue after HCI attack ' m=monochnic, t=tetragonal, c=cubic measurement was used At the end of the heating, the sample was cooled in dry hydrogen (liquid-nitrogen trap) After evacu- ation, the sample was transferred to the lateral tube, sealed off under vacuum and submitted to the magnetic measurement without exposure to the atmosphere Chemical analysis The manganese content was determined by atomic absorption spectroscopy (AAS) A known amount of each ZMnx sample was dissolved in 2 5 ml of concentrated (40%) HF solution After dilution, the solution was analysed using manganese standard solutions containing ZrO, dissolved in HF in a concentration similar to that of the sample The amount of manganese rinsed away with HC1 [Mnrlns (%)I was similarly determined using standard solutions with appropriate HC1 concentrations The analytical results are collected in Table 1 X-Ray diffraction measurements A Philips diffractometer (PW 1725) and Cu-Ka (Ni-filtered) radiation were employed The 28 angular region from 10 to 150" was scanned To determine changes in the unit-cell parameters of tetragonal or cubic zirconia, the following nine reflections were used (620), 28= 146", (006), 125", (153), 124", (044), 116", (511), 103", (422), 95", (042), 84", (133), 81", (400), 74" The graphical method of extrapolation against the Nelson and Riley function was adopted For each sample the position of the diffraction peak was measured several times and showed an angular uncertainty of +_O 05" The axial ratio c/a= C was determined by minimizing the deviations of the values of the zirconia unit-cell parameter a following a least-squares pro- cedure At least two determinations were made for each specimeq It was observed that the errors in a and C are +O 002 A and &O 001, respectively This means th%t the unit- cell parameter c could be measuredo to +O 007 A and the zirconia unit-cell volume, V,to k0 3 A3 Raman spectroscopy Raman spectra were obtained in back-scattering geometry from self-supporting disks using a SPEX 1877 TRIPLEMATE spectrograph equipped with a liquid-nitrogen cooled EG & G PAR CCD multichannel detector and an Ar ion-laser source (line 514 5 nm) The laser power at the sample position was 5 mW to avoid sample modification due to irradiation The experimental Raman spectra were treated by GRAMS/386 software (by GALACTIC) in order to remove the baseline and to fit the curve Magnetic susceptibility determination The magnetic determinations were carned out by the Gouy method at 4000, 6000 and 8000G in the temperature range 98-298 K A Mettler balance reading to &O 01 mg was employed The instrument was calibrated with Hg[Co(CNS),] l9 The magnetic susceptibility of ZrO,, treated as the manganese-containing samples, was determined with the same apparatus 404 J Muter Chem , 1996,6( 3), 403-408 Thermogravimetry The thermogravimetric apparatus consists of an electrobalance (Cahn RG) reading to 0 01 mg connected to a vacuum line to perform experiments in controlled atmosphere During the experiment the sample was kept in a constant flow (20 ml min-l) of gas (oxygen, nitrogen or hydrogen) The temperature was changed by a linear programmer at a rate of 2 K min-' XPS measurements X-Ray photoelectron (XP) spectra were obtained with a Leybold-Hereaus LHS 10 spectrometer interfaced with a Hewlett-Packard 2113 B computer, using Al-Ka (1486 6 eV) radiation The sample, in the form of a fine powder, was pressed onto a tantalum plate attached to the sample rod The spectra were recorded in the sequence Mn 2p, 0 Is, C Is, Zr 3d Zr 3d,,, binding energy (EB)at 1825eV was taken as reference Results X-Ray diffraction Fig 1 shows the XRD patterns of ZMnx samples, and the results of the phase analysis are given in Table 1 Undoped ZrO, showed essentially the monoclinic form The addition of manganese caused a sharp decrease in the fraction of the monoclinic modification and a parallel increase in the fraction of the higher-symmetry polymorphs (tetragonal and/or cubic) The XRD patterns of the two more copcentrated samples shoy the most intense MnO lines (d=2 57 A, 28 = 34 91 ",d =2 22 A, 28=40 63", Fig 1) These lines are no longer observed on the diffraction patterns of the same samples submitted to the HCl treatment (not shown) The zirconia unit-cell parameter a and the axial ratio C =c/a, are also given in Table 1 Their values were not affected by the HCl rinsing The effect of the manganese addition on 10 I5 20 25 30 35 40 45 50 55 60 2Ndegrees Fig.1 XRD patterns for 'as-prepared' ZMnx samples a, ZrO, (monoclinic modification), b, ZMn4 18, c, ZMn7 79, d, ZMnll 50, e, ZMnl4 74 *, MnO reflections 5.120 \ \ 5.110 2 5.100 Ip,P Fig. 2 Zirconia unit-cell constants for ZMnx samples as a function of the manganese content in the solid residue. X, a; 0,c. the zirconia unit-cell constants is shown more clearly in Fig. 2 where a and c are plotted against the manganese content (atom%) in the solid residue. The latter quantity was obtained from Mnsol res(atom%)= 100{A/[A+( 100-70.94A)/123.22]} 129.0 " 200 300 400 5IIo 600 700 800 900 TIK Fig. 4 Zirconia unit-cell volume, V, for ZMnx samples rinsed with HCl and heated at the specified temperatures for 2 h, in oxygen.A, ZMn7.79; 0,ZMnll.50. Table 2 Wavenumbers (in cm-') of the bands in the Raman spectra for manganese-doped zirconium oxide samples (for designation of samples see text) ZMn7.79 ZMn4.18 ZMnll.50 ZMn6.70 ZMn14.30 m-Zr0220,21 t-Zr0222 c-Z~O,,~ where A =,,,(Mnsol %)/54.94, and 54.94, 70.94 and 123.22 are 200w 192vsthe Mn, MnO and Zr02 molar masses, respectively. As Fig. 2 268s 264w 272s 275wshows, while a exhibits only a small contraction, c is found to 224w 309wdecrease strongly up to a manganese content of about 16 atom% where it reaches the value of a. For higher manganese 325m 319m content an axial ratio c/a= 1 is observed.The variation of the zirconia unit-cell volume, V,is shown in Fig. 3. The volume V exhibits a small but significant contraction, mostly due to the decrease in c. The zirconia unit-cell volume, V, for ZMnx samples heated at 753 K for 5 h in a oxygen stream is also plotted in Fig. 3. The heating in oxygen causes an additional contraction of V. For comparison, data from ref. 17 are also shown. The change of V as function of heating at a given temperature (2 h, in oxygen) is shown in Fig. 4 for the samples ZMn7.79 and ZMnl1.50 rinsed with HCl, as typical examples. For thermal treatments up to 673 K a shrinkage of I/ was found. For higher temperatures (up to 823 K, the highest temperature studied), V remains constant. No effect on the axial ratio c/a was observed, and no extra lines, due to formation of separate phases, were found. Raman spectroscopy The frequencies of the bands present in the Raman spectra are collected in Table 2 and compared with data from literature relating to monoclinic,20*21 tetragona122 and zirconia.The Raman analysis suggests that the samples may be divided into two sets according to the manganese content. Sample ZMn4.18 and ZMn6.70 showed seven bands while ZMn7.79, ZMnll.50 and ZMn14.30 have simpler spectra characterized 133.0 132.0 130.0 335m 355w 349w 380w 383m 440w 420w 463m 464m 476s 503w 539w 565m 561w 597m 590m 604w 617mw 639m 644s 647s Peak intensity: vs, very strong; s, strong; m, medium; mw, medium weak; w, weak.by broad and weak bands. As shown in Table2, the spectra of samples of the first group match the Raman bands of tetragonal zirconia,22 those of the second one show the Raman spectrum typical of cubic Zr02.23 Magnetic measurements The specific magnetic susceptibility values measured for Zr02, xgZr02,were positive with a small temperature dependence, but were independent of the treatment to which the zirconia was submitted. At 120 and at 292K the measured values were 6 x and 4 x erg GP2g-l, respectively. The specific magnetic susceptibility for undoped Zr02 is reported to be -1.1 x erg G-2 g-1.24 The positive values measured in our case may be accounted for in terms of small deviations from stoichiometry and/or in terms of paramagnetic contami- nations.The second hypothesis is more probable. In fact electron paramagnetic resonance (EPR) spectra showed a line +at g = 4.2 characteristic of Fe3 .25 However, other paramag- 129.0 t+'+,c netic species not detectable by EPR (e.g. Fe2+) cannot be excluded. The molar susceptibility of manganese, xMn,was obtained by subtracting the measured contribution of the matrix from the specific susceptibility of the sample, xgsample, using: Fig. 3 Zirconia unit-cell volume, V, as a function of the manganese XMn = 54.94{[XTrnPle-Zr02(%)x xgzro2]/Mn("/~)}content in the solid residue. 0, ZMnx samples; 0, ZMnx samples rinsed with HC1 and heated at 753 K, for 5 h in oxygen; 0,from ref.17. where 54.94 is the Mn molar mass and Zr02(%) and Mn(%) J. Muter. Chem., 1996, 6(3), 403-408 405 are the percentages of ZrO, and of manganese, respectively The molar susceptibility of manganese was found to be field- independent and to obey the Curie-Weiss law x=C/(T+@), where C is the Curie constant and 0 the Weiss temperature The constant C is related to the effective magnetic moment, Peff > by Peff =(8CP2 Table 3 shows the results of the magnetic measurements for samples before and after rinsing with HCl and for samples heated in oxygen at 753 K after HCl treatment For each sample the rinsing with HC1 had virtually no effect on the magnetic moment, while the heating in oxygen caused a decrease in this value The values of 0 were low and were not affected appreciably by the treatments For the sample ZMnll 50 reheated in hydrogen at 1073 K in the silica reactor, then cooled to room temperature in dry hydrogen and submit- ted to the magnetic determinations without exposure to the atmosphere, the values of peffand 0 were 5 93 pB (pB=Bohr magneton =9 274 x J T-'), and -30 K, respectively Thermogravimetry and XPS measurements Fig 5 shows the results of the thermogravimetry experiment for the ZMnll 50 sample, rinsed with HCl and heated in oxygen at 753 K for 5 h, taken as a typical example No change in mass was observed for ZrO, submitted to a similar thermal 00 05 10 i15 O2 02 Fig.5 TG curves for the ZMnll 50 sample, rinsed with HC1 heated at 753 K for 5 h, in oxygen (a) Heating and cooling in H, containing 0 2% water, followed by heating in N,, (b)heating in H, (2% water) and cooling in dry H, (liquid-nitrogen trap), (c) heating and cooling in 0, subsequent to (b) B, mass in dry H, treatment Fig 5(u) refers to the experiment in a hydrogen atmosphere, a decrease in mass was observed up to 873 K but for higher temperatures no mass variation was found The sample was then cooled to room temperature in the same apparatus and an increase in mass from 473 to 373 K was observed To test whether this behaviour was due to adsorption of water present in the gas, a subsequent heating in a nitrogen stream was performed no change in mass was observed up to 460 K The thermogravimetric behaviour of a second portion of the same sample is shown in Fig 5(b) The experimental conditions were similar to those for Fig 5(u), the only exception being the cooling carried out in a dry hydrogen stream (liquid- nitrogen trap) no mass increase was observed in the cooling range Fig 5(c) shows a thermogravimetry experiment in oxygen subsequent to that just described Note that an increase in mass was recorded at room temperature by changing the atmosphere from dry hydrogen to oxygen After equilibration, the temperature was raised to 753 K The mass increased, reaching a maximum at ca 573 K From 573 to 753 K the mass decreased and remained constant during the heating at this temperature for 5 h In the subsequent cooling to room temperature an increase in mass was found For some ZMn samples rinsed with HCl and heated in oxygen at 753 K for 5 h, Table 4 collects the mass variations in the thermogravimetry experiments in hydrogen (from room temperature to 1073 K), A%", An XPS analysis performed on sample ZMnl4 74, heated in hydrogen at 1073 K, rinsed with HCl, dried under vacuum and exposed to the atmosphere at room temperature, detects only the signal of Mn4+, the Mn 2p3,, EB value being 642 2t-0 2 eV 26 Discussion Samples prepared in H, The treatment in a hydrogen stream at 1073 K causes the incorporation of most of the manganese in the zirconia struc- ture The small fraction not in solid solution is present essen- tially as MnO The separate phase, directly revealed by XRD in the two more concentrated samples (Table 1)was quantitat- ively evaluated by rinsing the samples with HC1 The solid- solution formation is deduced from the variation of the zirconia unit-cell volume, V,as the manganese content increases (Fig 3) The incorporation of a high fraction of manganese into the zirconia structure explains the low values of the Weiss tempera- ture, @ (Table 3) In fact, the Mn2+-0-Mn2+ superexchange interactions, responsible for the an tiferromagne tic properties of MnO (0 610K),27 are markedly decreased in the solid solution in diamagnetic ZrO, The manganese incorporation opposes the formation of the otherwise thermodynamically stable monoclinic phase of zir- conia, and favours the tetragonal and the cubic modifications (Fig 1) As the manganese content increases cubic zirconia is formed at the expense of the tetragonal phase In fact, as Fig 2 Table 3 Magnetic data for manganese-containing zirconia samples subjected to different treatments treatments ZMn6 70 -7 5 78 6 50 -5 4 06 5 70 6 45 -5 4 94 ZMn7 79 -b -b 7 64 -10 3 89 5 58 7 58 -10 4 95 ZMnll50 -22 5 70 10 64 -20 4 03 5 68 10 57 -20 4 73 ZMnl4 74 -43 5 52 12 57 -35 408 5 71 12 41 -30 4 89 From Mnsol taking into account the mass increase during the heating in oxygen up to 1073 K (from TG) Not determined 406 J Muter Chem , 1996, 6(3), 403-408 Table 4 Composition of ZMn samples after HCl rinsing and heating at 753 K for 5 h in oxygen samples Co,/erg G-' mol-' K ZMn6.70 6.45 3.05 ZMnll.50 10.57 2.80 ZMn14.30 12.41 2.99 shows, starting from ca.16% atomic (ZMn7.79) the value of the zirconia unit-cell constant c equals that of the a parameter, whereas for lower manganese concentrations c is higher than a. Srinivasan et a[.,' have recently questioned the assignment of the cubic and tetragonal zirconia structures from XRD analyses performed on a restricted 28 range, since the two polymorphs, having similar lattice parameters, exhibit very similar front-reflection XRD patterns. In the present work, however, a firm assignment was reached since it was carried out by extending the analysis to the back-reflection region, and it was confirmed by Raman spectroscopy. The application of the latter technique is of particular help since tefragonal2, and monoclinic20.21 ZrO, have different spectra.As Table 2 shows, the Raman spectra of ZMn4.18 and ZMn6.70 match that of the tetragonal zirconia, while for samples with higher manganese content, starting from ZMn7.79 (manganese content = 15.97 atom%), the spectrum of the cubic modification is found. The stabilization of tetragonal and cubic structures by aliovalent ions was recently It was suggested that the aliovalent dopants stabilize the tetragonal and cubic structures because their effective charge alters the interatomic force constant. From magnetic measurements (Table 3) we conclude that in the 'as-prepared' samples most of the manganese is present as Mn2+, while a small fraction is in higher oxidized states (Mn3+ and/or Mn4+).In fact, for these samples the values of the magnetic moment, peff, are lower than 5.92 pB, the value expected for Mn2+. Moreover, since after rinsing with HCl, the magnetic moment remains lower than that expected for Mn2+, the oxidized manganese ions (Mn3+ and/or Mn4+) must be in solid solution. Their formation occurs during the cooling to room temperature at the end of the thermal treat- ment in a hydrogen stream and/or during exposure of the sample to air at room temperature. The presence of Mn4+ on the upper layers of these samples was indeed detected by XPS. On the other hand, direct evidence of the oxidation during the cooling in hydrogen is obtained from thermogravimetric experiments (Fig. 5). An increase in mass is observed from 473 to 373 K when the sample is cooled in the thermogravimetric apparatus [Fig.5(a)].This phenomenon cannot be ascribed to adsorption of water present in the gas stream since no loss in mass is recorded in the subsequent heating in nitrogen. Indeed, the mass increase is a consequence of the Mn2+ oxidation to Mn3+ and/or to Mn4+ according to the reactions: Zr, -,MnX2+0,-, +-Y 2 H20+ Zr, -xMnx-y2+Mny3+0, +-Y H,-x+y,2 2 Zr, -xMnx2+0,-+yH,O + Zr, -xMnx-y2+Mn,4+0,-+ +yH, as demonstrated by the fact that no increase in mass is observed if water is removed in the cooling step [Fig. 5(b)]. The value of the magnetic moment (5.93 pB)measured for the ZMnll.50 sample heated at 1073 K in hydrogen in a silica reactor, cooled in dry hydrogen and submitted to the magnetic measurement without exposure to air, confirms that if water is absent the thermal treatment in hydrogen is able to keep all the manganese in the +2 oxidation state.On the other hand, Fig. 5(c)shows that when this sample is exposed to air at room temperature, an increase in mass amounting to 0.41% is A%Hz Mn2+(%) Mn3+(%) ~n~+(%) 0.96 1.8 2.7 1.9 1.86 1.2 5.9 3.5 1.92 2.7 6.2 3.5 observed. Since no mass increase was found for undoped zirconia and the ZMnll.50 sample (cooled in dry hydrogen) contains solely Mn2+, the increase in mass corresponds to the oxidation of a limited amount of the latter ions. Therefore, an estimation of the solid-solution composition may be carried out, indicating with x,y,z the amount of Mn2+, Mn3+ and Mn4+ present in 100 g of sample after the air contact: x+y +z =Mnsolres (YO) (1) 16/54.94 (0.5y+~)=A% (2) 4.38~+3y + 1.892=CMn,,, res(Y~) (3) where 16 and 54.94 are the oxygen and manganese molar masses, respectively, C (=4.03) is the Curie constant for the sample ZMnll.50 rinsed with HC1 (Table 3), and 4.38, 3 and 1.89 are the Curie constant expected for Mn2+, Mn3+ and Mn4+, respectively.The solution of the system gives x= 8.4, y= 1.6 and z=O.6. This result points out that in our preparation conditions a large fraction of the incorporated manganese (almost 80%) remains in the +2 oxidation state. Since the other ZMnx samples exhibit values of the Curie constant similar to that found for ZMnll.50 (Table 3), the presence of a large fraction of Mn2+ in solid solution must be expected for them also.Therefore, on the !asis of the ionic radii of the involved spec@ (YZr4+(CN 8,=O.84 4, TMnZ+(CN ,,=0.96 A, YMn3+(CN 6)=0.645 A, rMn4+(CN 6)=0.53 A; CN =coordination number)30 an expan-sion of the zirconia unit-cell volume should be foreseen since Mn2+ is markedly present, and larger values for the ionic radii of Mn3+ and Mn4+ should be taken when eight-coordination is considered. In fact, for the transition-metal ions of the first series in the usual oxidation states, an increase in the ionic radius of around 15% is found passing from six- to eight- co~rdination.~~In contrast, Fig. 2 shows that a small but significant decrease in I/ is measured.This result shows that the incorporation of aliovalent ions into the zirconia structure cannot be described on the basis of a simple ionic picture. In fact, the incorporation of ions with a positive charge lower than +4 takes place with formation of anionic vacancies. Their number and their extent of association control the coordination number of cations and thus the value of their ionic radii. Therefore for a correlation between the zirconia unit-cell volume and the solid-solution composition it is neces- sary to have a deeper insight into the cation local structures and their variation with the solute concentration. It was recently sho~n,~',~~ that Vegard's law may be applied to the zirconia solid solution by considering an ion-packing model based on defect clusters.By extrapolating data concerning zirconium oxide solid solutions of several metal oxides, taken from the literature, to zero dopant content it was shown that the ynit-cell volume for undoped cubic zirconia is I/= 133.1 A3.33 This value is in goo{ agreement with the value extrapolated in our case, V= 133.2A3 (Fig. 3). Samples heated at 480 "Cin oxygen for 5 h As already shown, in the rinsed samples most of the manganese is present as Mn2+ and only a small fraction is present in higher oxidation states. When these samples are heated in oxygen, the fraction of oxidized manganese increases as demon- strated by the thermogravimetric behaviour exemplified in Fig. 5(c) for the ZMn11.50 sample.The small decrease in mass J. Mater. Chew., 1996, 6(3), 403-408 407 in the range 573-753 K suggests that for temperatures higher than ca 573 K, Mn4+ (formed during the heating and/or present at room temperature) tends to transform to Mn3+, according to the chemistry of manganese oxides 33 in particular, MnO, is reported to decompose at higher temperatures giving Mn203 at ca 873 K We note that in the present work the transformation takes place at a lower temperature This may be related to the higher mobility of oxygen ions in the zirconia structure When kept at constant temperature the sample composition remains fixed, at least on a 5 h timescale [Fig 5(c)], while an oxidation occurs when the sample is cooled to room temperature Therefore the composition at room temperature is different from that at 753 K As Table 3 shows, magnetic moment values close to 4 9 pB are measured for samples after the heating in oxygen, thus confirming that the fraction of oxidized manganese has increased On the other hand, the redox process takes place in solid solution In fact, while no separate phase is observed on the XRD patterns of these samples, Fig 4 shows that the zirconia unit-cell volume V decreases as the temperature of the heating in oxygen increases As already noted, it is not possible to correlate quantitatively V to the composition, nevertheless the volume contraction is a clear indication that Mn2+ is converted to ionic species of smaller ionic radius (Mn3+ and/or Mn4+) Fig 4 also shows that for samples heated in the range from ca 600 to 823 K, V (measured at room temperature) is constant This behaviour indicates that the sample composition at room temperature does not depend significantly on the actual tem- perature of the thermal treatment We believe that this is a consequence of the oxidation taking place during the cooling from the higher temperature range [Fig 5(c)] The composition of the samples may be evaluated by combining magnetic, thermogravimetry and analytical data Since the manganese is completely reduced to Mn2+ by heating in hydrogen at 1000 K, the mass variation in the thermogravi- metry experiment in hydrogen up to this temperature is related to the conversion to Mn2+ of the oxidized manganese (Mn3+ and Mn4+) present in the sample at room temperature Therefore, indicating with x,y,z the amount of Mn2+, Mn3+ and Mn4+ in 100 g of sample after the heating at 753 K for 5 h in oxygen, eqns (l), (2) and (3) may be considered, substituting A% with A%H2(the mass change in the thermo- gravimetnc experiment in hydrogen) in eqn (2), and C with Co2 (the experimental Curie constant for samples heated in oxygen at 753 K) in eqn (3) The results are listed in Table 4 The simultaneous presence of Mn2+, Mn3+ and Mn4+ is apparent, in spite of the experimental errors in the various determinations In this context, note the results of a recent study on the manganese oxide-zirconium oxide system l7 The samples, pre- pared by heating in air a mixture of co-precipitated zirconium and manganese hydroxides in the range 773-1073K were studied, by XRD among other techniques The values of the zirconia unit-cell volume as a function of the manganese content are shown in Fig 3 and agree with our data, but in ref 17 the incorporation of manganese, essentially as Mn4+, was claimed Conclusions By heating in hydrogen at 1073K, manganese is incorporated into the zirconia structure as Mn2+ However, the presence of water in the gas stream during the cooling step and the exposure of the sample to the atmosphere (at room tempera- ture) cause the oxidation of a limited fraction of Mn2+ to Mn3+ and/or Mn4+ The incorporation of the foreign species in the solid solution favours the tetragonal modification of zirconia, which becomes cubic for a manganese content higher than ca 16 atom% In spite of the value of the Mn2+ ionic 408 J Muter Chew, 1996, 6(3), 403-408 radius [rM,,Z+(CN 8)=0 96 h]being larger than that of the host cation [rZr4+(CN 8) =0 84 A], the incorporation of manganese causes a decrease of the zirconia unit-cell volume This phenom- enon indicates that the system cannot be treated on a simple ionic basis Heating in oxygen up to 753 K induces an oxidation of Mn2+, partially transformed to Mn3+ and/or Mn4+ The oxidation takes place in solid solution and causes a contraction of the zirconia unit-cell volume On the basis of magnetic susceptibility determinations, thermogravimetry and analytical data, the composition of the oxidized samples was deduced Financial support from the Italian MURST (Finanziamenti progetti di ricerca di interesse nazionale-quota 40%) is gratefully acknowledged References 1 M Valigi and D Gazzoli, Reactivity of Solids, ed P Barret and L-C Dufour, Elsevier, Amsterdam, 1985, p 1081 2 P D L Mercera, J G Van Ommen, E B M Doesburg, A J Burggraaf and J R H Ross, Appl Catal, 1990,57,127 3 P D L Mercera, J G Van Ommen, E B M Doesburg, A J Burggraaf and J R H Ross, Appl Catal, 1991,71,363 4 A Cimino, D Cordischi, S De Rossi, G Ferraris, D Gazzoli, V Indovina, G Minelli, M Occhiuzzi and M Valigi, J Catal, 1991,127,744 5 F Kapteijn, A D van Langeveld, J A Moulijn, A Andreini, M A Vuurman, A M Turek, J-M Jehng and I E Wachs, J Catal, 1994, 150,94 and references therein 6 M Roper, W Keim, J Seibring and G Kolle-Gorgen, Eur Pat, 208 102,1987 7 W Keim and W Falter, Catal Lett, 1989,3, 59 8 D K Koh, J S Chung, Y G Kim, J S Lee, I-S Nam and S H Moon, J Catal, 1992,138,630 9 T Ishihara, N Horiuchi, K Eguchi and H Arai, Appl Catal, 1990,66,267 10 T-C Cheung, J L DItri and B C Gates, J Catal, 1995, 151,464 11 V Adeeva, G D Lei and W M H Sachtler, Appl Catal A, 1994, 118, L11 12 C-H Hsu, C R Heimbuch, C T Armes and B C Gates, J Chem SOC Chem Commun, 1992,1645 13 C-H Lin and C-Y Hsu, J Chem SOC Chem Commun ,1992,1479 14 G Bacquet, J Dugas, C Escribe and A Rouanet, J Solid State Chem, 1976,19,251 15 H Nishizawa, T Tani and K Matsuoka, J Muter Scz, 1984, 19,2921 16 I Voigt and A Feltz, Solid State Ionzcs, 1993,63-65, 31 17 A Keshavaraja and A V Ramaswamy, J Muter Res ,1994,9,837 18 V P Dravid,V Ravikumar, M R NotisandC E Lyman,J Am Ceram SOC , 1994,77,2758 19 B N Figgis and R S Nyholm, J Chem SOC, 1958,4190 20 V G Keramidas and W B White, J Am Ceram SOC , 1974,57,22 21 C H Perry, F Lu, D W Liu and B Alzyab, J Raman Spectrosc, 1990,21,577 22 D P C Thackeray, Spectrosc Acta, 1974,301,549 23 V G Keramidas and W B White, J Phys Chem Solids, 1973, 34,1873 24 P W Selwood, Magnetochemistry, Interscience New York, 2nd edn, 1956, p 78 25 K E Swider and W L Worrel, J Am Ceram SOC, 1995,78,961 26 J S Ford, R B Jacmman and G C Allen, Phzlos Mag A, 1984, 49,657 27 H Bizette, C F Squire and B Tsai, C R Acad Sci ,1984,207,449 28 R Srinivasan, S F Simpson, J M Harris and B H Davis, J Muter Scz Lett, 1991,10,352 29 A N Cormack and S C Parker, J Am Ceram SOC, 1990, 73, 3220 30 R D Shannon, Acta Crystallogr Sect A, 1976,32,751 31 M Yashima, N Ishizawa and M Yoshimura, J Am Ceram SOC , 1992,75,1541 32 M Yashima, N Ishizawa and M Yoshimura, J Am Ceram SOC, 1992,75,1550 33 R D W Kemmit, in Comprehensive Inorganic Chemistry, ed J C Bailar, H J Emeleus, R Nayolm and A F Trotman-Dickenson, Pergamon, Oxford, 1993, vol 5, p 802 Paper 5/05210F, Received 4th August, 1995

ISSN:0959-9428    

DOI:10.1039/JM9960600403

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Gas-sensitive resistors: surface interaction of chlorine with semiconducting oxides Darryl H. Dawson and David E. Williams* Department of Chemistry, University College London, 20 Gordon Street, London, UK WC1H OAJ The electrical response of WO, and FeNbO, to the presence of small concentrations of C1, in air, in the presence and absence of water vapour, is reported. The response of WO, was very large and fast; that of FeNbO, was smaller and slower. The electrical conductivity of the niobate could be remarkably decreased by making it slightly niobium-deficient, but this had little effect on the chlorine response. A simple, surface-trap-limited model describes reasonably well the variation of electrical resistance with gas concentration. A competitive, dissociative chemisorption of chlorine and oxygen to form the surface-trap states Oa&-and clad, -along with OH,,,- (from water vapour) is postulated and the enthalpy of displacement of surface oxygen by chlorine derived from the temperature dependence of response: AHads=(-152& 28) kJ mol-' on FeNbO,.The different behaviour of the different materials is rationalised in terms of differing trap-state energies for Oads-, clad,- and OH,,,-. At elevated temperature, semiconducting oxides show electrical conductivity changes in response to small concentrations of reactive gases in air. A study of a wide range of materials' showed that this is a phenomenon common to many oxides, with a general pattern of response behaviour1.2 indicating the mediation of a common surface species, presumed to be dissociatively adsorbed oxygen acting as an electron Combustible gases are presumed to cause a response by reacting with the surface oxygen species and thereby changing the surface-trap density.In the cases of highly reactive gases such as chlorine and nitrogen dioxide, one might expect the model of responses mediated through a single surface oxygen species to fail, and that a variety of reactions and surface species might be involved. There is, indeed, the expectation that oxides could react, leading perhaps to low-melting-point products and consequent decomposition of the material. What is perhaps surprising is that many oxides show a conductance change at elevated temperature in the presence of chlorine or nitrogen dioxide which is rapid on introduction of the gas into the air, reversible on removal of the gas and is, for some materials, extremely large in the presence of very small (ppm level) concentrations of the gas.' For WO,,, for example, the response to NO2 is approximately 10, times that to the same concen- tration of combustible gases; to NO, the response is about 20 times that to combustible gases.Two ideas can be developed for the response. First, taking the idea that chlorine, in particular, will react with metal oxides to form chlorides, the gas might react with or displace lattice oxygen at the interface, or might sorb onto surface oxygen vacancies [eqn. (l)]: C1, + 2OOx + 2n' (2) NO, + 0,' + (NO,),' + n' (3) These oxidising gases give a conductance decrease on n-type oxides, so eqns.(2) and (3) predict the wrong sign of the response. Alternatively, these strongly oxidising gases can be considered to adsorb competitively with oxygen and provide a trap state at a different energy with respect to the band edges of the material. This latter analysis has been applied by Gopel et al.576 to the effect at elevated temperature on the conductivity response of SnO, to NO, at low concentration in air. By considering the effect of grain size on the electrical response of WO, to NO,, Tamaki et aL7 inferred an effect of adsorbed NO2 on the depletion layer thickness in the oxide. Tungsten(v1) oxide has previously been investigated chiefly for its response to and H,S.879 The work of Antonik et aL8 further indicated the importance of grain boundaries for the gas response of WO,: the response of rf-sputtered films of thickness 50 nm developed following a heat treatment at 300 "C which caused the films to become textured with crystallite sizes of around 100nm.The effect of surface decoration with pre- cious metals has been explored."." In all of this work, we can remark that the response to gases which might be considered to give a signal as a consequence of reaction with surface oxygen species has been much smaller than that to a gas (NO,) which might be considered to give a response as a consequence of a competitive adsorption with surface oxygen; it therefore seemed reasonable to explore the behaviour with respect to another gas (Cl,) which might be considered to behave similarly to NO,.Here, we report a study contrasting three materials with very different magnitude and rate of electrical response to trace concentrations of chlorine in air: FeNbO,, FeNbO, doped with a small amount of W03, and W03. We show that a simple competitive adsorption/surface trap model rationalises the behaviour of all three materials and offers some insight into the origin of the different behaviours. Experimental Methods Devices for study were prepared on alumina substrates (typi- cally 3 mm square) carrying on one side an interdigitated gold electrode pattern (200 pm line- and gap-width) and on the other a platinum heater track, both fabricated by screen printing.The ceramic powder dispersed in an organic vehicle [cellulose acetate-terpineol for screen printing or poly(viny1 butyral) for tape fabrication] was coated over the electrodes and fired in air to decompose the vehicle, sinter the ceramic and achieve adhesion to the electrodes. The final thickness of the porous ceramic layer was typically 60 pm. Tungsten(v1) oxide (Aldrich) was used without further purification. The niobates were prepared by mixing the oxide powders (Aldrich) by ball-milling in acetone then firing at 1200 "C for 12 h, then grinding. Particle sizes in the final devices were much larger than those quoted by Tamaki et aL7 or by Antonik et ~2.:~ typically 2 pm or more (Fig. 1). The FeNbO, preparation was demonstrated by X-ray diffractometry (XRD) and scanning electron microscopy (SEM) to be a single-phase material.In the tungsten-doped preparation, tungsten did not substitute on the FeNbO, lattice to any detectable degree. Instead, small particles of unreacted WO, were observed dispersed throughout the material. In consequence, the niobate should have been niobium-enriched, and this was confirmed by energy dispersive X-ray analysis J. Muter. Chem., 1996, 6(3),409-414 409 Fig. 1 Microstructure of the W03 preparation (EDXA) in the SEM image. The XRD was unchanged from that of FeNbO,. One difficulty in making comparisons between different works in this field is in making precise comparisons of the electrical behaviour: this is in part because the electrical conductivity is dependent on the microstructure, and also in part because of the effect of trace impurities.An apparent electrical resistivity of WO, is not given by Tamaki et ~l.:,,~the electrical resistance of small devices (mm-scale with mm-scale electrode spacing) was around lo6-lo8 R. An approximate finite-difference calculation was used to obtain the apparent resistivity of the materials used in the present study from the measured resistance of the devices: (p/Q cm) e00.4(R/R). It is very evident that the resistivity of the tungstic oxide used in our study was much less than that used by Tamaki et a[.;' this could in part be attributed to the presence of impurities such as Na [detected in our material by X-ray photoelectron spectroscopy (XPS)] charge-compensated by WIV, giving donor states which would enhance the electrical conduc- tivity.In contrast, the conductivity quoted by Antonik et aL8 was somewhat less (ca. lop3S cm-' at 300°C) than that of the material used here; presumably, the rf-sputtering technique pro- duced a slightly oxygen-deficient material with a conductivity consequently greater than that for a pure, stoichiometric material. A difficulty in the study of sensor responses to gases such as chlorine is that of establishing reliably a given gas concen- tration in the apparatus. We used an all-glass construction, Fig. 2(u), in which the concentrated chlorine stream (50ppm chlorine in 21% 0,-79%N2, BOC Ltd) and the air diluent (laboratory air, filtered, scrubbed of CO, and dried) were fed separately into the base of the apparatus through diametrically opposed ports: this avoided problems with the adsorption of chlorine on the walls of the feed tubes (PTFE).The relative humidity of the gas mixture was controlled by passing a portion of the air stream through distilled water at room temperature and mixing the wet and dry air streams before introducing them into the apparatus. Since the chlorine stream was dry, and the maximum gas concentration studied was 25 ppm, the maximum relative humidity obtainable was 50%. Mass-flow controllers (Tylan) regulated the flow rates of the three gas streams. Experiments always started from the highest concentration, allowing time to saturate the internal surfaces of the chamber with chlorine, then working downwards in concentration.That the gas concentration was reliably estab- lished was checked by checking the consistency of results from the different sensor ports. The chlorine concentration of the cylinder gas was checked by bubbling this through aqueous KI solution for a known time then titrating the iodine pro- 410 J. Muter. Chem., 1996, 6(3),409-414 Port = B14 Cone or SQ28 Fitting Gas Outlets Port 1 Port 4 Port 2 Port 5 Port 3 Port 6 ptfe socket I T 25mm 13cm sensor / 1 air gas out Fig. 2 (a) All-glass apparatus for the study of gas response of multiple self-heated sensor elements (b) Glass cell for response time measurement duced. Measurement of response time used a cell of small internal volume, Fig. 2(b), in which the gas flows were con- trolled manually using needle valves and rotameter tubes with ruby floats (Platon).The temperature of the devices was measured by measuring the resistance of the platinum heater track, which had a repeatable and reliable temperature coefficient of resistance close to that of pure platinum. The temperature was controlled by incorporating the heater into a Wheatstone bridge circuit, using the out-of-balance signal to regulate the power applied to the heater: in this way, the heater was controlled at constant resistance and therefore constant temperature. The device resistance measurement was a two-terminal dc measurement using an auto-ranging digital multimeter (Keithly model 197).This instrument applied a potential difference across the sensor varying from 0 to 2V as the measured resistance progressed from the bottom to the top of any given range. We have noted that the resultant current flow can cause drifts in the measured resistance for some sensor materials, also marked by apparent jumps in resistance when the meter switches range. No such effects were observed for the materials whose behaviour is reported here. Results Fig. 3 shows a typical resistance-time-gas concentration trace for WO,. There are two points: first, the response of the J OD05tl I 1 I 0 2500 5000 7500 10000 12500 IS000 17500 20000 22500 tls Fig. 3 Effect of chlorine on the electrical resistance of WO,; resistance, R us.time, t. Dry gas 395 "C, air baseline resistance 6 x lo3Q chlorine concentrations: (a)25; (b)10; (c) 5; (d)4; (e) 3; (f)2; (g)1 ppm; starting and finishing in dry air. electrical resistance of the material to changes in the gas concentration was large and fast; and secondly, there were some longer-term, irregular and irreproducible variations in response to changing gas concentration which could generally be attributed to the difficulty of establishing reliably a stable, low concentration of chlorine in the apparatus (see Experimental Methods for further discussion). In Fig. 4 is shown the time course of the response to 2 ppm chlorine with the device mounted in a small chamber: the time for response to 50% of the final resistance was less than 10 s, and that for response to 90% of the final resistance was less than 50 s: again, the time-scale to establish a stable gas concentration in the cell was unknown.The response increased strongly with decreasing temperature, and at lower temperature it was greatly decreased by the presence of water vapour in the atmosphere (Fig. 5). The response varied approximately with the square root of the gas concentration at lower and higher temperature, and approximately linearly at intermediate temperatures. This behaviour is very similar to that reported for responses to N02.4For the same gas concentration and temperature, the response (RIR,) to chlorine determined here was approximately 18 times that reported for response to NO,, an observation which was confirmed by measurement of the response of our own devices to NO2.The behaviour of FeNb04 was different in some notable respects. Again, there was a strong response of the electrical resistance to the presence of small concentrations of chlorine (Fig. 6 and 7), although the effect was smaller than that shown by W03. The response was, however, much slower. -SOT) -0" 0 ? I 0 c2oo . ? 1 100 200 300 400 500 600 tIs Fig. 4 Response of W03 to the introduction of chlorine, measured in a small cell. Dry gas, 400°C. 70 60 50 40 30 20 10 0 1 I I 1 1 I 2 4 6 8 10 12 1 I I I I 0 2 4 6 8 10 12 45 40 35 30 25 20 15 10 "y00 1 t I I 1 0 2 4 6 8 10 12 Fig.5 Temperature- and gas concentration (c)-dependence of response of WO, to chlorine. Open symbols: dried gases; filled symbols: 50% relative humidity at 20 "C. (a) 230 "C, line: response cc (b)341 "C, line: response cc c; (c) 516 "C, line: response cc c~'~. J. Muter. Chem., 1996,6(3), 409-414 411 7000 r I 1 6000 + 5000 + 4000 ! 3000 zoo0 1000 I 10 0 10000 20000 30000 40000 soooo tIs Fig. 6 Effect of chlorine on the electrical resistance of FeNbO, at 398 "C 1, dried gases, 11, 50% relative humidity at 20 "C Chlorine concentrations (a)25, (b)10, (c)5, (d) 4, (e) 3, (f) 2, (g)1 ppm, starting and finishing in air 401 I I I I , ,/ I I 1 1 0 2 4 6 8 10 CJPPm Fig. 7 Chlonne concentration dependence of response of niobates, derived from steady-state values of resistance with gas concentration decreasing (400 "C) -, square-root fitted lines, 0, FeNbO,, 0,FeO 9SWO 0SNb04 Comparison with the results for W03 confirms that these were not effects due to the time required to establish a stable gas concentration in the apparatus There were two effects an initial response to about 50% of the final resistance, occurnng within some tens of seconds, and a slow subsequent further increase of resistance The time-scale of this slow process was strongly dependent on the temperature and water vapour pressure, taking from 2h at higher temperature in a dry atmosphere, to more than 5 h at lower temperature or in a wet atmosphere, to reach a stable value (Fig 8) It is also evident from Fig 6 that the time to stabilise the electrical resistance upon decrease of the gas concentration seemed significantly less than that required upon increase of the gas concentration In further contrast to W03, the response was increased by the presence of water vapour Although, as noted in the Experimental Methods, the tung- sten-doped FeNbO, was a two-phase mixture in which the effect of addition of tungsten(v1) oxide seemed to be a slight niobium enrichment of the niobate, there was a noticeable effect on the electrical behaviour (Fig 9) First, the electrical conductivity was significantly decreased, by a factor of approxi- mately lo3 at 400°C with respect to the undoped niobate Since the conductivity of the W03 was significantly higher 412 J Muter Chem, 1996, 6(3), 409-414 12 5000 0 5030 lOOS0 15003 2 00 tls Fig.8 Time-dependence of resistance, R, of FeNb0, following the introduction of chlorine (25 ppm), expressed as a fraction of the resistance change after 5h exposure Device at 0, 398"C, dry air, 0,398 "C, 50% relative humidity in air, V,367 "C, dry air, V,367 O -,r 50% relative 600000 500000 400000 9 JOOO00 p: 200000 80000 90000 1DooW tls Fig. 9 Effect of chlorine on the electrical resistance of niobium-enriched (W0,-admixed) Feo 95Nb0,, us time, t, 384 "C, 50% relative humidity at 20 "C Chlonne concentrations (a)25, (b)10, (c) 5, (d) 4, (e) 3, (f) 2, (8)1 ppm, starting and finishing in air than that of the niobate, this observation could not have arisen simply as a consequence of the presence of the second phase The effect on the response to chlorine was, however, small (Fig 7) The response again showed an initial rapid phase followed by a slow final phase, the fraction of the response in the initial phase seemed somewhat larger than for the undoped niobate (Fig 9) Discussion The grain size of W03 studied here was significantly larger than that in the study of Tamaki et a1' The inference of that work was that the conductivity of materials with grain size larger than ca 40nm was controlled by charge transport across a Schottky barrier at the grain junctions, rather than being surface-trap limited However, as we have previously remarked,3 l2 it can be difficult to distinguish between these two limiting cases by measurement of gas response alone Furthermore, with current transport through a porous body limited to the necks between the grains, depending on the size of the necks, a surface-trap-limited model may provide the proper description of current flow through the necks3 Therefore, here we have chosen, for ease of development, to adopt a simple surface-trap-limited model, finding that it describes reasonably well the variation of electncal resistance with gas response, and that the analysis leads to some insight concerning the interaction of chlorine with the surface of the material.We presume a dissociative chemisorption, and formulate the equilibrium of the gas with sites, S, on the oxide surface: 2s + 0, s 0 + 0 (4) 2s + c1, $ c1+ c1 (5) Reactions (4) and (5) are formulated as a competitive adsorp- tion of chlorine and oxygen: 0 + iC1, =$c1+40, (6) (7) where 8 denotes the fractional coverage over the surface.It is useful to introduce the dimensionless variable: We suppose that the uncharged adsorbed species act as surface electron traps, with energy in the bandgap of the oxide. Charge trapping in the surface states is then represented as the equilibria: 00 --=KOn’0, As has often been stated, in eqns. (9) and (lo), since charge is transported across the interface, the equilibrium constant depends upon the interfacial potential difference, which in turn increases to a limit determined by the potential-energy differ- ence between conduction band edge and surface state, as charge is trapped in the surface states.A consequence is that the surface coverage of the charged surface-trap states depends upon the interfacial capacitance and can be small. Furthermore, the rate of charge trapping should decrease markedly as the interfacial potential difference increases, since surface trapping requires a transfer of charge across the surface potential barrier. An alternative but not exclusive point of view has been advanced by Barsan,’, who proposes a site-limited model in which the chemisorption sites are surface oxygen vacancies; the surface coverage of these is presumed low but they are considered to be completely filled by the charge-trapping species.To develop the ideas, we assume that the surface is completely covered by the uncharged species and that the fractional coverage of charged species is small and fixed by equilibrium with donor states in the bandgap. That is: and 8”-+ eo-= (14) where q is a constant. Then: With the assumption of a surface-trap-limited conductance and q constant, the left-hand side of eqn. (15) is proportional to the resistance, R, of the material, so by considering the limits of low and high chlorine concentration, we have: where Roand R, respectively denote the resistance in the limit of zero and high chlorine concentration, and: =-re1 Kcl KO Further manipulation of eqn.(16) yields: or, if KadsX112 < 1, again from eqn. (16): The data for WO, at lower and higher temperature fit eqn. (20) (Fig. 5);those for FeNb04 fit both eqns. (19) [Fig. lO(a)] and (20) (Fig. 7); those for the niobium-deficient material Fe(Nb,W)04 fit eqn. (19) [Fig. 10(b)]. The assumption which largely determines the form of the response as one proportional to the square root of concentration is that of a competitive adsorption of chlorine and oxygen [eqns. (4)-(6)]. Elaborating the model to include different charge states of adsorbed oxygen, or to change the assumption behind eqn. (13) to one in which the coverage of the uncharged species is small, changes the detail of the derived response, but in some limit the result generally follows eqn.(20). Such analysis illustrates how rela- tively insensitive a diagnostic of the mechanism is the observed response law. We presume that the linear response illustrated in Fig. 5 for WO, at intermediate temperature is a consequence of observation over a rather limited range of concentration. From eqn. (19) and Fig. 10, we obtain the ratio: intercept slope -Kads The increase of baseline resistance of the niobium-enriched niobate meant that the intercept according to eqn. (20) could reliably be distinguished from zero, and its temperature depen- dence determined. Fig. 11 shows the temperature variation of Kad, so derived. The enthalpy of displacement of oxygen by chlorine on the niobium-deficient FeNb04 is thereby calcu- lated: AHad, = (-152& 28) kJ mol-l; displacement of oxygen by chlorine is exothermic.Eqns. (19) and (20) indicated that a discussion of the different magnitude of the response for the different oxides could be framed in terms of the magnitude of Kads, and particularly of Krel. Because there is a resistance increase on exposure to chlorine, Krel> 1, meaning that the state represented as clad,- lies lower in energy than that represented as Oads-. Furthermore, since Krel would vary exponentially with the energy difference between the two surface states, small changes in relative energy could have an extremely large effect upon the sensitivity, rationalising the oxide composition effects on sensitivity reported here.The same argument would rationalise the different sensitivity to NO,. The influence of adsorbed water on the response can be discussed in terms of the relative energy of the surface states represented by OH (from water) and C1: thus we interpret that on W03 the surface trap represented as OHad,- lies lower in energy than that rep- resented as Clads-, so chlorine does not displace hydroxide (from water) from the surface; on the niobate, the interpretation would then be the opposite, that OHad,- lies higher in energy than Clads-, so chlorine does displace hydroxide from the surface. With increasing temperature, the coverage of OHad, -would decrease relative to that of Oads-: the temperature J. Muter. Chem., 1996, 6(3), 409-414 413 ‘W0000 01 02 05 O04 05 06 07 08 ~-I~/pprn-’~ Fig.10 Dependence of resistance, R, on chlorine concentration, c, according to eqn (20), 420 “C, (a) FeNbO,, (b)Fe, 95Nb04 0 0014 0 001 5 0 0016 0 0017 0 0018 T-’IK-’ Fig. 11 Temperature-dependence of Kads determined from Fig 10(b) according to eqn (21) dependence of the effect of water vapour on the response is thereby rationalised Self-evidently, some structural model is needed for these surface states if any further progress is to be made in interpretation Tungsten(w) oxide was insoluble in the mobate, so the effect of the attempted doping was smply to prepare a mobium-ennched FeNb04 Goldschm~dt,~~studying the Nb,05-Fe,03 system showed a sohd-solution range of 33-55 mol% Fe203 We presume that the donor states determining the conductivity of the niobate were Fe2 + charge-compensated by oxygen vacanaes Consideration of the charge balance in the lattice would then mdicate that excess Nb would dmmsh the Fez+ concentration and hence the electncal conductivlty, as was indeed observed A discussion of the factors influencing the response time is necessanly speculative The model of vacancy migration in the electric field present at the interface, proposed by Gopel,’ or of annealing of sub-surface defects by NOz adsorption, pro- posed by Schierbaum et ul to explain a slow dnft of response, would imply that the recovery after exposure to the gas should be equally as slow as the response, or that there would be a semi-permanent baseline shift caused by gas exposure the niobates do not show such an effect and the response of WO, is fast and reversible One possible model for the fast initial and slow final approach to a steady-state response shown by the niobates could be that the total surface charge increases as a consequence of the adsorption of chlorine, q in eqn (14) not being constant Another possible model is that the presumed displacement of OHad,- by Clads- is slow Conclusion A simple surface-trap-limited model describes reasonably well the variation of electrical resistance with gas response, and the analysis leads to some insight concerning the interaction of chlorine with the surface of the oxide The different behaviour of the different oxides can be rationalised in terms of differing trap-state energies for Oads-, clad,- and OHad,-Thus, Oad,-lies higher in energy than clad,- and the energy difference between these states is greater on WO, than on FeNbO, On wo, the surface trap represented as OHad,- appears to lie lower in energy than that represented as Clads-, so chlorine does not displace hydroxide (from water) from the surface, on the niobate, the opposite appears to apply, that OHad,- lies higher in energy than Clads-, so chlorine does displace hydrox- ide from the surface Gas-sensitive resistors based on semi- conducting oxides evidently offer good prospects for the development of low-cost, robust devices sensitive to very low concentrations of chlorine in air we have noted, for example, that the behaviour of the materials studied here did not degrade over time, despite long exposure to polluted air or to atmospheres of high relative humidity This work was supported by the Engineering and Physical Sciences Research Council and Capteur Sensors and Analysers Ltd References 1 P T Moseley, A M Stoneham and D E Williams, in Techniques and Mechanisms in Gas Sensing, ed P T Moseley, J 0 W Norris and D E Williams, Adam Hilger, Bristol, 1991 2 D E Williams and P T Moseley, J Muter Chem ,1991, 1,809 3 D E Williams, in Solid State Gas Sensing, ed P T Moseley and B C Tofield, Adam Hilger, Bristol, 1987, and refs therein 4 M Akiyama, J Tamaki, N Miura and N Yamazoe, Chem Lett, 1991,1611 5 W Gopel, Sens Actuators, 1989,16, 167 6 K D Schierbaum, H D Wiemhofer and W Gopel, Solid State Zonics, 1988,28-30, 1631 7 J Tamaki, Z Zhang, K Fujimori, M Akiyama, T Harada, N Miura and N Yamazoe, J Electrochem SOC ,1994,141,2207 8 M D Antonik, J E Schneider, E L Wittman, K Snow, J F Vetelino and R J Lad, Thin Solid Films, 1995,256,247 9 E P S Barrett, G C Georgades and P A Sermon Sens Actuators B, 1990,1, 116 10 H-M Lin, C-M Hsu, H-Y Yang, P-Y Lee and C-C Yang, Sens Actuators B, 1994,22,63 11 P J Shaver, Appl Phys Lett, 1967,11,255 12 J F McAleer, P T Moseley, J 0 W Norm and D E Williams, J Chem Soc ,Faraday Trans 1,1987,83,1323 13 N Barsan, Sens Actuators B, 1994,17,241 14 H J Goldschmidt, Metallurgia A, 1960,62A, 211 Paper 51059266, Received 7th September, 1995 414 J Muter Chem, 1996,6(3), 409-414

ISSN:0959-9428    

DOI:10.1039/JM9960600409

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

~~ Synthesis of three-dimensional compounds from alkali-metal ion-exchanged y-titanium phosphate Miguel A. Salvad6," Pilar Pertiema,' Santiago Garcia-Granda," Marta Suarez,' Maria Luz Rodriguez,b Ricardo Llavona,' JosC R. Garciab and Julio Rodriguez*b "Departamento de Quimica Fisica y Analitica, Universidad de Oviedo, 33071 Oviedo, Spain 'Departamento de Quimica Organica e Inorghnica, Universidad de Oviedo 33071 Oviedo, Spain Treatment of y-titanium phosphate with 1mol dmP3 solutions of MCl (M =Na, K, Cs) results in layered crystalline phases with compositions at 50 "C of:y-Ti(Nao~5H,.,P04)( P04).+H20.P04)-H20, ?-Ti( KHPO,)( P04)-+H20 and ~-T~(CS~.~H,,,PO~)( Heating at 900 "C gives rise to three-dimensional crystalline compounds. The composition of the new materials depends on the type of alkali-metal present in the solid. The K phase originates KTi2( PO4),, the Cs phase a-CsTi3P,019, and the Na phase produces mixtures of NaTi2( PO4), and TiP207.Moreover, amorphous alkali-metal phosphates which are soluble in acidic media are formed. Samples of y-Ti(KHPO,)(PO,)-$H,O (A), KTi2(P04), (B) and a-CsTi,P,O,, (C)have been studied using X-ray powder diffractjon. The lattice parameters of A [a=5.093( 3) A,b =6.2?7( 1)A,c =11.502( 5) A,p =109.30(4)"], B [a=8.3569(2) A, c =23.0725( 3) A] and C [a =6.320(1) A, b = 14.122( 2) A, c =9.333(2) A, p = 105.82( 2)"] have also been calculated. Titanium and zirconium Nasicon-type phosphates, Analytical procedures AM2(P04)3 (A =Li, Na, K; M =Ti, Zr), have been studied as The analysis of the concentration of phosphorus in the solid fast alkali-metal ion conductors'*2 and low thermal expansion was carried out gravimetrically as magnesium pyrophosphate, ceramics., These compounds are usually obtained by ceramic while the titanium was determined with cupferron." Theor sol-gel methods.Another procedure is the thermal treatment released phosphate groups were measured spectrophoto-of ion-exchanged phases of lamellar phosphates of tetravelent metrically2' using a Perkin-Elmer 200 instrument. The alkali- using this method a multitude of phases can be metals in solution were determined by atomic absorption produced depending upon the type and the amount of metal spectrometry (AAS) using a Varian model spectrAA-300 spec- present between the layer^.^.^ trometer.The solid density was determined using an auto- There are two lamellar phosphates of titanium with pycnometer Micromeritics model 1320. ion-exchange properties,'-'' a-Ti( HPO,),.H,O (a-Tip) and y-Ti( H2P04)( PO4)*2H20(y-Tip). The AM2( PO,), type com- pounds can only be obtained from the half-exchanged phases, Structural characterization procedures a-TiMH( P04)2~nH20 and y-Ti( MHPO,)( P0,).nH20. The samples were ground gently in an agate mortar and side- The half-exchanged phases of alkali-metal ions in a-TiP can loaded in order to minimize preferred orientation. The powder be prepared by treatment of a-TiP with the alkali-metal patterns were collected at room temperature (approx. 25 "C) hydroxide, resulting in partial hydrolysis of the e~changer.'~.'~ on a Philips automated diffractometer.The diffractometerowas Thermal treatment of these compounds gives rise to materials equipped with a copper target (Cu-Ka radiation, 1.5418A), a of the AM2(PO4), type containing TiO, as an impurity. &compensating slit, and a graphite-diffracted beam mono-y-TiP retains alkali-metal ions in acidic media without chromator, and was run in steps of 28=0.02" at 5 s step-' Crystalline phases with different com- in the range 3-80" [y-Ti(KHP0,)(P0,).$H20], 20 s step-'noticeable degradati~n.'~ positions depending on the metal type are obtained by treat- in the range 3-90" [KTi,(PO,),] and 20s stepw1 in the ment of the solid with aqueous solutions of the alkali-metal range 3-110" [a-CsTi3P,Ol9] at 40 kV and 30 mA.The chlorides: ?-Ti( M,.,H,.,PO4)( P04).nH20 (M =Na, Cs)l5,I6 diffractometer was externally calibrated against an Si standard. and y-Ti( KHPO,)( P0,).H,0.17 Thermal treatment of these Indexing of the X-ray powder diffraction patterns was materials was found to produce three-dimensional (3D) performed using the TREOR program2' starting from well compounds. resolved peak diffraction positions. The least-squares refine- ment of the lattice constants was realized with the program LSUCRE,22 using all unambiguously indexed reflections. Refinement of the crystal structure of KTi,(PO,), from powder Experimental data was done using a local version of the DBWS program.23 Electron diffraction patterns and micrographs were recorded Materials in a JEOL 2000 EX-I1 electron microscope equipped with a All chemicals used were of reagent grade.y-TiP was obtained side-entry goniometer stage with tilt angles of 28", operating by using 16.5mol dmP3 H3P04 and a reflux time of 10 days, at 120-200 kV. The samples were dispersed in ethanol by as described previously.18 Portions of y-TiP were equilibrated ultrasonic treatment. Drops of the resultant suspension were with 1mol dm-3 MCl (M=Na, K, Cs) solutions over 2 h. The collected on a perforated carbon film supported on a Cu grid. solids were centrifugally separated. This process was repeated three times. The products were washed with deionized water Results and Discussionuntil all the chloride had been eliminated (test with AgNO,) and dried at 50°C. These solids were then heated at 900°C y-TiP retains lithium ions only when the treatment is with for 24 h.The resultant products were treated with 1mol dm-3 LiOH solutions.24 Since our purpose was to prepare 3D HC1, washed and dried. compounds with high purity the Li phases were not synthesized J. Muter. Chew., 1996, 6(3), 415-419 415 because they are prone to yield amounts of T10, due to the low y-TiP degradation during the ion-exchange process y-TiP treatment with MC1 (M =Na, K, Cs) solutions results in ion-exchanged phases with definite compositions Their subsequent washing until all the chloride had been eliminated results in the production of high-purity 3D phases after the corresponding thermal treatment Potassium compounds The Hf/Kf substitution process in y-TiP occurs in three stages, with crystalline phases of 25, 50 and 100% conversion being formed 25 Over the entire composition range, only two different crystalline phases coexist The first two exchange steps were reached by the addition of (KCl+HCl) solutions The third step was only reached when the half-exchanged phase was in contact with (KCl+KOH) solutions, the fully exchanged phase being obtained The half-exchanged phase dned at 50°C has the y-Ti( KHPO,)( PO,) +H20 composition An X-ray diffraction pattern of the sample heated at 900°C and treated with HCl corresponds to the phase KTI~(PO,)~ The unwashed com-pound exhibits peaks at the same angles but they are broader and of lower intensity According to the data obtained by Clearfield in the study of ion-exchange phases of a-zirconium phosphate,, these results can be explained on the basis of eqn (1) Washing in acidic media gives rise to the potassium metaphosphate solubil- ization26 Despite the fact that this compound was produced as a glass and hence did not show up in the X-ray pattern, its presence distorts the KTi,( lattice y-Ti( KHPO,) (PO,) +H20 ++KTi,( PO,), ++KPO, +H20 (1) As might be expected, the chemical analysis of the washed sample (Table 1) indicates a K Ti P ratio of 1 2 3 In the case of the unwashed sample the ratio is 1 1 2 Therefore, the loss of mass in the treatment at 900°C is associated only with the dehydration process In fact, the expenmental loss of mass (6 38%) concurs with the theoretical value (6 27%) Sodium compounds The H+/Naf substitution in y-TiP also occurs in three steps 27 The first stage was reached by the addition of (NaCl+HCl) solutions The second exchange step was only formed when the phase with 25% conversion was in contact with (NaCl +NaOH) solutions originating the half-exchanged phase When this stage was completed, subsequent additions of (NaC1+NaOH) yielded the fully exchanged phase The 25% substitution phase after washing at 50°C has a composition of y-Ti(Na, ,HI 5P04)(P04)H,O An X-ray pat- tern of the sample heated at 900°C and treated with HC1 exhibits reflections charactenstic of the NaTi,( PO4), and TiP,O, phases The unwashed compounds present the same signals but with broader and of lower intensity, due to the presence of sodium metaphosphate glass These results can be explained on the basis of eqn (2) Table 1 Chemical analysis of matenals treated at 900 "C before and after washing with HCl solution metal M conditions M(%) TI(Yo) P(Y0) K unwashed 14 05 17 87 22 91 K washed 9 13 22 92 22 03 Na unwashed 4 05 21 00 25 53 Na washed 2 66 21 97 26 01 cs unwashed 22 69 16 45 21 24 cs washed 18 03 19 59 21 21 y-Ti(Na, 5H1 5P04)(P04) H20++NaTi,(P0,), ++T1P207+iNaPO, + 1$H20 (2) According to this assumption, chemical analysis (Table 1) of the washed sample indicates a Na Ti P ratio of 1 4 7 In the case of the unwashed sample the ratio is 1 25 47 A mixture of NaTi,(PO,),, T1P207 and NaPO, in the pro- portions of eqn (2) will result in a ratio of 1 2 4 The sodium deficiency must be related to the partial volatility dunng thermal treatment at 900°C This agrees with the fact that the loss of mass undergone during this process (15 22%) is higher than that expected for a simple dehydration process (11 71 YO) A similar effect was observed by Clearfield and Jir~stithipong~ in the thermal decomposition of the sodium ion-exchanged phases in a-zirconium phosphate Caesium compounds In acidic media the ready retention of the Cs+ by y-TiP led to the formation of the 25% substitution phase Higher conver- sions with partial decomposition of the exchanger and precipi- tation of TiOz nH20 were obtained by using (CsCl+CsOH) solutions l6 The 25% phase, stabilized in air at room temperature, has two molpdes of water of hydration and an interlayer distance of 11 6 A Upon treatment at 50 "C it p?rtially dehydrates, its basal space decreases to 11 OA, with the y-Ti(Cso 5H1 ,PO,)(PO,) +H20 compound being formed At temperatures close to 100°C it transforms t? the anhydrous compound with an interlayer distance of 9 9 A X-Ray diffraction patterns of the solid obtained by heating the 25% phase both before and after washing with HC1 solution, are similar but with higher intensities observed after washing The pattern corresponds to the a-CsTi3P5019 phase 28 Chemical analysis of the washed sample presents a Cs Ti P ratio of 1 3 5, confirming the CsTi3P5019 composition The unwashed sample has a ratio 1 2 4 indicating that thermal treatment only results in loss of water According to this, the experimental loss of mass (7 53%) concurs with that expected for both dehydration and condensation processes in the 25% phase (7 14%) This clearly shows that a soluble amorphous caesium compound, Cs,O 2P,05, was formed by the heating process [eqn (3)] Dehydration mechanism The thermal behaviour of the materials containing sodium and potassium is similar to that observed in ion-exchanged phases of a-zirconium phosphate and it can be explained on the basis of eqn (4) The Cs-compounds behaviour is different due to the ion caesium size y-Ti( MXH2-,PO,)( PO,)++xMTi2( PO,), +( 1-x)MP207++xMP03+( 1-+x)H20 (4) A previous paper on the ion-exchange properties of y-T1P14 reported the difficulty of obtaining a Cs half-exchanged phase, in contrast with the ease of obtaining this phase in the case of potassium This can be explained by taking into account that the presence of a potassium ion associated with each dihydro- genphosphate group does not distort the y-layer while the presence of a large caesium ion does produce a distortion These steric factors are clearly exhibited in the thermal behav- lour of the 25% phases at temperatures <25O"C The Cs phase undergoes a dehydration process as can be seen in eqn (5), while the sodium and potassium phases present a different behaviour [eqn (6)] The splitting of the 25% substi- 416 J Mater Chem, 1996,6(3), 415-419 tution phases is a slow process at 80°C; however, at 200°C this splitting is rapidly achieved.,, The process is reversible, so by exposure to the air at room temperature, the monohydrated phases are regenerated.This behaviour can be explained if there is alkali-metal ion migration into the phosphate layer resulting in the formation of coexisting half-exchanged and non-substituted phases (Fig. 1). The caesium 25% phase does not behave in this way due to the low ability of this cation to form half-exchanged phases. y-Ti(CsO.5Hl.SP04) (p04 )*2H20 +y-Ti (CS0.5H1 .SP04) ( p04) (5) M0.5H1 .5p04)( P04).H20 (M0.5H1.5P04) ( +?/-Ti( MHP04)( PO4) + +y-Ti( H2P04)( PO4) (6) Treatment of the 25% substitution phases at higher tempera- tures (900 "C) leads to different compound types depending on the alkali-metal ion exchanged (Na' or Cs').In the Na compounds, the condensation of the dihydrogenphosphate groups [Fig. l(c)] will give rise to TiP,O,, while the half- exchanged phase (as in the K compounds) will produce NaTi,( P04)3.The structural arrangement of the Cs com-pounds [Fig. l(b)] is different, giving rise to CsTi3P,0,,. Structural studies Samples of y-Ti(KHP04)(P04)-+H20, KTi,( PO4), and a-CsTi,P,O,, were studied using powder X-ray diffraction (XRD). The experimental patterns are shown in Fig. 2-4. y-Ti( KHPO,)(PO,).+H,O. Powder XRD data are shown in Table 2. After the full cell refinement, Smith and Snyder3' and de Wolff31 figures of merit were F2,=78(0.006,40) and M20=28.0The pattern yas indexed on monoclinic cell: a= 5.063(3) A, b=6.297(1) A, c= 11.502(5)A, p=109.30(4)".This cell is close to the y-TiP cell, suggesting that K+ ion intercal- ation does not modify the structure of the layers. Fig. 5 shows a TEM image and electron diffraction (ED) pattern along the [OOl] zone axis of y-Ti(KHP04)(P04).~H20. The a and b values obtained from the ED results concur with those calcu- lated by X-ray diffraction. KTi,(P04)3. The pattern was indexed in a rhombohedra1 cell, coincident with the known powder data PDF 34-131.32 A structural Rietveld refinement was performed using as input the structure previously reported from single-crystal data:33 0alkali-metalcation intertayer space associated with0each H,PO< gaup Fig.1 Scheme showing the dehydration process of the y-Ti(M0,5H1,5P04)(P04)nH20 phases: (a) hydrated compounds, (b) anhydrous compounds and (c) ionic migration with the formation of y-Ti( MHP04)( PO4) and y-Ti( H2P04)(PO4) phases Table 2 Powder XRD data for y-Ti(KHP04)(P0,).+H20 h k 1 dobs/ti 2OO,,/degrees I/I, AO"/degrees 00 1 10.933 8.08 45 0.038 01 1 5.457 16.23 83 0.009 1 0 -1 5.043 17.57 5 0.008 10 0 4.792 18.50 6 0.009 10 -2 4.381 20.25 18 -0.018 01 2 4.1 18 21.56 4 -0.001 1 1 -1 3.936 22.57 68 -0.002 1 1 -2 3.600 24.71 30 0.002 1 0 -3 3.504 25.40 26 0.011 02 0 3.152 28.29 100 0.010 1 1 -3 3.059 29.17 15 -0.023 02 1 3.030 29.45 21 0.038 1 0 -4 2.794 32.00 27 -0.012 00 4 2.719 32.91 13 -0.016 1 2 -1 2.670 33.53 4 -0.020 12 0 2.632 34.03 2 -0.001 1 1 -4 2.556 35.08 2 0.007 2 0 -1 2.520 35.60 16 -0.007 02 3 2.377 37.81 17 -0.024 11 3 2.339 38.46 4 -0.001 21 0 2.237 40.29 6 -0.038 12 2 2.213 40.73 5 -0.068 10 4 2.088 43.29 5 0.013 03 1 2.062 43.86 9 -0.002 03 2 1.962 46.23 15 0.061 01 6 1.743 52.46 10 -0.006 3 0 -3 1.683 54.49 5 0.077 3 0 -1 1.660 55.29 4 -0.011 3 0 -4 1.632 56.32 2 -0.014 04 0 1.575 58.54 7 -0.001 3 0 -5 1.556 59.34 2 0.013 0 3 -5 1.511 61.29 4 -0.013 5 4 s t7 Q3 F) 0,22 .-g2 v) 1 0 10 20 30 40 50 60 70 1 3 2Bldegrees Fig.2 Experimental XRD pattern of y-Ti(KHPO,)( PO4).+H2O space group R3c and isotypic with NaTi,(PO,),.The zone below 15" was excluded from refinement due to the absence of any peaks. The final agreement parameters were RWP=10.5% and R,=5.0%. The ce!l parameters oktained from the refinement, a =8.3569( 2) A, c =23.0725( 3) A, and the atomic coordinates do not show significant differences from the single- crystal results. u-CsTi,P,O,,. This compound was first obtained by Ono28 who reported powder data for two phases, PDF 39-713 (a, pattern quality '0' because it was unindexed) and 39-714.,, These phases were prepared by heating stoichiometric mixtures at 1055-1090 and 1100-1275 "C, respectively. Powder XRD data are shown in Table 3. After the full cell refinement, figures of merit were F30 =39(0.008,96) and M,, = J. Muter.Chem., 1996, 6(3), 415-419 417 1 Table 3 Powder XRD data for CsTi,P,OI9 41 i Fig. 3 Experimental (dots), calculated (solid line) and difference (below) XRD patterns of KTi,(P04), 16 I 0 10 % 30 40 50 60 70 80 90 100 110 2Bldegrees Fig. 4 Experimental XRD pattern of a-CsTi,P,O,, Fig.5 TEM image (left) and [00l] zone axis ED pattern (right) of ?-Ti( KHPO,)( P04).+H,0 28. The patter? was indexed iq a monoclinic cell: a= 6.320(1)A, b= 14.122(2)A, ~=9.333(2) A, p= 105.82(2)". Differences exist between our measurements and those of the PDF 39-713. Ono presented diffraction peaks in the 28 range 11-37", while we indexed diffraction peaks up to 63". Moreover, there are substantial differences between some of the line positions from this study and those reported by Ono.All the values reported satisfy the systematic absence rule OkO (k =2n) and h01 (I =2n), suggesting the possible space group P21/c. The intensities were measured as peak heights above the background and are expressed as a percentage of the strongest line, and were previously converted to fixed-slit values using the equation provided in the Philips documen- 418 J. Muter. Chem., 1996, 6(3), 415-419 h k 1 2OobSldegrees I/Io A2Bldegrees 01 1 7.58 11.66 10 0.008 02 0 7.06 12.52 100 0.005 100 6.09 14.55 7 0.006 -1 2 0 4.6 1 19.23 43 0.017 002 4.49 19.75 18 0.006 111 4.289 20.69 19 0.019 03 1 4.168 21.30 91 -0.007 022 3.790 23.45 63 0.009 -1 3 1 3.661 24.29 60 -0.002 -1 2 2 3.61 1 24.64 28 0.000 131 3.252 27.40 29 0.003 102 3.219 27.70 19 0.006 -2 1 1 3.077 28.99 55 0.007 140 3.053 29.23 21 -0.005 -1 1 3 2.966 30.12 18 -0.011 01 3 2.928 30.50 28 0.001 042 2.775 32.23 30 -0.004 -2 2 2 2.693 33.25 60 -0.007 21 1 2.622 34.18 5 0.003 03 3 2.527 35.49 8 0.019 -2 1 3 2.462 36.47 4 -0.015 113 2.400 37.44 3 0.001 142 2.377 37.81 2 -0.010 -2 4 1 2.353 38.25 9 0.002 23 1 2.321 38.78 5 -0.009 123 2.303 39.08 4 0.013 004 2.246 40.12 4 0.01 1 -2 3 3 2.207 40.85 9 -0.027 160 2.195 41.09 3 -0.003 024 2.139 42.21 8 -0.007 -2 0 4 2.101 43.02 2 0.012 -1 3 4 2.083 43.40 4 0.02 1 -1 5 3 2.067 43.76 3 -0.003 -1 6 2 2.053 44.06 6 0.005 -2 5 2 2.027 44.66 4 -0.010 -3 2 0 1.948 46.58 4 0.001 -1 4 4 1.940 46.78 6 0.010 -3 3 1 1.922 47.26 4 0.020 04 4 1.895 47.97 4 0.01 1 -3 2 3 1.873 48.57 3 0.013 -1 6 3 1.860 48.93 3 0.01 1 -2 6 2 1.830 49.79 3 -0.028 -2 4 4 1.804 50.54 2 -0.018 -3 4 2 1.786 51.11 4 -0.025 -3 0 4 1.763 51.81 9 0.028 -2 7 1 1.699 53.92 6 -0.010 07 3 1.672 54.8 5 8 -0.022 34 1 1.652 55.57 2 -0.019 262 1.626 56.56 4 -0.007 -4 0 2 1.576 58.51 4 -0.002 -3 6 2 1.555 59.37 4 0.021 09 1 1.546 59.75 1 0.028 313 1.491 62.22 2 -0.002 092 1.481 62.68 1 -0.017 244 1.464 63.51 1 -0.010 tation. The most intense lines are of the type k+1=2n, suggesting an occupation of special positions (on a symmetry element) by a significant part of the structure. The experimental density value (D,=2.96 g cmP3) indicates Z =2 formula units per crystal cell (D,=3.078).Fig. 6 shows a TEM image and the ED pattern along the [OlO] zone axis. This pattern can be indexed on the basis of the monoclinic cell. No extinctions caused by a centred cell were observed. Conclusions The ion-exchanged phases of lamellar acid phosphates of tetravalent metals are converted to 3D phosphates by calci- nation at high temperatures. The composition of the new materials depends on the degree of ionic substitution in the 10 A.Clearfield, Chem. Rev., 1988,88,125. 11 J. R. Garcia, R. Llavona, M. Suarez and J. Rodriguez, Trends Inorg. Chem., 1993,3,209. 12 M. Suarez, J. R. Garcia and J. Rodriguez, J. Phys. Chem., 1984, 88, 159. 13 J. R. Garcia, M. Suarez, R. Llavona and J. Rodriguez, J. Chem. Soc., Dalton Trans., 1984,2605. 14 R. Llavona, M. Suarez, J. R. Garcia and J. Rodriguez, Inorg. Chem., 1989,28,2863. 15 R. Llavona, J. R. Garcia, C. Alvarez, M. Suarez and J. Rodriguez, Solvent Extr. Ion Exch., 1986,4, 567. 16 E. Gonzalez, R. Llavona, J. R. Garcia and J. Rodriguez, J. Chem. SOC., Dalton Trans., 1989, 1825. 17 R. Llavona, J.R. Garcia, M. Suarez and J. Rodriguez, Thermochim.Acta, 1986, 101, 101. 18 R. Llavona, J. R. Garcia, M. Suarez and J. Rodriguez, Thermochim. Acta, 1985,86,281. Fig.6 TEM image (left) and [OlO] zone axis ED pattern (right) of 19 M. Kolthoff, E. B. Sandel, E. J. Meehan and S. Bruckenstein, a-CsTi,P,O,, Quantitative Chemical Analysis, Nigar, Buenos Aires, 1972, p. 666. 20 0.B. Michelsen, Anal. Chem., 1957,29,60. initial solid and the counter-ion characteristics. Compounds of high purity are obtained from this synthetic method, which is another alternative to the classic methods of high-tempera- 21 22 P. E. Werner, L. Eriksson and M. Westdahl, J. Appl. Crystallogr., 1985, 18,367. D. E. Appleman and H. T. Evans, NTIS Document No. PB- 216188. ture materials preparation. 23 D. B. Wiles and R.A. Young, J. Appl. Crystallogr., 1981,14,149. 24 E. Gonzhlez, R. Llavona, J. R. Garcia and J. Rodriguez, J. Chem. We thank the CICYT (Spain) for financial support, Research Project no. MAT94-0428. 25 SOC.,Dalton Trans., 1989, 829. R. Llavona, C. Alvarez, J. R. Garcia, M. Suarez and J. Rodriguez, Solvent Extr. ion Exch., 1985,3, 931. 26 T. P. Whaley, in Comprehensive inorganic Chemistry, ed. J. C. Bailar, H. J. EmelCus, R. Nyholm and A. F. Trotman- References 27 Dickenson, Pergamon Press, Oxford, 1975,vol. 1, p. 481. R. Llavona, J. R. Garcia, C. Alvarez, M. Suarez and J. Rodriguez, 1 2 3 4 H. Y-P. Hong, Muter. Res. Bull., 1976, 11, 173. F. D’Yvoire, M. Pintard-Screpel, E. Brety and M. De la Rochere, Solid State ionics, 1983, 10, 851. S. I. Limaye, D. K. Agrawal and H. A. McKinstry, J. Am. Ceram. Soc. Commun., 1987, C232. A. Clearfield and P. Jirustithipong, Muter. Res. Bull., 1980, 15, 1603. 28 29 30 31 32 Solvent Extr. Ion Exch., 1986,4, 585. A. Ono, Bull. Chem. Soc. Jpn., 1985,58,3039. R. Llavona, M. Suarez, J. R. Garcia and J. Rodriguez, Anal. Chem., 1986,58, 547. G.J. Smith and R. L. Snyder, J. Appl. Crystallogr., 1979,12,60. P. M. de Wolff, J. Appl. Crystallogr., 1968, 1, 108. Powder Diflraction Files, International Centre for Diffraction A. Clearfield, Annu. Rev. Muter. Sci., 1984, 14,205. A. Clearfield, Eur. J. Solid State Inorg. Chem., 1991,28, 37. A. Clearfield, Muter. Chem. Phys., 1993,35,257. 33 Data, Swarthmore, PA, 1987. E. S. Lunezheva, B. A. Maksimov and 0. K. Mel’nikov, Kristallografiya, 1989,34, 11 19. G. Alberti, Chem. Rev., 1978, 11,163. inorganic ion Exchange Materials, ed. A. Clearfield, CRC Press, Boca Raton, FL, 1982. Paper 5/02857D; Received 3rd May, 1995 J. Mater. Chern., 1996, 6(3), 415-419 419

ISSN:0959-9428    

DOI:10.1039/JM9960600415

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Synthesis, crystal structure and spectroscopic properties of the NH,NiPO,vzH,O (n=1,6) compounds; magnetic behaviour of the monohydrated phase Aintzane Goiii," JosC Luis Pizarro,b Luis M. Lezama,' Gaston E. Barberis,' Maria Isabel Arriortuab and Teofilo Rojo*" "Dpto.Quimica Inorgcinica (ISEM), Universidad del Pais Vasco, Bilbao 48080, Spain bDpto.Mineralogia y Petrologia (ISEM), Universidad del Pais Vasco, Bilbao 48080, Spain NH,NiPO4-6H,O and NH4NiP04.H,0 have been obtained by adding different concentrations of H$O, to dilute solutions of NiCl2-6H20 with special attention to the control of the pH in the solvent medium, which was regulated by addition of NH40H. The NH,NiP04.6H,0 compound crystallizes in the orthorhombic Pmn2, space group with cell parameters a =6.9032(8), b=6.0907(5) and c= 11.1402(8)A,V=468.39(7) A3, 2=2, R=2.3 and R,=2.3%.The structure is three-dimensional and consists of Ni[O(W)]6 octahedra [O(w) =oxygen from a water molecule] linked to PO, and NH, tetrahedra by hydrogen bonds. All polyhedra are quite regular in this compound. The NH4NiPO4*H20 space group with cell phase crystallizes in the Pmr~2~ parameters a =5.5698(2),b =8.7668( 2) and c =4.7460( 2) A. The structure of this compound has been refined with the Rietveld method using the coordinates of the KMnPO,.H,O phase as a starting model. The final residual factors were R, =7.37, RB=2.68%. The structure is formed from sheets of distorted NiO, corner-sharing octahedra bridged through the oxygen atoms of the phosphate tetrahedra.These layers are pillared along the b direction and are interconnected by hydrogen bonds with the NH4+cations, which are inserted between the sheets. The spectroscopic properties of both compounds are in good agreement with the symmetry observed in each phase. The values of the nephelauxetic ratio, p, are 0.89 and 0.94 for the hexahydrated and monohydrated compounds respectively. Magnetic susceptibility and specific heat results obtained for NH,NiPO4-H20 show an essentially two-dimensional antiferromagnetic exchange coupling, which becomes of a more three-dimensional behaviour with decreasing temperature. Nickel(I1) phosphates offer a considerable number of different structures which can give rise to practical applications such as ion exchange, ionic conductivity, etc.,' and interesting magnetic properties.The choice of synthetic method is important, as it can lead to the production of several phases with predetermined structure types. Water solution chemistry procedures can gener- ate nickel phosphates with a variable number of H20 molecules coordinated to the metal. This number depends on the reaction conditions, such as pressure and temperature. The affinity of Ni" and other divalent transition metals to coordinate water molecules sometimes prevents other ligands from forming intermetallic bridges. When there is a large number of water molecules in the formula of a compound, most of the coordi- nation positions of the metal are occupied by water molecules, leading to three-dimensional (3D) structures built via hydrogen However, a small number of water molecules in the coordination sphere of the metal allows the bonding of other groups as PO,, which can form strong intermetallic bridges.This can lead to the formation of compounds with interesting magnetic behaviour. 5,6 The well known series of compounds M'M"PO,.H,O (MI= NH,, K; M"=Mn, Fe, Co, Ni),7 has been of interest because of the strongly defined layered crystal structures of these phases. The divalent metal ions are bridged by the oxygen atoms of the phosphate groups, leading to the formation of magnetic planes separated by NH,' ions. This arrangement should afford interesting two-dimensional (2D) magnetic inter- actions. However, the study of the magnetic behaviour of the manganese compound' showed a crossover in the power law dependence of the magnetization with temperature.This result was attributed to a crossover in the lattice dimensionality from 2D to 3D at low temperatures. On the other hand, magnetic and Mossbauer studies of the NH4FeP0,.H,0 compo~nd~-'~ showed, at low temperatures, two different regimes. One of them involves a short-range ordered region at 70 > T/K >26 and the second is a long-range ordered region below 26 K, with an uniaxial antiferromagnetic state with the axis nearly parallel to the layer stacking direction. These results may be explained by the existence of superexchange pathways between the magnetic layers through the intercalated NH,' ions. In this article, we report the phase diagrams in water solution obtained for the [Ni2+ /H,PO,/NH,OH] system, in which the presence of three different nickel@) phosphates has been observed: Ni3( P04)2*8H20, NH4NiPO4.H2O and NH4NiPO,.6H,O.The Ni3( P04)2.8H20 compound has the vivianite structure and has been studied extensively." In this work, the other two ammonium nickel phosphates are ana- lysed, and their magnetic properties and the specific heat of the monohydrated phase are discussed. Experimental Synthesis A systematic investigation of the reactions in water solution for the [Ni2+/H3P04/NH40H] system was carried out. The phase diagrams obtained are shown in Fig. 1. The precipitates were characterised by elemental analysis. The nickel and phosphorus content were determined by atomic absorption spectroscopy (Perkin-Elmer 3030B) and thermogravimetric techniques, respectively.Characterisation by X-ray powder diffraction was also performed. As can be seen in Fig. l(a), mixtures of phases exist in different pH ranges and Ni:P ratios, which may due to the overlap of areas of thermodynamic equilibrium or the low solubility of the nickel phosphates. The phase diagram undergoes important variations when the precipitate is maintained in contact with the solvent medium at room temperature and pressure for a week [see Fig. l(b)]. In this case, mixtures of phases and the NH,NiPO,-H,O phase are not observed. This result indicates that the solid $ solution equilibrium favours the evolution of this species to form the NH4NiP04*6H20 compound which is thermodynamically more stable.The title compounds were synthesized by adding NiCl2-6H,O (4x mol dm-3) to different solutions of H,PO, and NH,OH at 90°C. Green and yellow polycrystal- J. Mater. Chem., 1996, 6(3), 421-427 421 I :40 1:40 - 5 130 1:30- z 1:20 1:20 - ... ... 1:lO ... . .. . . .... . .. 1:lO - 11 ., . .... .. . .. . . .,.... .. .. 1:1 - I 1 2 3 4 5 6 7 8 91011 1 2 7 A S 6 7 8 91011 NH4NiP0,.6€120 0 Ni3(P0,).8H,0 0 No precipitate Fig. 1 Phase diagrams in water solution obtained for the [NiZf/H,PO,/NH4OH] system: (a)the precipitate was filtered off immediately; (b)one week after synthesis (see text) line samples were obtained for NH,NiP0,.6H20 and NH,NiPO,-H,O respectively, but only recrystallization of NH,NiP0,.6H20 from water solution gave crystals of X-ray diffraction quality.Structure refinement of NH4NiP04*6H20j-A prismatic single crystal of NH,NiP0,.6H20 (0.12 x 0.15 x0.22 mm3) was selected for structure determination. Preliminary cell dimensions were calculated by oscillation and Weissenberg photographs. Diffraction experiments were per- formed on an Enraf-Nonius CAD-4 automatic diffractometer using graphite-monochromated Mo-Ka radiation. The orien- tation matrix and the final lattice constants were determined from 25 high-angle reflections (14<28 <27"). Crystal and data collection parameters are summarized in Table 1. Two standard reflections were recorded every 2 h.Their intensities showed no statistically significant change over the duration of the data collection. Lorentz and polarization corrections, as well as an empirical absorption correction (DIFABS program12) were applied to the data. The crystal structure was refined using as the starting structural model the fractional coordinates and the space group of the struvite mineral, NH,MgP04.6H20,13-15 by the full-matrix least-squares method (SHELX76I6). The origin was fixed by keeping the z coordinate of the P atom constant. Further anisotropic refinements followed by a difference Fourier synthesis allowed the location of all H atoms. Atomic scattering factors were taken from the International Tables for X-ray Crystal10graphy.l~ All non-hydrogen atoms were refined with anisotropic thermal parameters and the hydrogen atoms with isotropic ones.The final difference Fourier map showed no peaks higher or deeper than k0.38 e A-3. The final fractional atomic coordinates and displacement parameters are given in Table 2. Selected interatomic distances are shown in Table 3. The geometric calculations were per- formed with the programs PARST" and BONDLAI9 and molecular illustrations were drawn with the ATOMS2' program. Structure refinement of NH4NiP04*H,0j-The X-ray powder diffraction pattern of NH,NiPO,.H,O was measured at room temperature on a Stoe (Darmstadt) diffractometer using Ge-monochromated Cu-Ka, radiation in reflection mode. The data were collected in the 28 range 5-1 10" in steps of 0.02" (Table 1). The Rietveld refinement of the crystal structure was undertaken by starting from the structural t Single-crystal data are available from Fachinformationszentrum Karlsruhe, D-76344 Eggenstein-Leopoldshafen, Germany.422 J. Muter. Chem., 1996, 6(3), 421-427 model of KMnP04.H20.21 The structure was refined using the program FULLPROF2' with a pseudo-Voigt function used to model the peak shape. A March model for preferred orientation correction was applied because of the plate habit of the microcrystals, parallel to the (010) plane, as well as an asymmetry correction to the low-angle reflection. Scale and background variables were refined initially followed, in sub- sequent iterations, by the zero point of 28, the cell constants, the peak-shape parameters and the atomic and isotropic thermal parameters.The origin was fixed by keeping the z coordinate of the Ni atom constant. The residual factors dropped by successive refinement cycles to R,, =7.37% and RB=2.68%. Fig. 2 shows the observed, calculated and difference powder X-ray diffraction (XRD) patterns of NH,NiPO,.H,O. The final fractional atomic coordinates, equivalent displacement parameters and selected interatomic distances are given in Tables 2 and 3. Physical measurements IR spectra were obtained using KBr pellets (0.5%) on a Nicolet FTIR 740 spectrophotometer in the 4000-400 cm-I region. Reflectance spectra were measured at room temperature on a Cary 2415 spectrometer in the range 5000-40 000 cm-'.Thermogravimetry (TG) measurements were carried out with a Perkin-Elmer 7 system. Crucibles containing 20 mg of sample were heated at 5°C min-' under a nitrogen atmosphere. Magnetic measurements were performed on a polycrystalline sample using a Quantum SQUID magnetometer, in the 300-2K temperature range. The magnetic field used for the magnetic measurements was 1000G. Specific heat measure- ments were carried out using a quasi-adiabatic method, in the temperature range 1.5-30 K. The calorimeter was made of an adiabatic chamber with a germanium resistance thermometer and an evaporation deposit Cr/Ti heater. The sample holder was suspended with a Nylon thread. Results and Discussion Crystal structure of NH4NiP04.6H20 The crystal structure of NH,NiP0,.6H20 consists of isolated Ni[O(w)16 octahedra bonded to the PO, and NH, tetrahedra by hydrogen bonds (Fig.3). The Ni[O(w)16 octahedra are fairly regular, with a range of Ni -0(w) distances from 2.092( 5) to 2.040(3) A, and O(w)-Ni-O(w) angles ranging from 95.9(1) to 87.8(1)". The PO, tetrahedra show a high local symmetry caused by the special position of the P atom, located in a mirror plane. The P-0 distances with a mean value of 1.535A do not show important variations (<la), and the Table 1 Crystallographic data for the NH,NiPO,.nH,O (n= 1,6) compounds NH4NiP0,.6H,O NH,NiPO,.H,O Mr 279.8 189.74 crystal system orthorhombic orthorhombic sp!ce group (no.) Pmn2, (31) Pmn2, (31) a/+ b/+CIh 6.9032( 8) 6.0907(5) 11.1402( 8) 5.5698(2) 8.7668(2) 4.7460( 2) VIA3 468.39( 7) 231.75(1) z 2 2 F (000) 292 Pobsli? cm - 1.96(3) 2.68(5) Pcalclg cm -TIK 1.98 295 2.72 295 p/cm - 22.81 diffractomete; Enraf-Nonius CAD4 Stoe (Darmstadt) radiation (A/A) Mo-Ka (0.71069) Cu-Ka, (1.5406) scan type wf28 w/28 h,k,l range 9, f8, k15 28 rangeldegrees 4-60 5-110 reflections collected 3035 unique reflections 771 reflections with 132.541) 675 no.of parameters 105 R 0.023 RP 0.052 Rw 0.023 RWP 0.074 RB 0.027 Table 2 Fractional atomic coordinates and equivalent displacement parameters (A2) for the NH,NiPO,.nH,O (n= 1,6) compounds NH,NiPO, .6H,O NH,NiPO,.H,O atom X Y Z Beqa/A2 atom X Y z B,,"/A2 0 0.37285(9) 0.37222( 13) 1.62(1) 0 -0.00548( 18) 0.00190 1.47(2) 0 0.3696( 10) 0.7347( 6) 2.93(11) 0 -0.0222(5) --0.1355 (4) 2.14(7) 0 -0.2379(6) 0.0557( 3) 2.03(7) 0.1826( 4) 0.1157(4) 0.0443 (3) 2.07(5) 0 0.6779( 7) 0.2907(5) 3.24(12) 0 0.0661(7) 0.4566(5) 2.91(10) 0.2194(4) 0.2675(5) 0.2639( 3) 2.20(6) 0.2055( 4) 0.4785(5) 0.491 3 (3) 2.72( 7) Table 3 Selected bond lengths (A) and angles (degrees) for the NH,NiPO,.nH,O (n= 1,6) compounds NH,NiP0,.6H20 NH,NiPO,.H,O Ni coordination octahedra Ni P N 0(1) O(2) 0 0 0 0 0 -0.0189( 2) 0.1902(2) 0.5290(6) 0.1624(5) 0.3659(4) 0.00427 -0.4278(8) 0.1255(15) -0.7481( 13) -0.3628( 12) 1.26( 2) 1.26(2) 1.26(2) 1.26(2) 1.26(2) O(3) O(w) 0.2239(9) 0 0.1135(3) -0.2010(5) -0.2779( 10) -0.3015( 12) 1.26(2) 1.26(2) 0-P-0 angles are in the range from 108.8( 1) to 110.4( 1)O. So, the PO4 groups can be described as low-distorted tetra- hedra.The NH4+ tetrahedFa are rather regular too, with mean N-H distances of 0.88A and mean H-N-H angles of 109( 3)". The 3D structure of the title compound is formed by a complicated scheme of hydrogen bonds established by the water molecules and ammonia groups. The hydrogen- bonding arrangement of the four independent water molecules in the structure contains eight different bonds. Seven of them are among the strongest hydrogen bridges given by the H20 molecule in crystalline hydrates, with values ranging from 2.607(4) to 2.691(4) A.The eighth bond, O(wl).-.O(~2)~ (f= x,1+y, z), however, is near to !he upper limit for this type of bond,23 with a value of 3.00(7) A. The hydrogen bridges of the ammonium group are rather different, ranging from a shorte!t one N...O(l)g (g=x,y, ++z) with a distance of 2.790(7)A, to two bifurcated weak bonds, N-..O(w4)/0(w4)', and N...O(w3)j/O(w3)"(h=+-x, 1-y, ++z; i= -x,y, z;j=x-3, 1-y, z++),with distances of 3.131(7) and 2.957(6) A,respect-ively. No bond was detected between N-..O(w2), in contrast to the model proposed for NH4MgP04.6H20r1' because the observed distance, 3.607( 8) A, is too long. Crystal structure of NH4NiP04*H20 NH4NiP04.H20 presents a layered structure, with nickel phosphate sheets separated by NH4+ cations, as shown in J.Muter. Chem., 1996, 6(3), 421-427 423 Ni-0(w1 ) Ni -0(w2) Ni-O(w3)/O(w3)" Ni-0(w4)fO (w4)" 0-Ni-0 (ca. 90") 0-Ni-0 (ca. 180") phosphate tetrahedra P-O(1) P-O(2) P-O(3)/O(w3)" O(1)-P-0(2) O(1)-P-0(3)/0(~3)' 2.068(5) Ni-O( l)b 1.977(5) 2.092(5) Ni-O(w) 2.158(5) 2.040( 3) Ni- O(3)/0( 3 r 2.172(4) 2.046( 3) Ni-O( 3)d/0( 3)" 2.023(5) 90(2) 0-Ni-0 (ca. 900) 90(7) 177(2) 0-Ni-0 (ca. 180") 169(5) 1.534(5) P-0(1) 1.540( 7) 1.538(4) P-0(2) 1.571 (4) 1.535(2) P-0(3)/0(3)' 1.593(5) 109.2(1) 0(1)-P-0(2) 110.4( 3) 109.8(2) O(l)-P-0(3)/0(3~ 112.0(2) O(2)-P-O(3)/0(w3)" 108.8 (2) O(2)-P-O(3)/0( 3 r 109.1 (2) 0(3)-P-0(3)" 110.4(1) 0(3)-P-0(3)' 104.2(3) ammonium tetrahedra and hydrogen bonds N-H 0.88(2) (hydrogen atoms not located) H-N-H 109( 3) N-acceptor 2.8(2) N.-.0(2) 2.85( 1) H- donor -H 107(5) 0(2)*..N..-0(2) 109( 18) Symmetry codes: a -x, y, z; b~ y,, l+z; '-x, y, z; d~-+7 -Y7 s+z; =+-x, --y, ++z.Fig. 2 Observed, calculated and difference powder XRD pattern of NH,NiPO,*H,O Fig. 3 View of the crystal structure of NH,NiP0,-6H20 Fig. 4. The inorganic layers are formed by the NiO, corner-sharing octahedra, crosslinked by the phosphate tetrahedra. The Ni(O), octahedra are highly distorte!, with a range of Ni-0 distances from 1.977(5) to 2.172(4) A, and 0-Ni-0 angles ranging from 98.1(2) to 86.4(2)". The four equatorial bonds Ni-0(3)/o(3)~/o(3)d/0(3)eare 2.172(4), 2*172(4), 2.023(5) and 2.023(5) A respectively (Table 3) are responsible for the.'crosslinking in the sheets. The axial 'oxygen Ojl)b is provided by the PO4 tetrahedron [Ni-O( l)b1.977(5)A] and the remaining axial vFrtex is the O(w) of the coordinated water [Ni-O(w) 2.158(5) A]. 424 J. Mater. Chem., 1996, 6(3), 421-427 Fig. 4 View of the crystal structure of NH,NiPO,.H,O The PO4 groups can be described as distorted tetrahedra, even though the special position of the P atom is located in a mirror plane as in NH4NiP04.60H,0. The P-0 distances with a mean value of 1.57(2)A range from 1.540(7) to 1.593(5) A, and the 0-P-0 angles are in the range 104.2(3)-112.0(2)". The PO4 groups are connected to the nickel octahedra by three oxygen vertices, 0(1), 0(3), 0(3)', with one common edge, 0(3)...0(3)".The fourth oxygen corner, 0(2), is directed towards the interlayer space. The NH4+ ions inserted between the inorganic layers estab- lish hydrogen bonds with the symmetrical O(2) atoms of the four adjacent phosphate groups of two facing layer!. The N-0(2) distances range from 2.733(8) to 2.933(2)A, and the 0(2)-P-0(2) angles are in the range 98.5(2)-143.4(2)". Thermogravimetricstudy The thermogravimetry and differential thermogravimetry (TG and DTG) curves of the title compounds obtained under a nitrogen atmosphere from room temperature to 600 "C are represented in Fig. 5. The thermal decomposition study of I I I I I I 1 100 200 300 LOO 500 TIT Fig. 5 TG (-) and DTG (---) curves of (a) NH4NiP04-6H,0 and (b)NH4NiP04.H20 NH4NiP04-6H20 shows only one endothermic step in the temperature range 80-300 "C, which corresponds to the mass loss of one ammonia and six water molecules (exptl.44.0%, calc. 44.72%). The TG study of NH4NiP04.H20 shows two non-solved steps, in the temperature range 170-300°C, with a total mass loss of 19%. This loss can be attributed to the water and ammonium molecules present in the structure (calc. 18.5%). The two overlapped steps can be solved by using atmospheres of ammonia or humid Finally, a gradual mass loss is observed in the TG curves of both compounds to give nickel(I1) pyrophosphate, Ni2P20,, at temperatures higher than 400 "C. Spectroscopic properties Selected bands obtained from the IR spectra of NH4NiP04.6H20 and NH4NiP04.H20 are given in Table 4.Both phases present a set of bands in the range 3500-3000cm-', which can be assigned to the stretching vibration modes of the OH and NH groups, in good agreement with the presence of the NH4+ group and the water molecules coordinated to the metal in the structures. In the case of NH4NiP04.6H20, these bands are strong because of the presence of six water molecules in the compound. The IR spectra of both compounds show the bending mode of the 0-H group around 1615cm-'. The bands which appear at 1470 and 1440cm-' in the spectrum of NH4NiP04.H20 and at 1455 cm-' in the NH4NiP04-6H20 phase can be ascribed to the bending mode of the NH4+ group. In the case of the hexahydrated compound this band is not split because of the NH4+ tetrahedra present high symmetry in the structure.However, for the monohydrated compound a splitting is observed which is indicative of the existence of distortions in the NH, polyhedra.+ The vSt(PO4) bands appear in the range 1100-950 cm- ' for NH4NiP04.H20. These bands are split owing to the distortion of the PO4 tetrahedra observed in the structure. However, in the hexahydrated compound only one band assigned to this vibration mode is present at 1015 cm-', which indicates that the phosphate tetrahedra are quite regular, in good agreement with the structural features. These results are confirmed by the presence of the PO4 bending modes which are located at 625-560 cm-', and 575 cm-' for the monohydrated and hexa- hydrated compounds, respectively.The reflectance spectra of both compounds exhibit three strong absorptions corresponding to the three spin-allowed transitions expected for a d8 ion in octahedral symmetry C3A2,j3T2,, 3A2, j3T1,, 3A2, +3Tl,(P)], and two more sig- nals which can be attributed to the spin-forbidden tran-sitions 3A2g-*'Eg and 3A2,+'T2,, respectively. For the NH4NiP04.6H20 compound strong absorption bands were observed at 8400, 13800 and 25000 cm-', whereas in the spectrum of the monohydrated compound these signals are displaced to lower frequencies, 7200, 12900 and 23900 cm-', as a consequence of the distortion in the NiO, octahedra. The values determined for the lODq (8400 cm-') and Racah param- eters, B (918 cm-') and C (4229 cm-'), corresponding to the NH4NiP04.6H20 compound are typical for an Ni2+ ion in an octahedral environment of six water with a value C/B=4.61 whcih close to the value for the ideal octahedral geometry (C/B=4.46).However, for the NH4NiP04.H20 compound these values are lODq = 7200cm-', B=967cm-' and C=3720cm-', which give a value of C/B=3.85, characteristic of the Ni2+ ion in a distorted Table 4 Selected bands (v/cm-') obtained from the IR spectra for the NH4NiP04~nH,0(n= 1,6) compounds NH,NiP04 .H,O 3390m 1610w 1470m 1 lOOm 625m 3210w 1440m 1085m 560m 3060w 1050s 2930w 950s NH4NiP04 .6H,O 3450m 1620w 1455m 1015s 575s 32 15s 3100s 2935s w =weak, m =medium, s =strong.J. Mater. Chem., 1996, 6(3), 421-427 425 octahedral environment These results are in good agreement with the structural data of both phases Magnetic properties of the NH,NiPO,*H,O compound The thermal variations of both the magnetic susceptibility and x,T for the NH4NiP04 H20 compound are shown in Fig 6 The thermal evolution of xrn satisfies the Cune-Weiss law at high temperatures with C=12 emu K mol-I and 8= -27 44 K, and exhibits a xmaxcentred at 140 K This result, together with the continuous decrease in the x,T values, is indicative of antiferromagnetic exchange couplings in the compound Owing to structural findings of this phase, two different antiferromagnetic models could be considered for the analysis of the magnetic behaviour the 2D square-planar Heisenberg system for S=1, considering only magnetic interactions between the nickel(1r) ions arranged inside the layers, or the 3D Heisenberg system for S= 1, which also considers inter- actions between layers through the NH4+ groups In the first case, the validity of the 2D Heisenberg model was tested by has been also carned out The thermal variation of the magnetic contribution (C,) to the specific heat, calculated by substrac- tion of the lattice contribution to the expenmental values, showed a broad maximum centred at 8 8 K [Fig 7(u)] The magnetic entropy calculated for the reached maximum tem- perature gave a value of AS= 8 46 J rno1-l K-l, which is close to the theoretical value for AS=R ln(2S+ 1) =9 13 J mol-' K-', with S=l [Fig 7(b)] 59% of the total entropy is acquired above the maximum (at 8 8 K) indicating a high degree of short-range interactions Taking into account that the 3D ordering causes in the magnetic specific heat a A-type second-order transition, the rounded maximum observed in Fig 7(a) for NH,NiPO, H20 cannot be interpreted as being due to the existence of an order of 3D character So, the existence of a 3D order at temperatures higher than 1 8 K, which is the lower temperature limit reached in the measurements, should be discarded However, this fact does not exclude the existence of weak antiferromagnetic 3D interactions In this way, it was not possible to fit the exper- imental values of the thermal variation of C, by using the 2D model for magnetic specific heat of Rushbrooke and using the expressions described by De Jongh and M~edema~~ as was also observed in the case of the thermal variation of for the calculation of the exchange coupling value, J/k Xmax IJ I Ns2P2 =00521 Two different J/k values, 3 18 K and 3 68 K, were obtained using expressions (1) and (2), respectively These results indi- cate no agreement between the experimental data and the used model In the same way, it was not possible to fit the xm experimental values (Fig 6) using the analytic expression reported by Lines28 for a 2D square-planar Heisenberg system from a high temperature expansion series studied by Rushbrooke and Wood 29 x= 2Ng2P2[1+AX+Bx2+Cx3+Dx4+Ex5+Fx6]-'(3)~ 3kT where x =J/kT, A= 5 333333, B =9 777778, C =9 481482, D =19 06173, E =45 08971 and F =25 46392 On the other hand, the magnetic behaviour of this phase was also studied with a 3D Heisenberg model using the expression of Rushbrooke and given in eqn (3), with the values A= 8, B =14 66667, C =14 2222, D=61 185, E = 162 449 and F =1127 96 As can be seen in Fig 6 these results show no good agreement between the experimental data and the 3D model In order to clarify the magnetic behaviour of NH4NiP04 H20, a study of specific heat for this compound 0 025 0 --L 002 a$ E Y 0015 63 E,"N t-001 4N 0 005 2 0t 0 50 100 150 200 250 300" TIK Fig.6 Thermal variation of xrn and x,T for NH4NiP04H20 the circles are the expenmental values and the full lines represent the theoretical values for 2D and 3D Heisenberg systems 426 J Mater Chem, 1996,6(3), 421-427 magnetic suscepti bill ty Considering the structural data, the exchange pathways of the 3D interactions must involve the NH4 groups This path- way implies that the hydrogen bonds between the oxygen atoms of the PO 4g+roups are directed toward the interlayer space and the NH4 ion inserted between the layers, leading to weak antiferromagnetic interactions With respect to the intralayer magnetic exchange, two different pathways could be deduced One of them involves the d,z-yz orbitals from the N106 octahedra linked through the O(3) atoms (see Fig 4), leading to antiferromagnetic couplings The second pathway involves the PO roup as was observed for other transition metal phosphate:! giving ferromagnetic interactions One can conclude that the 2D simple model does not explain the magnetic behaviour of the layered NH4NiP04 H20 com- pound and that an increase of the 3D interlayer ordering occurs when the temperature decreases, which agrees with the results obtained for other related compounds 25 These results lead us to deduce that a 2D-3D intermediate model for the (a)6-5-Lz4--33 02-':.L.,.. I ....I .,..I ' ...I...,* 0 5 10 15 20 25 30 TIK Fig.7 Thermal variation of (a) the magnetic specific heat, C,, and (b) AS for the NH4NiP04H20 compound fitting of the magnetic behaviour of this compound might be necessary.5 6 C. Calvo and R. Faggiani, Can. J. Chem., 1975,53,1516. T. Rojo, L. Lezama, J. M. Rojo, M. Insausti, M. I. Arriortua and G. Villeneuve, Eur. J, Solid State Inorg. Chem., 1992,29,217. 7 D. Tranqui, A. Durif, J. C. Guitel and M. T. Averbuch-Pouchot, Conclusions Two phase diagrams in water solution were obtained for the [Ni2+ /H3P04/NH40H] system, in which the presence of two ammonium nickel(I1) phosphates is observed: NH4NiP04-H20 and NH4NiP04.6H20. The crystal structure of the hexahy- drated compound consists of isolated Ni [0(w)l6 octahedra bonded to the PO4 and NH, tetrahedra by hydrogen bridges. 8 9 10 11 Bull. SOC.Fr. Mineral. Crystallogr., 1968,91, 10. S. G. Carling, P. Day and D. Visser, Solid State Commun., 1993, 88,135. J. E.Greendan, K. Reubenbauer, T. Birchall, M. Ehlert, D. R. Corbin and M. A. Subramanian, J. Solid State Chem., 1988,77,376. S. G. Carling, P. Day and D. Visser, Acta Crystallogr., Sect. A, J. L. Pizarro, M. I. Arriortua, L. Lezama, T. ROJO and G. Villeneuve, Solid State Ionics, 1993,63-65, 71. 1990,46, C-278. It adopts a three-dimensional structure in which all polyhedra are linked together by a complicated scheme of hydrogen bonds. The NH4NiP04.H20 compound presents a layered structure, with nickel phosphate sheets separated by NH4+ cations. The inorganic layers are formed by the NiO, corner-sharing octahedra, crosslinked by the phosphate tetrahedra. 12 13 14 15 N. Walker, and D. Stuart, Acta Crystallogr., Sect. A, 1983,39, 158. A. Whitaker and J. W. Jeffery, Acta Crystallogr., Sect.B, 1970, 26, 1440. F. Abbona, M. Calleri and G. Ivaldi, Acta Crystallogr., Sect. B, 1984,40,223. G. Ferraris, H. Fuess and W. Josswig, Acta Crystallogr, Sect. B, 1986,42,253. The spectroscopic data for the hexahydrated compound con- firm the high symmetry of the NiO,, PO, and NH, groups, whereas the data for monohydrated compound show the existence of important distortions in these polyhedra. The antiferromagnetic behaviour observed in the layered NH4NiP04.H20compound cannot be explained by either the 2D or the 3D Heisenberg models because of the existence of weak exchange interactions between layers. 16 17 18 19 20 G. M. Sheldrick, SHELX76, Program for Crystal Structure Determination, University of Cambridge, 1976. International Tables for X-Ray Crystallography, Kynoch Press, Birmingham, 1974, vol.IV, pp. 72-98. M. Nardelli, PARST, Comput. Chem., 1983,7,95. J. M. Stewart, F. A. Kundell and J. C. Baldwin, The X-ray 70 system, computer science center, University of Maryland, College Park, MA, 1976. E. Dowty, ATOMS: A Computer Program for Displaying Atomic Structures. Shape Software, 521 Hidden Valley Road, Kingsport, This work was carried out with the financial support of the Basque Government (PI-9439) which we gratefully acknowl- edge. (G. E. B., on sabbatical from UNICAMP (Brazil) 21 22 TN, 1993. D. Visser, S. G. Carling, P. Day and J. Deportes, J. Appl. Phys., 1991,69,6016. J. Rodriguez Carvajal, FULLPROF Program. Rietveld Pattern acknowledges the Ministerio de Educacibn (Spain) and CNPq (Brazil) for financial help. We thank R. Kuentzler and Y. Dossmann (Strasbourg) for the specific heat measurements. 23 24 Matching Analysis of Powder Patterns., 1994. M. Falk and 0. Knop, Water: A comprehensive treatise, vol. 2. Plenum, New York, 1973. J. Pysiak, E. A. Prodan, V. V. Samuskevich, B. Pacewska and N. A. Shkorik, Thermochim. Acta., 1993,222,91. References 25 A. Gofii, J. Rius, J. L. Pizarro, M. Insausti, L. Lezama, M. I. Arriortua and T. Rojo, Chem. Muter., submitted. 1 T. Kanazawa, Inorganic Phosphate Materials. Elsevier, Tokyo, 1993. 26 27 A. Bose and R. Chatterjee, Proc. Phys. SOC. London, 1963,82,23. L. J. De Jongh and A. R. Miedema, Adv. Phys., 1974,23,1. 2 A. Jouini, M. Dabbabi and A. Durif, J. Solid State Chem., 1985, 28 M. E. Lines, J. Phys. Chem. Solids, 1970,31, 101. 60, 6. 29 G. S. Rushbrooke and P. J. Wood, Mol. Phys., 1963,1,257. 3 M. Berraho, C. R'Kha, A. Vegas and M. Fafiq, Acta Crystallogr., Sect. C, 1992,48, 1350. 30 L. Lezama, K. S. Suh, G. Villeneuve and T. Rojo, Solid State Commun., 1990, 79,449. 4 A. Whitaker and J. W. Jeffery, Acta Crystallogr., Sect. B, 1970, 26, 1429. Paper 5/04974A; Received 26th July, 1995 J. Muter. Chem., 1996, 6(3), 421-427 427

ISSN:0959-9428    

DOI:10.1039/JM9960600421

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Crystal structure and physical properties of UAuSi and UAu, Rainer Pottgen," Vinh Hung Tran,b Rolf-Dieter Hoffmann," Dariusz Kaczorowskib and Robert Trocb "Anorganisch-Chemisches Institut, Universitat Miinster, Wilhelm-Klemm-Strasse 8, 0-481 49 Miinster, Germany bW.Trzebiatowski Institute for Low Temperature and Structure Research, Polish Academy of Sciences, PO Box 937, 50-950 Wroctaw, Poland UAuSi and UAu, have been prepared by arc-melting of the elemental components and subsequent annealing. The crystal structure of UAuSi, which has been previously reported to crystallize in the TiNiSi-type structure, was determined from X-ray powder diffraction data: LiBaSi-type structure, P6m2, u=419.5(2) pm, c=397.2(2) pm, 1/=0.0605 nm3, Z= 1and R,(I)=0.018. It is derived from the AIB,-type structure by an ordered arrangement of the gold and silicon atoms on the boron sites.The binary compound UAu, (AlB,-type) is confirmed. Magnetic susceptibility measurements indicate spin-glass behaviour and paramagnetism for UAuSi and UAu,, respectively. Both compounds are metallic conductors. The binary uranium silicide p-USi, adopts the hexagonal structure of A1B2.172 The same structure was also reported for UAu, .3,4 However, Dommann and Hulliger' reported UAu, with the CeCd,-type structure,' which is derived from that of AlB, by simply shifting the gold position from z=OS to z= 0.45, resulting in slightly puckered layers. Thus, the silicon atoms in P-USi, should be substitutable by gold atoms. We recently reported on U,AUS~~,~ the first compound in the pseudo-binary system US,-UAu, .Similar U2TSi3 compounds exist with other transition metal^.^-^ Another compound in this system is UAuSi," previously reported to adopt the orthorhombic structure of TiNiSi; sus-ceptibility measurements on this compound indicated spin- glass behaviour. We have reinvestigated the crystal structure of UAuSi and observed some inconsistencies with the previous data. Herein we show that UAuSi adopts the hexagonal structure of LiBaSi, an ordered derivative of the AlB,-type structure. In addition we report in detail on the physical properties of UAuSi and UAu,. Experimental Starting materials were uranium platelets (Merck, 'nuklearr- ein'), gold wire (Degussa, >99.9%) and silicon lumps (Merck, >99.9%).The uranium platelets were cleaned with concen- trated nitric acid to remove oxide impurities and were then kept under argon. The samples were prepared by arc-melting the elements in an argon atmosphere. The argon was purified by repeatedly melting titanium sponge prior to the reactions. The molten buttons were turned over and remelted three times on each side to ensure homogeneity. The mass losses after several meltings were always <0.5%. The samples were sub- sequently wrapped in tantalum foil and annealed in evacuated sealed silica tubes at temperatures between 650 and 800°C for 2 weeks. Guinier powder patterns of all samples were recorded with Cu-Ka, radiation and a-quartz (a =491.30 pm, c =540.46 pm) as internal standard.Air-sensitive UAu, was ground to powder with dried paraffin oil and placed between two scotch tapes to prevent hydrolysis. The lattice parameters (Table 1) were obtained by least-squares refinements. The indexing was facili- tated by intensity calculations." Powder diffraction measurements of the UAuSi samples were performed on a STOE STADI/P focusing monochromatic beam diffractometer with a rotating very flat sample in the symmetric transmission mode. Cu-Ka, radiation was used with a linear position-sensitive detector, a step width of 0.02" (28) and a counting time of 20 s per step. The Rietveld refinements were performed with the RIETAN program.12 The magnetic susceptibilities of polycrystalline samples were measured in the temperature range 4.2-300K using an RH- Cahn electrobalance.Electrical resistivity measurements were performed in the temperature range 4.2-300 K with a conventional four-point technique. The sample voltage was measured automatically every 20 s with an accuracy of & 1 pV. The measurements were repeated with different samples, and the results were reproducible. Results and Discussion Crushed buttons of the samples are all light grey with a metallic lustre; powders are dark grey. The silicon-rich samples up to composition UAuSi are stable in air over several months; however, UAu, was less stable and showed some hydrolysis. These samples were kept under vacuum. The sensitivity against traces of humidity of such gold compounds was also observed for the corresponding thorium compound13 as well as for the alkali-metal gold compounds KAu, ,I4 NaAu,," RbAu16 and Rb,Au, .I7 Table 1 Lattice parameters of the hexagonal binary and ternary compounds in the pseudo-binary system USi, -UAu," compound structure type alpm c/Pm cla v/nm3 ref ~ ~ ~ ~~~~~~~~ p-USi, AlB, 402.8( 1) 385.2( 1) 0.956 0.0541 1 U,AuSi, AlB, 414.5(3) 398.9(2) 0.962 0.0593 6 UAuSi LiBaSi 419.5(2) 397.2( 2) 0.947 0.0605 this work UAu, AlB, 475.6(2) 311.0( 1) 0.654 0.0609 this work UAu, AlB, 475.4 310.7 0.654 0.0608 4 UAu, AlB, 475.6 311.0 0.654 0.0609 3 UAu, CeCd, 475.6 310.9 0.654 0.0609 5 " Standard deviations in the positions of the least significant digits are given in parentheses throughout the paper.J. Muter. Chem., 1996, 6(3), 429-434 429 Crystal structures The crystal structure of UAu, is confirmed The lattice para- meters [a=475 6(2) pm, c= 311 0(1) pm] are in excellent agreement with the previous data (Table 1) 'The powder patterns of our UAu, samples always showed the AlB2 struc- ture, independent of the annealing processes, in agreement with the work of Palenzona and Cirafici3 and Tran and Troc However, Dommann and Hulliger5 reported a CeCd,-type structure for UAu, In this structure type, the gold atoms are not situated on the mirror plane at z =O 5, they are shifted to zz0 45, resulting in slightly puckered layers We have calculated theoretical powder patterns with LAZY- PULVERIX" in order to check the two possibilities For the CeCd,-type structure two reflections at 28 =63 94" and 76 74" should occur with intensities of 65 and 48 (scaled at an intensity of 100 for the strongest reflection), respectively Both reflections are calculated with intensities of 0 1 for the A1B2- type structure None of our samples showed these reflections on the Guinier powder patterns However, for comparison, the 001 and 11 1 reflections with calculated intensities of 4 8 and 7 5, respectively, are clearly visible on the patterns We therefore conclude that UAu, adopts the A1B2-type structure (Fig 1) and there is no evidence for a puckering of the hexagonal gold nets in UAu, Positional parameters and interatomic distances are listed in Table 2 For UAuSi we find a different symmetry than that reported previously lo All our powder patterns (melted samples and those annealed at 650 and 800°C) show hexagonal symmetry and an intensity distribution resembling the A1B2-type struc- ture This result was also reproduced for other samples The refined lattice parameters of the sample annealed at 800 "C are a=4195(2) pm and c=3972(2) pm However, Tran and Troc" reported the TiNiSi-type structure for UAuSi (arc- melted sample annealed at 650°C for 7 days) with the ortho- rhombic lattice constants a =418 1 pm, b =798 2 pm and c = 724 1 pm These values correspond to the doubled orthohex- agonal Cell of AlB2 Uortho =ahex, bortho =2Chex and Cortho = ahex$ A reinvestigation of the previously reported'' powder data indeed showed that UAuSi has the small AlB,-like hexagonal cell (Fig 2) In order to establish whether or not the gold and silicon atoms are ordered within the hexagonal network, we have performed powder diffraction measurements on the arc-melted and the annealed samples The order can be detected directly from the sub-cell intensities, since the ordered structure has the translationenglezche18 subgroup P6m2 The crystallographic relationship for such order-disorder transitions has already Fig.1 Crystal structure of UAu, 0, uranium, 0, gold The two- dimensional gold network is outlined Table 2 Positional parameters and interatomic distances (in pm) of UAu, atom P61mmm X Y Z U la AU 2d u 12 2 6 Au 315 6(1) U 311 0(1) U 475 6(2) Au 3 Au 2 Au 6U 274 6( 1) 311 0(1) 315 6( 1) 430 J Mater Chem ,1996, 6(3), 429-434 Fig.2 Crystal structure of UAuSi projected along the z direction All atoms are situated on mirror planes at z=O(U) and z= 1/2 (Au, Si) connected by thin and thick lines, respectively been discussed in detail for isotypic ThAuSi l3 For the sample annealed at 800 "C, we clearly established the ordered LiBaSi- type structure l9 2o The intensities resulting from the Rietveld powder refinements are given in Table 3 Positional parameters and interatomic distances are listed in Table 4 In contrast, the collected powder diagrams of the arc-melted sample and the sample annealed at 650°C show two very similar hexdgonal cells Several reflections already showed some splitting on the Guinier powder patterns From this result we assumed that only a certain part of the samples is ordered The Rietveld powder refinements (assuming two phases) showed about 70% LiBaSi-type and about 30% AlB,-type structures in the sample annealed at 650°C The refined lattice parameters (powder diffraction data) were a =419 25( 3) pm, c=398 14(2) pm, V=O 0606 nm3 for the LiBaSi portion and a =420 79( 5) pm, c =395 98( 5) pm, V= 0 0607 nm3 for the AlB, portion As could be expected, the cell volume of the disordered AlB, part is slightly larger than that of LiBaSi The order within the LiBaSi-type structure expresses itself by the smaller lattice parameter u and a larger lattice parameter c We conclude from this refinement, that (1) the annealing temperature of 650 "C was slightly too low to achieve complete order, or (ii) the annealing time of 2 weeks was too short for this temperature Lattice constants The lattice parameters and the cell volumes of the compounds in the pseudobinary system USi,-UAu, are given in Table 1 As can be seen easily from these data, the a and c parameters increase slightly, when some silicon atoms in p-USi, are replaced by the much larger (metallic radii 131 9 pm for Si and 1442 pm for Au, both for coordination number, CN, 12),l gold atoms The increase of both is small up to UAuSi and then the situation is different While a rises dramatically up to UAu,, c behaves in the opposite manner Crystal chemistry A part of the pseudobinary system USi2-UAu2 has been investigated by X-ray powder diffraction All patterns show hexagonal symmetry The border phases p-US121 and UAU,~ ' adopt the AlB,-type structure For UAuSi we have determined an order between the gold and silicon atoms They form a hexagonal BN-like network The previously reported inter- metallic U2AuSi3 did not show any order between the Au and Si atoms, however, ordered structures for 2 13 silicides were recently reported for the compounds Ln,RhSi, ,22 23 Ln,Pds~,,~ and U2RuSi3 The silicon atoms in P-USi, are thus substitutable by gold Table3 X-Ray powder data of UAuSi (sample annealed at 800°C); the observed (I,) and calculated (I,) intensities for both refinements (disordered AlB,-type and LiBaSi-type) are listed together with the corresponding residuals R,(I)" ~~~~~~ ~ AlB, type LiBaSi type hkl ~~~~~~ ~~ 280,,/degrees dCdA I0 IC 10-Ic I0 I, Io-Ic 100 24.45 3.6334 22350 16314 6036 24324 24783 459 101 33.35 2.6823 100000 100567 567 100OOO 98485 1515 110 43.06 2.0977 35665 36527 8 62 33121 32943 178 002 45.55 1.9883 9600 9929 329 908 1 9086 5 111 49.03 1.8554 689 716 27 145 146 1 200 50.14 1.8167 2567 2676 109 3615 364 1 26 102 52.38 1.7442 4633 4693 60 4846 482 1 25 201 55.54 1.6524 17274 17515 24 1 17760 1769 1 69 112 64.49 1.4431 15893 14683 1210 15605 15138 467 210 68.21 1.3733 342 316 26 276 264 12 202 70.08 1.3412 2128 2016 112 2170 2086 84 003 71.03 1.3256 125 119 6 22 22 0 21 1 72.77 1.298 1 12497 11776 72 1 13461 12918 543 103 76.39 1.2453 5307 4853 454 5670 5373 297 300 78.96 1.2111 3419 3092 327 3706 349 1 215 301 83.31 1.1586 658 598 60 119 112 7 212 85.92 1.1300 2590 2372 218 1776 1663 113 113 86.82 1.1206 619 570 49 116 108 8 203 9 1.97 1.0708 2572 2404 168 2792 269 1 101 220 94.48 1.0489 1594 1503 91 1917 1847 70 302 96.24 1.0343 2929 2797 132 3545 3464 81 22 1 98.81 1.0142 474 468 6 95 93 2 310 99.67 1.0077 667 660 7 510 503 7 RB(I)=0.049 RB(I)=0.018 The final residuals for the refinement with the ordered LiBaSi structure are R,, =0.1025, RE=0.0652 and x2 =2.4714.Table 4 Positional parameters and interatomic distances (in pm, calculated from the Guinier data) for UAuSi atom ~6m2 X Y Z B/A2 U Au Si la If Id 0 213 113 0 1/3 213 0 112 112 2.2(7) 3.6( 16) 1.7(35) U: 6 Au 313.2(2) Au: 3 Si 242.2(1) Si: 3 Au 242.2(1) 6 Si 2 U 6 U 313.2(2) 397.2(2) 419.5(2) 6 U 313.2(2) 6 U 313.2(2) atoms up to the other border phase, UAu,.This substitution reveals not only a drastic change in the lattice parameters as discussed above, but also a change in the coordination spheres, resulting in distinct differences in chemical bonding. A compari-son of the interatomic distances in p-USi2, U,AuSi,, UAuSi and UAu, is given in Table 5. In p-USi, the U-U distances of 385.2 and 402.8 pm are similar.This is totally different when all the silicon atoms are substituted by gold atoms. The U-U bond lengths in UAu, amount to 311.0 and 475.6 pm, respectively. While both U-U distances in p-USi, may not be considered as direct U-U interactions, the small U-U contact of 311.0 pm in UAu, is certainly strongly bonding. Similar short U-U distances have also be observed in the intermetallics U,Mn3Ge,25 U,Ti,26-28 Table5 Comparison of the interatomic distances (in pm) in the structures of p-USi,, U2AuSi3, UAuSi and UAu, compound u-u U-Au/Si Au/Si-Au/Si B-USi, 385.2 302.0 232.6 402.8 U,AuSi, 398.9 311.5 239.3 414.5 UAuSi 397.2 3 13.2 242.2 419.5 UAu, 311.0 315.6 274.6 475.6 U3Si29 and UHg,.,' They are in the same range as in a-U (average U-U bond length of 313.7 ~m).~' The coordination of the uranium atoms in P-USi, consists of 12 silicon atoms.Owing to the short U-U contacts, this coordination number increases to 14 in UAu,; 12 Au and 2 U. The U-Au/Si distances increase from 302.0 pm (p-USi,) to 311.5 pm in U,AuSi3 and then increase only slightly to 315.6 pm in UAu,. The slightly larger U-Au distance in UAu, also reflects the larger coordination number of 14. The interatomic distances within the hexagonal network increase from 232.6 (p-USi,) to 242.2 pm (UAuSi) and then increase markedly to 274.6 pm in UAu,. The increase of only 9.6 pm (the difference of the metallic radii amounts to 12.3 pm2') from P-USi, to UAuSi reflects the strong bonding between the gold and silicon atoms.The Au-Si distance of 242.2 pm is about 12% smaller than the sum of the metallic radii. It is even smaller than the Au-Si bond lengths of 249, 249,252 and 246 pm in ScAuSi (ScAuSi-type), YAuSi (LiGaGe- type), LuAuSi (S~AuSi-type)~, and ThAuSi (LiBaSi-type, iso- typic with UAuSi),13 respectively. In the Au, rings of UAu,, the Au-Au distances of 274.6 pm are about 5% smaller than the Au-Au bond lengths of 288 pm in gold rnetaL3' The coordination number for the atoms in the hexagonal network increases from CN 9 in p-USi, (6 U +3 Si) to CN 11 in UAu, (6 U+3 Au+2 Au). Physical properties UAuSi. In order to derive the influence of atomic ordering of Au and Si atoms on the physical properties of UAuSi, the temperature dependence of the magnetic susceptibility and electrical resistivity was studied on three samples of UAuSi: (i) as-quenched, (ii) annealed at 650°C and (iii) annealed at 800 "C.We have measured x(T) in magnetic fields up to 0.6 T employing both the ZFC (zero magnetic field cooling) and FC (field cooling) conditions. It appears that these three samples show almost identical temperature dependencies of the mag- netic susceptibility, similar to that already reported in ref. 10. In Fig. 3 we present the magnetic data for UAuSi (iii) only. J. Muter. Chem., 1996, 6(3), 429-434 431 / 0 50 100 150 200 250 300 TIK Fig. 3 Temperature dependence of the reciprocal magnetic susceptibil- ity of UAuSi The left-hand inset shows the FC and ZFC susceptibility behaviour at B=O05 T The field dependence of the freezing temperature, T,, is shown in the nght-hand inset As seen from the inset, the FC susceptibility shows a weak saturation tendency at low temperatures, while the ZFC suscep- tibility exhibits a maximum at Tf (freezing temperature) Thus, the temperature dependence of the magnetic susceptibility for this compound appears to be strongly dependent on the sample cooling conditions The observed anomaly at Tf may indicate a magnetic phase transition, however, this maximum is broad and its magnitude is too large to be ascribed to a simple antiferromagnetic phase transition Such a maximum might signal a transition into an anisotropic ferromagnet, but the magnetization behaviour of UAuSi (iii) taken at various tem- peratures below 40 K in magnetic fields up to 4 T does not show characteristics of a ferromagnet As seen from Fig 4, the magnetization, a@), does not show saturation at 4 2 K At the highest obtainable magnetic field of B=4 T, only the very Fig.4 Magnetization us magnetic flux density of UAuSi at (a)42, (b) 20 and (c) 40 K 432 J Mater Chem , 1996,6(3), 429-434 small magnetic moment of p=O 19 pB(atom U)-' is observed However, at temperatures between 42 and 20 K, a small remanence is present in o(B) As suggested in ref 10, the x anomaly is thought to originate from spin-glass (SG) behaviour Therefore, we have investi- gated the irreversibility line from xzFc(T) and zFC(T)taken in several magnetic fields Defining the freezing temperature, T,, as the splitting point of the xZFCand zFCcurves, we have determined Tf as a function of the applied magnetic field (see Fig 3 inset) It is clear that Tf follows well the theoretical Almeida and Thouless line (AT line)33 in the low magnetic field limit, strongly supporting the SG-like behaviour for UAuSi The extrapolated T,value to B=O is about 17 K This value of Tf is almost the same as that found in our previous electrical resistivity measurements lo It should be mentioned that the AT straight line was predicted for the case of an Ising- type SG in the infinite-range, random-bond model33 On the other hand, as shown above, the magnetic properties of UAuSi are found not to depend on the degree of the atomic order and therefore such an interpretation for UAuSi is not quite adequate Also, at present it is difficult to claim that any non- stoichiometry of this compound is a possible reason of the SG-behaviour observed SG-behaviour is found for numerous uranium- and rare-earth-metal-based intermetallics crystallizing in hexagonal structures, such as CeCd,-type (UCUS~),~~AIB,-type (U2TSi3),7 and Ln(Al,Ga), 35 For all these hexagonal phases the XY-type mechanism seems to be responsible for the SG- formation Also, the existence of some randomness in the interactions between the U-U or Ln-Ln atoms, since a statisti- cal distribution of the non-magnetic atoms in the unit cell of a given compound can lead to the SG properties This mechanism has been postulated for CePd3B0336 and CePtGa, 37 For all samples of UAuSi, the magnetic susceptibility is reversible above T, and reaches almost the same values As seen from Fig 3, the x-'(T) function shows modified Curie- Weiss behaviour, yielding pexp=3 14 pB (U atom)-', @= 17 1 K and xo=O 3 x emu mol-' These values are in agreement with those reported in ref 10 In contrast to the x(T) behaviour, the temperature depen- dence of the electrical resistivity changes for different sample preparation methods The electrical resistivity measurements as a function of temperature, p(T), for the three samples of UAuSi are given in Fig 5 While the resistivity of the samples which are non-annealed or annealed at 650 "C increases with decreasing temperature in a Kondo-like manner, the resistivity of the sample annealed at 800 "C, I e of that showing full 1 a0921 ' ' ' ' ' 0 50 100 150 200 250 300 TIK Fig.5 Temperature dependence of the reduced electrical resistivity for the three different UAuSi samples The spin freezing temperature is indicated by arrows 1, as-cast, 2, annealed at 65OoC, 3, annealed at 800 "C crystallographic order, decreases distinctly with decreasing temperature (metal-like behaviour). Also important differences occur in the magnitude of p. The resistivity of samples (i) and (ii) is about an order of magnitude larger than that of sample (iii). These differences certainly reflect some variation in the degree of the crystallographic order of the Au and Si atoms.Nevertheless, there exists a common feature for all these samples, namely, the p anomaly occurring at about 17 K. This temperature should be associated with the Tf value deduced from the extrapolation to B=O T (see Fig. 3). Note that the occurrence of the resistivity maximum is one of the most distinctive features of the metallic-type SG.38 Thus, this behav- iour may be characteristic of samples (i) and (ii). UAu,. The low-temperature properties of UAu, have already been reported in a number of previous investigations of the U-Au system.4.39-41However, serious controversy concerning the magnetic properties observed for this compound still remains. For example, Ott et were the first group to report Pauli paramagnetism of UAu,, but they assigned the CeCd,-type structure for their sample.On the other hand, the annealed UAu2 sample had the hexagonal A1B2-type structure, tivity pointed to the existence of a weak ferromagnetic compo- 1 0 TIK Fig. 7 Temperature dependence of the specific resistivity of UAu, (annealed, 800"C, 2 weeks, B = 0 T). The inset shows the derivative dpldT us. T below 50 K. 2 r I 1 I 1I , .?01 and both the magnetic susceptibility and the electrical resis- --................... . . . ...... nent below 25 K4 Moreover, Kondo-type behaviour or spin- fluctuation effects have also been considered by Canepa et aL4' or recently by Kontani et respectively. In the present investigation we have used the annealed UAu, sample for the measurements. Its magnetic susceptibility us temperature curve is displayed in Fig.6. Above T= 50 K the susceptibility behaviour for this sample obeys the Curie-Weiss law yielding peXp= 2.98 pB(atom U)-' and 0= -190 K. These values agree well with those previously reported for a non- annealed sample4' but they are slightly different from those found for an annealed sample by Kontani et aL41As illustrated in Fig. 6, the susceptibility of our UAu, sample exhibits a pronounced upturn below T = 30 K. This behaviour, which is a consequence of oxidation, will be analysed in detail in a forthcoming paper.42 The temperature dependence of the electrical resistivity for the annealed sample of UAu, is displayed in Fig.7. This curve is similar to that recently reported by Kontani et aL4' p(T) shows a pronounced increase in resistivity at low temperatures, and then a broad shoulder followed by saturation in the high- temperature region. This result resembles the p( T) behaviour characteristic of materials dominated by spin-fluctuation effects.43 The temperature derivative of the resistivity, dp(T)/dT exhibits a distinct maximum at 18 K. Some tiny anomalies are also observed at 8 and 40 IS. They are probably due to impurities and were also observed by Kontani et 400 300 c 100 Lo -I 60 100 150 200 250 300 TIK Fig. 6 Temperature dependence of the inverse magnetic susceptibility of UAu, (annealed, 800 "C, 2 weeks, B= 0.7 T) 0 30 60 90 120 150 TIK Fig.8 Temperature dependence of the magnetoresistivity of UAu, (annealed, 800"C, 2 weeks, B= 1 T). The field dependence of the magnetoresistivity at 4.2 K is given in the inset. In order to clarify the nature of the anomalies observed for UAu, we undertook measurements of the electrical resistivity at B= 1T. It is interesting to note that the transition observed at 40 K starts to develop distinctly at B= 1 T, but no visual change has been observed in the transitions at 8 and 18 K, respectively. The magnetoresistivity, Ap/p = [p(T,lT)-p(T,O)]/p(T,O), measured as a function of temperature, is displayed in Fig. 8. At 4.2 K, the magnetoresistivity is negative and decreases linearly with increasing applied magnetic field without any tendency to saturation (see Fig.8 inset). As the temperature is increased, the Ap/p curve increases, starting from a value of -8%, goes through zero at about 15 K, and then reaches a distinct maximum at 18 K. At higher temperatures, this curve remains almost unchanged, showing a positive value. The negative magnetoresistivity observed at low temperatures is in agreement with some ferromagnetic character of the sample. We thank Professor Dr. W. Jeitschko for his interest and support of this work. We are also indebted to Dr. W. Gerhartz (Degussa AG) for a generous gift of gold metal and to the Fonds der Chemischen Industrie for a stipend to R. P. References 1 A. Brown and J. J. Norreys, Nature, 1959,183,673. 2 K. Remschnig, T. Le Bihan, H.Noel and P. Rogl, J. Solid State Chem., 1992,97,391. 3 A. Palenzona and S. Cirafici, J. Less-Common Met.,1988,143, 167. J. Muter. Chem., 1996, 6(3), 429-434 433 4 V H Tran and R Troc, 20iimes Journees des Actinides, Abstract P7, Prague, Apnl 17-20,1990 5 A Dommann and F Hulliger, J Less-Common Met ,1988,141,261 6 R Pottgen and D Kaczorowski, J Alloys Comp ,1993,201,157 7 D Kaczorowski and H Noel, J Phys Condens Mutter, 1993, 5,9185 8 R Pottgen, P Gravereau, B Darriet, B Chevalier, E Hickey and J Etourneau, J Muter Chem ,1994,4,463 9 B Chevalier, R Pottgen, B Darriet, P Gravereau and J Etourneau, J Alloys Comp ,in press 10 V H Tran and R Troc, Physzcu B, 1993,186-188,744 11 K Yvon, W Jeitschko and E Parthe, J Appl Crystullogr , 1977, 10,73 12 F Izumi, Rietveld analyses programs RIETAN and PREMOS and special applications, in The Rietveld Method, ed R A Young, International Union of Crystallography, Oxford University Press, 1993, pp 236-253 13 J H Albering, R Pottgen, W Jeitschko, R -D Hoffmann, B Chevalier and J Etourneau, J Alloys Comp ,1994,206, 133 14 U Zachwieja, J Alloys Comp, 1993,196,187 15 U Zachwieja, J Alloys Comp ,1993,196,171 16 U Zachwieja, Z Anorg Allg Chem ,1993,619,1095 17 U Zachwieja, J Alloys Comp ,1994,206,277 18 H Barnighausen, Commun Math Chem, 1980,9,139 19 H Axel, K H Janzon, H Schafer and A Weiss, Z Nuturforsch, Tell B, 1968,23,108 20 G Wenski and A Mewis, Z Anorg Allg Chem ,1986,535,110 21 E Teatum, K Gschneidner Jr and J Waber, Rep LA-2345, US Department of Commerce, Washington, DC, 1960 22 B Chevalier, P Lejay, J Etourneau and P Hagenmuller, Solid State Commun ,1984,49,753 23 R E Gladyshevskii, K Cenzual and E Parthe, J Alloys Comp 1992,189,221 24 P A Kotsanidis, J K Yakinthos and E Gamari-Seale, J Mugn Mugn Muter, 1990,87,199 25 R Pottgen, B Chevalier, R-D Hoffmann, W Jeitschko and J Etourneau, J Alloys Comp ,to be published 26 M C Udy and F W Boulger, Trans AIME, 1954,200,207 27 A G Knapton, J Inst Metals, London, 1955,83,497 28 B W Howlett, J Nucl Muter, 1959,3,289 29 G Kimmel, B Sharon and M Rosen, Actu Crystullogr, Sect B 1980,36,2386 30 R E Rundle and A S Wilson, Actu Crystullogr , 1949,2, 148 31 J Donohue, The Structures of the Elements, Wiley, New York, 1974 32 M L Fornasini, A Iandelli and M Pani, J Alloys Comp 1992, 187,243 33 J R L de Almeida and D J Thouless, J Phys A, 1978,11,983 34 V H Tran and R Troc, J Mugn Mugn Muter, 1990,86,231 35 A R Ball, D Gignoux, D Schmitt and F Y Zhang, J Magn Mugn Muter, 1992,1048~107,170 36 S K Dhar, K A Gschneider Jr ,C D Bred1 and F Steglich,Phys Rev B, 1989,39,2439 37 J Tang, K A Gschneider Jr ,R Caspary and F Steglich, Physicu B, 1989,163,201 38 P J Ford and J A Mydosh, Phys Rev B, 1976,14,2057 39 H R Ott, E Felder, A Schilling, A Dommann and F Hulliger, Solid State Commun , 1989,71, 549 40 F Canepa, A Palenzona and R Eggenhoffner, Physicu B, 1990, 160,297 41 M Kontani, T Nishioka, Y Hamaguchi, H Matsui, H A Katori and T Goto, J Phys SOC Jpn , 1994,63,3421 42 R Troc et a1 ,to be published 43 T Moriya, in Spin Fluctuations in Itinerant Electron Magnetism, Springer, Berlin, 1985 Paper 5/04446D, Received 7th July, 1995 434 J.Mater Chem., 1996, 6(3), 429-434

ISSN:0959-9428    

DOI:10.1039/JM9960600429

出版商:RSC    

年代:1996

数据来源: RSC