J. MATER. CHEM., 1994, 4(8), 1293-1301 Phases in the ZrxTal -x(O,N),, System, formed by Ammonolysis of Zr-Ta Gels: Preparation of a Baddeleyite-type Solid Solution Phase ZrxTal -xO1+xN1 0 <x <1 Jekabs Grins,* Per-Olov KaII and Gunnar Svensson Department of lnorganic Chemistry, Arrhenius Laboratory, Stockholm University, S-10697 Stockholm, Sweden Phase formation in the system ZrxTa, -x(O,N)y has been studied by ammonolysis of Zr-Ta gels, prepared by the sol-gel technique, at temperatures between 700 and 1000 "C. The starting gels and observed phases were characterised by X-ray powder diffraction (XRPD), scanning and transmission electron microscopy (SEM and TEM) and thermogravimetric (TG) analysis. Oxynitride phases of compositions Zr,Ta, -,O, +,Nl -,, 0 <{ d 1, with the baddeleyite-type structure, were prepared at 800 "C.The unit-cell volume increased linearly from 127.8 A3 for TaON (x =0) to 140.9 A3 for ZrO, (x=1). The structure was verified for the composition Zr,,,Tao,01.6No., (x =0.4) by a Rietveld refinement (R, =3.5%) using Cu-Ka, XRPD data. An orthorhombic oxynitride phase was observed in preparations at 700°C for 0.26dxdO.90 in Zr,Ta, -,01+,N, -,. Unit-cell parameters and powder X-ray reflection intensities agree with an orthorhombic ZrO, type structure. According to X-ray data, a cubic solid solution phase with a fluorite related subcell is present in materials prepared at 900 O0Cfor 0.26 5x d 0.68. However, electron microdiffraction patterns suggest a metrically monoclinic unit cell with a =6.1 A, b =14.1 A, c =7.1 A and p= 125".The Ta,N, type of structure was found to incorporate up to ca. 18 atomo%Zr at 900 and 1000 "C. Transition-metal nitrides with metallic properties and com- positions M,N, with n 61 are relatively well characterised with respect to structure and properties, while considerably less is known about more covalently or ionically bonded nitrides and oxynitrides containing one or more transition metals. A series of oxynitride perovskites have, however, been prepared, including ATa0,N with A=Ba, Sr and Ca,' BaNb0,N' and LnWO,N,-, (Ln=La and Nd).2 Compounds LnW0,N (Ln =Nd, Sm, Gd and Dy) having the CaWO, scheelite structure have also been ~ynthesized,~ as well as LnA1203N compounds (Ln=Nd, La and Sm) with the K,NiF,-type ~tructure.~Other rather recently reported oxynitride phases are LnA11,02,N compounds with the PbFe1,Ozg magnetoplumbite structure5 and BaCeLn(O,N), compounds (Ln=La and Ce) possessing the CaFe,O, structure.6 The increased interest in these types of compounds emanates from the obvious possibility to prepare new compounds with interesting thermal, electrical and magnetic properties.Two Ta compounds are presently known which contain nitrogen and Ta5+, viz. Ta,N, and TaON. The red Ta,N5 was reported in 18767 and can be prepared by ammonolysis of Ta205 at temperatures <1000oC.8 The crystal structure was determined by Strahle' by single-crystal X-ray diffraction and was later refined from neutron powder diffraction data obtained at 16 K." The orthorhombic structure is isotypic with pseudo-brookite Fe2Ti0511 and contains irregular TaN, octahedra sharing corners and edges. The yellow oxynitride TaON is formed as an intermediate phase when Ta205 is ammonolysed.8 The crystal structure was determined from powder neutron diffraction data recorded at 4.2 K', and was subsequently confirmed by single-crystal X-ray diffraction.', The structure is homeotypic with monoclinic ZrO,', and shows a complete ordering of 0 and N atoms in alternating layers perpendicular to the c axis.Ta nitrides with Ta in oxidation states below +5 are obtained when Ta205 is ammonolysed at temperatures above ca. 1000"C;15 Ta4N5 has an N?Cl-related tetragonal structure with a =6.831 A, c 7 4.269 A an! Ta,N, a hexagonal structure with a =5.176 A, C= 10.353 A.The brown non-conducting zirconium nitride Zr,N, was prepared by ammonolysis of ZrX, (X =C1, Br, I) at 700 OC;', its crystal structure is not known. Three Zr oxynitrides have been reported by Gilles and co-worker~;'~~'~ Zr,ON, (y), Zr708N4(p) and Zr701,N, (p'). The compounds were pre- pared at temperatures near 1000, 1100 and 2000°C, rcspect-ively, by ammonolysis of ZrO,, or by heat treating mixtures of Zr0, and ZrN in an NH, atmosphere. The crystal structure of Zr20N2 is allegedly isotypic with the C-type rare-earth- metal structure." Recent studies by Ohashi et showed, however, that Zr,ON, has a variable composition acci brding to the formula Zr02-2xN4x,3, with 0.5dxd0.8 at 950°C. Zr708N, and Zr,O,,N, are isotypic with Zr,Sc,O, and Zr5Sc2013,2' respectively, with structures that can both be derived from the fluorite structure by ordered omission of oxygen atoms.No Zr-Ta oxynitrides containing Zr4+ and Ta5+ havc been reported. The present work is a part of a programme aimed at synthesis of new transition-metal oxynitrides and nitrides. This study is an investigation of the possibility of synthesizing non-metallic oxynitrides containing both Zr and Ta by tiitrid- ation of appropriate oxide mixtures with NH, gas, i.e. by ammonolysis. Because the oxides ZrO, and Ta205 are unreac- tive at low temperatures, Zr-Ta xerogels were used as starting materials for the ammonolysis. The results from ammonolysis experiments on Ta-Zr xerogels of different compositions at temperatures between 700 and 1100"C are described below.Experimental A JEOL JSM-880 scanning electron microscope and a JEOL JEM-2000 FX-I1 transmission electron microscope with EDX (energy-dispersive X-ray) microanalysis systems LINK AN 10000 and LINK QX200, respectively, were used in character- isations of Ta-Zr starting gels and ammonolysed materials. Metal compositions were determined by averaging 15-20 EDX point analyses, The statistical errors in each EDX analysis were ca. 1 atom% in the SEM and ca. 4 atom% in the TEM. Approximate dimensions sampled in the analyses were ca. 0.5 pm in the SEM and G0.05 pm in the TEhL TG recordings were obtained with a Perkin-Elmer TGS-2 TG analyser, operated in air with a heating rate of 10"C min-'.The synthesized materials were characterised by their XRPD, recorded with a Guinier-Hagg camera, using Cu-Kq radiation and Si as an internal standard. The patterns were evaluated with a film-scanning system.22 The relative pro- portions of the phases present were estimated visually from reflection intensity ratios. XRPD data from baddeleyite-type were collected on a STOE STADI/P diffractometer, using Cu-Ka, radiation, a rotating sample in symmetric transmission mode and a linear position-sensitive detector covering 4.6" in 28. The step-length was 0.02" and the 20 range 15-92" was covered over a period of 72 h, yielding an intensity of ca. 8500 counts for the strongest Zro,4Tao&1.4No,6 peak.Preparation and Characterisation of Zr-Ta Gels Xerogels containing Ta and Zr were prepared with TaC1, and Zr n-propoxide as precursors in a dry-box containing an oxygen- and water-free atmosphere. Appropriate amounts of TaC1, (Merck p.a.) and a calibrated 70 wt.% solution of Zr n-propoxide in propanol (Merck) were weighed into 100 ml beakers. The beakers were sealed with Parafilm and removed from the dry-box. Dry ethanol (ca. 25 ml) was then added to each beaker and the clear solutions were stirred for 10min. The solutions were then hydrolysed by rapid injection of water. The amount of water corresponded to twice the amount of alkoxide present. Hydrolysis was carried out both under prevailing acidic conditions and under basic conditions.The latter was achieved by adding NH, to the water. The basic solutions gelled rapidly, whereas viscous oils were produced under acidic conditions. Xerogels were then obtained by evaporation of the solvent on a hot-plate. An SEM image of a typical gel, hydrolysed under basic conditions, is shown in Fig. 1. The gel exhibits a large surface area and is composed of agglomerates of granules of ca. 0.1 pm. The higher magnification image shows that these granules are composed of particles of ca. 50-100 A. This fine structure is more distinctly revealed in TEM images, as shown in Fig. 2(a).Gels hydrolysed under acidic conditions exhibited fragmen$ resembling crystallites, ranging from ca. 1 pm down to 100A.These fragments did not show the kind of fine structure observed in gels hydrolysed under basic conditions, as seen in Fig. 2(b). The gel morphology thus depends, as expected, on the hydrolysis procedure applied. The two types of gel produced no electron microdiffraction patterns and were thus amorphous. The materials formed when ammono- lysing the gels were in some instances conditioned by the type of gel used, as described below. Gels with the nominal compositions Zro.18Tao.82, zr0.68Tao.32 (made under acidic hydrolysis conditions) and Zro.60Tao.40 (made under basic hydrolysis conditions) were analysed by EDX in the SEM. The analyses showed that the gels were homogenous and that they contained substantial amounts of C1 from the use of TaCl, as precursor.No EDX signal from possible residual C1-could, however, be detected after ammonolysis of the gels. Elemental analysis of each gel yielded the compositions Zro.19~l~Tao.81~l~ ( 19 atom% 'l), zr0.70(1)Ta0.30(1) (28 C1) and zr0.63(1)Ta0.37(1) (21 atom% CI), respectively, with standard deviations in parantheses. The compositions agree well with the nominal ones. Corresponding EDX analyses of the last two gels in the TEM yielded the compositions: Zro.63(33Tao,37(3) and Zro.71(4)Tao.29(4),in good agreement with the SEM analysis. The TEM studies also showed that Zr and Ta were homogen- ously distributed. Typical TG curves showing the decomposition of the Zr-Ta J. MATER. CHEM., 1994, VOL. 4 Fig. 1 SEM micrographs of a typical Zr-Ta gel hydrolysed under basic conditions (composition Zro.6Tao,4) gels are shown in Fig.3. The gels hydrolysed under basic conditions showed weight losses of up to 55%, decreasing with Zr content. Gels hydrolysed under acidic conditions showed smaller weight losses, up to 20%, increasing with Zr content. The weight loss was found to take place at lower temperatures for the gels hydrolysed under basic conditions, up to ca. 5OO0C, compared with the gels hydrolysed under acidic conditions, up to ca. 800°C. Ammonolysis of Zr-Ta Gels The Ta-Zr xerogels were heat-treated in flowing NH, gas at temperatures between 700 and 1100"C. The reaction chamber consisted of a silica tube, length 1.5 m, id 3 cm, placed in a horizontal tube furnace.Gas connections for NH, and N2 were provided by glass fittings. The quartz tube was flushed with N, for ca. 1h before the ammonolysis. The ammonia gas was dried by passing it through a molecular sieve and CaH,. The gels were placed in small Pt or Al,O, cups, each J. MATER. CHEM., 1994, VOL. 4 Fig. 2 TEM micrographs of a Zr-Ta gel hydrolysed under (a) basic conditions (composition Zro,,Tao.4) and (b)acidic conditions (composition ZrO.52Tao.48) containing 20-50 mg. The temperature was measured with a Pt/Pt-Rh thermocouple near the sample. An NH, gas flow of 10-20ml s-' was used and the ammonolyses were ended by cooling the samples to ca. 100cC,in the furnace. Results The Ta-Zr gels were ammonolysed at 700, 800, 900, 1000 and 1100'C.Ammonolysis times of 12 and 96 h in most cases yielded similar results. The results from the 96 h runs are described below. Comparisons with runs using shorter ammonolysis times are given in cases where significantly different results were obtained. The nominal gel compositions used were ZrxTalPx with x=0,0.13, 0.18,0.26, 0.33,0.40,0.52, 0.60, 0.68, 0.80, 0.90 and 1: the sloping x values correspond to gels hydrolysed under acidic conditions. The samples in the 96 h runs were heated in an NH, atmosphere at a rate of ca. 200°C h-' to 7OO0C, prereacted at this temperature for 24 h, and then heat-treated at the various final tempcratures for 72 h. The XRPD patterns in general exhibited broad Bragg peaks with half-widths of 0.3-0.8" in 26, at 28 =50".TEM studies of the ammcnolysed gels revealed the presence of grain sizes down to 50 A, implying that the broad peaks can be attributed mainly to the presence of small crystallites. The phase relations observed at different preparation temperatures are summar- ised in Fig. 4. rI uu h80 8 \v E .cn 1P 40 I I I I I I 0 200 400 600 800 1000 T/"C Fig. 3 TG curves for the thermal decomposition of the Zr-Ta gels in air. Acid hydrolysis conditions: (a) x =0.18, (b) x =0.68; basic hydrolysis conditions: (c) x =0.80, (d)x=0.13. 00to 000 9$) Oooooo Oo0 11~1111111111 0.0 0.2 0.4 0.6 0.8 1.o x in Zr,Ta,-, Fig. 4Schematic illustration of observed phases at different prep- aration temperatures.Estimated relative amounts of the phases are indicated by the sizes of the symbols. 0, Ta,N,; A, Zr,O,N,; V,Ta4N,; A,Ta,N,; 0, baddeleyite; 0, cubic phase; 0,orthor-hombic phase; +, cubic phase. Phase Analysis 700 "C The samples with x G0.18 were biphasic, containing a baddele- yite-type phase, i.e. iso- or homeo-typic with monoclinic ZrO,, and small amounts of a phase isotypic with Ta,N,. Both the baddeleyite- and Ta,N,-type phases showed a cell expansion with increasing nominal Zr content and are thus solid solution (ss) phases. Ammonolysis performed for shorter times, e.g. 12 h, yielded materials containing only the baddeleyite phase. This indicates that the baddeleyite phase is formed first and then partly transformed to the Ta,N,-type phase.Preparations with 0.26 <x <0.90 consisted of mixtures of the baddeleyite phase and an ss phase, whose powder pattern could be indexed on bas@ of an orthorhombic unit cell with all cell axes around 5 A. Further characterisation of the orthorhombic phase is given below. The relative amounts of the two phases varied unsystematically with x. Preparations with x =0.52 and 0.68 were green-grey and contained mostly the orthorhombic phase, while the other preparations were light yellow and contained the baddeleyite phase as the major phase. The samples which contained mainly the orthorhombic phase emanated from Ta-Zr gels which had been hydrolysed under acidic conditions. The sample with x= 1.0 contained baddeleyite.J. MATER. CHF,M., 1994, VOL. 4 800 "C Preparations with xG0.18 consisted of mixtures of the baddeleyite-and Ta,N,-type phases. The amount of Ta,N,-type phase was larger than in the 700°C series but decreased with increasing x. Shorter amnionolysis times yielded, as above, monophasic samples of the baddeleyite phase. Monophasic samples of the baddeleyite phase were obtained for 0.26 <x < 1.00. The powder patterns for s=0.52 and 0.68 exhibited, however, one faint reflection from the orthorhombic phase appearing at lower preparation temperatures. The samples were of different colours, ranging from yellow or beige for 0.26 <x <0.40, green-grey for x =0.52 and 0.68, dark brown for x =0.60 to grey for x >,0.80.900 "C Preparations with x 60.18 were monophasic and yielded powder patterns from an Ta,N,-type phase. The powder patterns for 0.26 <x <0.80 showed the presence of two phases: an ss phase with apparent cubic symmetry and small amounts of the Ta,N,-type phase. The reflections of the cubic phase were considerably broader than those of the other observed phase, with half-widths of ca. 1" in 20. The :trong reflections could be indexed using a cell with az5 A. The intensities of these reflections showed the structure to be related to fluorite. Additional superstructure reflections were, however, present for compositions with 0.26 ,<x <0.68. This phase is further characterised below. The unit-cell dimensions of the Ta,N,-type phase were found to be constant in the biphasic materials.The sample with x =0.90 contained two phases, a baddeley- ite phase and a smaller amount of the cubic phase. The unit- cell volume of the baddeleyite phase showed that its composi- tion was close to x= 1. The unit-cell volume of the cubic phase could not be reliably assessed owing to overlap of all reflections with those of baddeleyite. The sample with x= 1.0 contained only baddeleyite. 1000"C The results obtained at this temperature were similar to those at 900°C. For xG0.18, the materials were thus found to be X-ray monophasic and consist of an Ta,N,-type phase. Biphasic materials were obtained for 0.26 dx <0.68: they contained an Ta,N,-type phase and the same cubic phase as described above.The colours of the biphasic materials showed no systematic variation with x.The majority of these samples had a grey to green-grey bulk colour and a brown surface colouration. The powder reflexions of the cubic phase were even broader than those found for the preparations at 900 "C. The apparent amount of the cubic phase did not increase systematically with x. Preparations with starting gels hydro- lysed under acidic conditions contained higher fractions of the cubic phase. The unit-cell volume of the Ta,N,-type phase increased with the nominal content of Zr in the single-phase region x G0.18, whereas approximately constant unit-cell volumes were observed for the biphasic materials with 0.26 6 x<0.68. The unit-cell volumes of the cubic phases were larger than those recorded for the 900 "C preparations.The sample with x=0.80 was orange-red and contained, in addition to the Ta,N,-type phase and the cubic phase, a small amount of a baddeleyite phase. The baddeleyite phase had a unit-cell volume corresponding to an estimated com- position near x= 1, i.e. ZrO,. The powder pattern for x =0.90 was the same as that observed in the 900°C series, showing the presence of a baddeleyite phase and a small amount of a cubic phase. The powder pattern for x= 1 showed only baddeleyite. J. MATER. CHEM., 1994, VOL. 4 1100 "C All preparations were black or very dark brown. The sample with x=O contained Ta,N, and a smaller amount 9f Ta,N,. The pbserved cel! parameters are a =5.178(1)A, c = 10.346( 6)0A, V =240.! A3 for Ta,N, and a =6.842( 2) A, c =4.266( 4) A, V=199.7A3 for Ta,N,, in fair agreement with the reported values,', given above.Preparations with 0.13<x 60.68 contained four phases: comparatively large amounts of Ta,N, and a baddeleyite phase, plus smaller amounts of Ta,N, and the cubic phase observed at lower temperatures. The unit-cell parameters of Ta,N, and the baddeleyite phase showed no variation with x. The unit-cell parameters of the baddeleyite phase were close to those of ZrO, and the parameters of Ta,N, were negligibly larger than those observed for x=O. The amount of the cubic phase varied unsystematically with x and was, as for preparations at 900 "C, higher for gels hydrolysed under acidic conditions.The samples with x =0.80 and 0.90 contained the baddeley- ite phase. a small amount of Ta,N, and a cubic phase with a face-centred cell with a =4.313(2)A. This unit-cell parameter and observed reflection ipensities are similar to those reported for cubic TaN, a=4.33 A (JCPDS no. 32-1283). The sample andwith x= 1 contained Zr708N4 (/I)a small amount of baddeleyi te. The observe! hexagonal c~ll parameteTs of Zr,O,N, are a=9.549(2) A, c=8.815(4) A, V=696.1 A3, in general agreement with the parameters given in ref. 18. A schematic illustration of the phases observed at different preparation temperatures is given in Fig. 4. Baddeleyite-type Phase Zr,Ta, -xO1+ xN,-0 d x <1 A phase with the baddeleyite (monoclinic ZrO,) type structure was obtained essentially X-ray monophasic for all composi- tions between the end-members TaON and ZrO, by ammon- olysis of the Ta-Zr gels at 800°C for 12 h.The stryngest reflection from the orthorhombic phase, 11 1 at dz 2.95 A, was present in the powder patterns for x=O.52 and 0.68. As described above, increasing the ammonolysis time resulted in the formation of the Ta3N5-type phase for compositions with x60.18. The unit-cell parameters obtained for the end-members TaON and ZrO, agreed well with literature data.l43l5 A linear increase in cell volume of ca. 10% is observed when going from TaON to ZrO, (Fig. 5). The increase is a consequence of the replacement of Ta5+ ions by larger Zr4+ ions, albeit counteracted to a small extent by the simul- taneous substitution of N3-ions by smaller 0,-ions.The Shannop-Prewitt ionic radii23 for the ions ar:: Ta5+ VII)=0.78 A, 0,-(IV) = 1.38 A and(VII)=0.69 A, 0Zr4+( N3-(IV)= 1.46 A. The a, h and c axes increase in a similar, essentially linear, manner with Zr content, as shown in Fig. 6. Deviations from linear behaviour are, however, discernible for 0.7<x <0.9; they might arise from different ordering of N3-and 02-ions (see below). The monoclinic B angle showed only a slight linear variation between the end compositions. The Ta:Zr ratios in the baddeleyite phase preparations were verified for two compositions, with nominal x values of 0.68 and 0.52, by EDX(SEM) analysis. The analyses yielded the experimental values x =0.66( 1) and 0.53( l), respectively, in good agreement with the nominal values.The thermal stability of the baddeleyite compounds in air was examined by TG heating runs (Fig. 7). The compounds were oxidised at elevated temperatures according to: Zr,Ta, -xO1 -x +3/4( 1 -x) 0, =ZrxTa1-x05,2-x,2+1/2( 1 -x)N2 1201 / II I I I II 0.0 0.2 0.4 0.6 0.8 1.0 x in Zr,Ta,-, Fig. 5 Unit-cell volumes of ss phases us. x in Zr,Ta,-,; e,baddele-yite-type phase; V,orthorhombic phase; 0,cubic phase; V. ortho-rhombic Zr0,;27 N, cubic Zr0214 5.4 5.3 5 c.$ 5.2 E E! (dQ --8 5.1 5.0 4.9 I I 1 I I I. 0.0 0.2 0.4 0.6 0.8 1.0 x in ZrxTal -fll+PI-cFig. 6 Cell parameters a (O), b (H), (V) t's.x for baddelFyite ZrxTa, -,O, +,N-,. EstimatFd standard deviations are <0.004 A for 0.26dx d 0.80 and <0.001 A for the remaining compositions. 6rc I I I __I 0 300 600 900 1200 TIOC Fig. 7 TG curves for the oxidation of baddeleyite phases Zr,Tal-,Ol+xN1-x in air. Heating rate 10°C min-'. (a) 'TaON, (h)x =0.34, (c) x =0.52, (d) x =0.68. For TaON (x=0) the oxidation effectively takes place in the temperature range 760-880 "C, although a small weight increase is noted to commence at cu. 620°C. For higher x values the oxidation is observed to begin at lower tempera- tures, cu. 500"C, and to occur over an increasingly broad temperature range with increasing Zr content. The observed increase in weight agrees well with values expected for complete oxidation of the nominal oxynitride compositions.The 0 and N contents for x=0.40 and 0.60 were also determined by the combustion method. Assuming nominal Ta: Zr ratios, the results of the analyses were con- J. MATER. CHEM., 1994, VOL. 4 The Rietveld refinement demonstrates that the Zro,4Tao,601.4No,6compound has a structure similar to the MX2 baddeleyite type. The baddeleyite structure is illustrated in a (100) projection in Fig. 9, with atomic coordinates for Zro~4Tao,601.4N0,6.Each M atom is coordinated by seven x atoms, a triangle of X( 1)atoms and a planar square group of X(2) atoms. The two planes defined by the X(1) and X(2) atoms, respectively, are roughly parallel and the structure displays layers of X( l)-M-X(2)-M parallel to (100).The X( 1) atoms are triangularly, and X(2) tetrahedrally, coordinated by M atoms. In the homeotypic structure of TaON, the 0 and and N atoms are completely ordered, with 0 atoms on X( 1)sistent with the compositions ~ro~40~ao,60~1~50~1~~o~5~~l~ Zro,60Tao,4001.69~1~No~37(1~,respectively. The N contents agree well with those obtained from the TG experiments, while the 0 contents are higher than expected. However, since the nominal and experimentally determined Ta :Zr ratios are found to agree well; the combustion analyses thus yield a total anion content which is too high. The determined 0 contents hence seem to contain a systematic error. Rietveld Refinement of Zro.,Ta,60,,No, having the Baddeleyite-type Structure The baddeleyite-type structure of Zro,4Ta,,60,.4No,6 was refined, from XRPD data, using a Rietveld package.24 The final refinement was carried out with a total of 22 parameters and using the pseudo-Voigt profile function (refined Lorentzian fraction =0.66).The number of theoretical Bragg reflections for 28<92" was 127 and the half-width of the peaks was 0.50" in 28 at 28=48". The X-ray data did not allow any distinction between different possible orderings of 0 and N atoms, and these were accordingly statistically distributed over the two available anion sites. Fig. 8 shows the fit between the calculated and observed patterns. A list of atomic coordinates is given in Table 1, with the esds multiplied by 5.5 in order to account for serial ~orrelation.~~An absorption correction, of the form exp(-pR sin O), yielded no significant changes in atomic coordinates and thermal parameters 1.7(1) and 1.3(2)A for the (Ta,Zr) and (0,N) atoms, respectively.10000 8000 v) -c2 6000s-.-4000 a,+ (I._ 2000 0 I 20 I 30 I 40 I 50 1 60 I 70 I 80 LJ 90 2Wdegrees Fig. 8 Observed and difference intensity X-ray patterns of zr0.4Ta0.601 .4N0.6 Table 1 htomic coordin$es for Tao.6Z~o.401,,No.6; monoclinic, (I= 5.043(2) A, b=5.101(2) A, ~=5.243(2)A, fi=99.49(2)" P2,/c, Z=4 atom x (Ta,Zr) 0.287( 1) (O,N)(1) 0.07(1) (O.N)(2) 0.44( 1) R, =4. 7 Yo, R,, =6.0Yo, R, Y Z B/A 0.0417(6) 0.33(1) 0.76( 1) 0.213( 1) 0.34( 1) 0.47 ( 1) LO( 1) 0.3(8) 0.3(8) =4.1Yo, R, =3.5Yo.sites and N atoms on X(2) sites, resulting in a layer sequence of N-Ta-O-Ta. The atomic array formed by the X(2) layer and the M atoms directly above and below this plane is the same as the atom arrangement in the fluorite (CaF,) or cubic zirconia structure. The major dissimilarity between the badde- leyite and fluorite structures is thus a different arrangement of atoms in the X( 1) layer. The M-X distances in the Tao,6Zro,401,4No.6 baddeleyite phase, M =(Fao,6Zro,4)X =(Oo,7No.3)2range between 2.02( 5) and 2.21(5) A, the average being 2.12 A. The average distances to the three X( 1) and four X(2) atoms can be compared with corresponding average distances in monoclinic Zr0,14 and TaON:" TaON (x =0) Zr,,,Ta, 6014No6 ZrO, (x = 1) average M-X/A average M-X( 1)/A average M-X( 2)/A 2.09 2.07 2.11 2.12 2.09 2.15 2.16 2.09 2.21 The average M-X( 1) distance is comparatively constant in the three compounds, while the average M-X( 2) distance increases with x.This implies a relative displacement of the M atoms towards the X(2) layer with decreasing x. For TaON this displacement has, together with observed short N-N distances, been interpreted12 as evidence for considerable covalency in the Ta-N bonds. The distances observed in ~ao,6~ro~4~l~4~o,6accord with a linear variation of the average M-X(2) distance with x, which would indicate that the N atoms in ~ao,6~ro~4~l~4~~~6 are located only on the X( 2) sites.This possibility cannot, however, be verified by the Rietveld analysis. Neutron data would help to distinguish possible ordering of 0 and N atoms and also increase the precision of the anion positions. Fig. 9 Baddeleyite (monoclinic ZrO,) structure projected on (100). The numbers represent the x coordinate in %. Anion positions X( 1) and X(2) are illustrated by unfilled and shaded circles, respectively. J. MATER. CHEM., 1994, VOL. 4 Orthorhombic Zr,Ta, -xO1+,N,-x Phases Materials obtained by ammonolysis of xerogels at 700°C were biphasic for 0.26 d x d 0.90, containing the baddeleyite- type phase and a phase exhibiting a powder pattern that could be, indexed with an orthorhombic unit cell with a,b,c=5 A. The orthorhombic phase was observed as the major phase only for +=0.52 and 0.$8, with unit-c$l parameters a=4.908(2) A, b=5.265(2) A, c=5.138(2) 4, 1/=132.8 A”for ~=0.52~anda=4.960(1)A, b=5.267(1)A, c =5.110(1)A, V= 133.5 A3 for x =0.68.The indexed powder pattern for x =0.68 is given in Table 2. The unit-cell volumes are smaller than those of corresponding baddeleyite phases (see Fig. 5). Two similar, alternative, structural models, an ortho-rhombic ZrO, type structure and an a-PbO, type structure, were both found to yield good agreement between calcu- lated and observed intensities for the orthorhombic phase. Orthorhombic ZrO, is a high-pressure modification, found in considerable quantities in transformation-toughened ceramics cooled to low temperatures.,, The unit-cell volume of this ZrO, modification is 3.3% smaller than that of the monoclinic modification.The orthorhombic ZrO, structure27 is very similar to the monoclinic baddeleyite modification; in both of them there is a coordination of Zr atoms by seven 0 atoms. The spaceogroup is Pbc2, ?nd the unit-Celt parameters are a =5.068( 1) A, b =5.260(1)A, c =5.077(1)A and V= 135.34A3. Similar unit-cell volumes and parameters are also found for oxides adopting the a-Pb0,-type structure.28 The a-Pb0,-type structure, with space group symmetry Pbcn, is commonly idealised as an hcp structure, with M atoms in edge-sharing MX, octahedra that form zigzag strings in the c direction. The structure can, however, be topologically trans- formed into the fluorite structure29 and consequently com- pounds with a-Pb0,-type structures are frequently found to be intermediate between the idealised a-Pb0, and the fluorite structure. Three sets of XRPD data were collected for the orthorhom- bic type phases.Rietveld refinements of the orthorhombic ZrO, structure model converged to RF=4%o for all three sets of data. However, only slightly higher RF values, ca. 5%0,were obtained using the alternative a-PbO, structure model. The two structure models are, in principle, distinguishable by hkO Table 2 Observed and calculated 28 values for the Guinier-Hagg diffraction pattern of orthorhombic Zro,,,Tao,320L,8No,32 up to the 20th observed line [A8 =28,,, -28,,,, i= 1.54098 A the corresponc- ing cell parameters are a =4.960( 1) A, b =5.267(1)A, c =5.1 10( 1) A] hkl 2O,,,/degrees A0/degrees dobs/A If10 110 24.607 -0.027 3.615 4 111 30.284 -0.027 2.952 100 020 34.039 0.022 2.632 12 002 35.069 -0.028 2.557 13 200 36.172 -0.016 2.48 13 10 02 1 38.419 -0.006 2.3412 3 112 43.335 -0.015 2.0863 2 022 49.707 0.026 1.8327 12 220 50.517 0.008 1.8052 16 202 51.258 -0.040 1.7809 14 22 1 53.820 0.012 1.7020 3 130 55.501 0.023 1.6544 4 131 58.589 0.006 1.5743 9 113 59.990 -0.017 1.5408 11 311 61.469 0.005 1.5073 12 222 63.015 0.024 1.4739 3 023 132 65.151 67.362 -0.028 -0.020 1.4307 1.3894 1 004 74.197 0.023 1.2770 3 04 1 74.328 -0.006 1.2751 2 reflections with h +k # 2n, which are allowed in Pbc2, for orthorhombic ZrO,, but systematically absent in Pbcii for an a-Pb0,-type structure.These reflections are, however, very weak for an orthorhombic Zr0,-type structure, and the presence or absence of them could not be unambiguously established. The correct structure could thus not be singled out on the basis of the X-ray intensity data. Further structural studies are in progress to settle the correct structure of the orthorhombic phases. A strong argument favouring the orthorhombic ZrO, model is provided by the observed unit-cell parameters. The observed unit-cell parameters of the orthorhombic phase are depicted in Fig. 10, together with the reported cell parameters for a number of a-Pb02-type structures.3w35 For the a-Pb0,-type structures the orthorhombic cell parameters decrease in the order b>c>a.Cell axis ratios reported in the litcrature range from 1.09 to 1.12 for b/c and 1.05 to 1.08 for c, a. The orthorhombic Zr,Ta, -,01+,N1 -,phases exhibit cell axis ratios which are much closer to unity, b/cz 1.02-1.03 and c/a~~l.02-1.05,and are similar to the axis ratios for ortho- rhombic ZrO,, b/c =1.04 and c/aFZ 1.00. Additional studies currently in progress, indicate that this orthorhombic phase is stabilised by additions of small amounts, 3-7 mol%, of Co, Cr and Fe. The unit-cell param- eters of these stabilised phases conform well with those given above for the Zr,Ta, -xO1+,N1-,phases. Cubic and Pseudocubic Zr,Ta, -x(O,N)y(y d 2) Phases For 0.26 d x <0.80, ammonolysis at 900 “C yielded biphasic materials.According to X-ray analysis these contarned a apparently cubic ss phase as major phase, together with small amounts of an Ta,N,-type phase. The same phase was also found, in smaller amounts, together with a baddeleyite phase in samples with x=O.9, prepared at 900 and 1000°C. The powder patterns of the phase were basically the same ;is that of the cubic fluorite type. The reflections were broad, which hampered the evaluation of the film data. The volumes of the fluorite type cells are shown in Fig. 5. No superstructure reflections were observed for x =0.80 and 0.90, and the structure thus appears to be of the cubic zirconia type for these compositions.The weight increase upon oxidation for the sample with x =0.80 corresponded, however, to a metal anion composition MX1.71. This suggests that the phase may be a disordered relative of the rhombo- hedral Zr oxynitride p phase, Zr,08N,. 4.41 1 I I J 0.60 0.65 0.70 0.75 average cationic radius/A Fig. 10 Cell parameters us. average Shannon-Prewitt cationic radius for orthorhombic ss Zr-Ta oxynitride phases (U), orthorhombic ZrOZ2’ (E) and selected cc-PbO, type phases; (i) Ti02,30 (ii) Taq. 9Feo.9Zn0. 0,.*F0.2,3 (iii) (Zr0.33Tio.67 )204,32 (iv) Hfli04, 33(v) ZrT10,,3~ (vi) ZrSno~,Ti,,,0,35 The powder patterns for 0.26 dx d0.68 contained five or six weak superstructure reflections, which could be indexed using a body-centred cubic cell with a doubled cell axis.The indices and observed intensities in YOof the strongest reflection are; (211) 2-4%, (411) 1-3%, (332) 1%, (541) 1-3% and (631) 1%. Powder patterns of compounds with the C-type rare-earth-metal structure, e.g. Zr,ON,, exhibit medium-intensity reflections at corresponding reflection positions. The observed intensities are, however, apparently much too weak in order for the Zr,Ta, -x(O,N)yphases to have this structure. The weight increase upon oxidation for x= 0.33,0.52 and 0.68 was found to correspond to a metal anion composition of MX1,85-1.87.These anion contents are likely to be somewhat too low, owing to the presence of small amounts of the Ta,N,-type phase. Electron microdiffraction patterns of this phase suggested two orthogonal cell axes of ca.14 and 7 A. Similar axes have been reported for the unit cell of the reduced and metastable rare-earth-metal oxide Tb,,0,036 or Tb01.875. The unit cell of Tb1,030 is metrically monoclinic and related to the fluorite cell a,, bf, cf by; a = a, + bf/2-cf/2 b=2bf+2cf c=-b f+Cf If a, = 5.00 A,the monoclinic cell parameters become a = 6.12, b= 14.1, c=7.07 A and b= 125.3". Electron microdiffraction patterns for the Zr,Ta, -x(O,N)y phase with x = 0.52 were found to be compatible with this type of unit cell. It appears plausible that the Zr,Ta,-,(O,N), phase has a metal anion composition MX1.875,as does Tb,6030, since the same type of unit cell is indicated for the phases.This implies that the unit cell contains two anion vacancies. A structural model may be arrived at by assuming that the two vacancies form pairs across metal atoms (along 1/2 [ ill],), a common vacancy arrangement in reduced rare-earth-metal oxides.37 Such a model cannot be verified presently, however, for lack of sufficient XRPD data. Ta,N,-type Phases X-Ray monophasic samples containing an Ta,N,-type phase were obtained at 900 and 1000°C for xG0.18. The colours changed from vermilion for x=O to red-brown for x=O.18. The unit-tell parameters qf Ta,N, were qbtained as a,= 3.8900(4) A, h= 10.224(1)A, C= 10.273( 1) A, V=408.6 A3. The unit-cell parameters were found to increase, as expected, with Zr content, but also to depend on the preparation temperature.When prepared at 900 "C, thc composition x; 0.18 thus yielfled a cell wiih u=3.919( 1) A, b= 10.258(2) A, c= 10.338( 1) A, V=415.6 A,, whereas the preparation at 1000 "C agave significanLly smaller celj parameter?: a = 3.909( 1) A, b= 10.244(2) A, c= 10.317(2) A, V=413.2 A,. The recorded weight gain upon oxidation for the latter sample was 8.0 wt.%. This agrees well with a calculated value of 7.9 wt.% for a composition Zro,53Ta2,47N4,4700,53,which accords with a replacement of Ta+N for Zr+0. The prep- aration at 900°C yielded a smaller weight gain, however, 7.5 wt.%. These observations suggest that the Ta,N,-type phases have varying anion compositions (cJ:Discussion). Discussion Ammonolysis studies of materials with mixed Zr-Ta composi- tions have not been reported earlier and the present results can thus only be compared with previous findings for com- pounds containing exclusively Ta or Zr.Considering the J. MATER. CHEM., 1994, VOL. 4 various parameters that may influence the formation of phases during ammonolysis, it is not surprising that somewhat differ- ent observations are made in different studies. Brauer et a!.,' when ammonolysing Ta205, found that the formation rate of Ta3N5 is dependent on temperature, flow rate of NH,, the efficiency of the removal of 0, and H20 from the reaction zone, and the amount and presynthesis of Ta,O,. During ammonolysis, the oxygen activity in the sample is, furthermore, expected to change continuously as oxygen is withdrawn.In the present study, the ammonolysed materials were cooled inside the furnace at a comparatively slow rate, introducing the possibility of phase transformations upon cooling. Brauer et al. prepared Ta,N, by ammonolysing Ta,O, at 800-900 "C. Nitrides with higher Ta : N ratios were found to form above 900 "C. Fontbonne and Gilles" prepared Ta,N, in the same way at 850°C. At 975 "C they obtained Ta,N,. Above 975 "C changes in the X-ray powder pattern of Ta,N, were observed, together with the appearance of peaks ascribed to E-TaN and d-TaN0.8,.,. Treatment of Ta,N, in argon at 775"C/12 h was found to yield a compound deficient in nitrogen, having the composition Ta3N4.66(2). The authors found, furthermore, that a cation-deficient Ta,N,-type phase formed when mixed Ta-Nb oxides were ammonolysed.Phases of the Ta,N, type may accordingly exhibit anion as well as cation deficiencies. In the present study we obtain Ta,N, as a dark red and X-ray monophasic material at 900 and 1000 'C, and a mixture of Ta,N, and a small amount of Ta,N, at 1100°C. Our finding that the Ta,N, type phase with metal composition Zro,18Tao,,2 exhibits a varying 0: N content, depending on preparation temperature, conforms with the previous report of non-stoichiometry in Td,N,-type phases. Brauer et al. prepared pure TaON by alternating and successively smaller nitridations and oxidations in atmos-pheres with different oxygen contents. It is not clear from studies made hitherto if the oxynitride TaON is stable in NH, or if it is formed as an intermediate phase, in which case the formation of monophasic samples may be connected with an oxygen buffering by the sample.Gilles et al. obtained Zr,08N4 (b)by ammonolysis of ZrO, at 1100"C.'7 At 950°C Zr20N2 (y) was obtained, with forming as an initial intermediate phase. Ohashi et a/.prepared both and y at 900-1000 "C, by reacting mixtures of b-ZrNC1 and ZrO, in an NH, atmosphere." In the present study, only baddeleyite was observed at 1000 "C,and at 1100"C a mixture of Zr708N, and a small amount of baddeleyite. Conclusions Four different solid solution phases are found in materials prepared by ammonolysis of Zr-Ta gels at temperatures between 700 and 1100°C: (i) a pure baddeleyite-type phase is obtained at 800°C between TaON (x=O) and ZrO, (x=1) for Zr,Ta, -,O, +xN1-,.(ii) An orthorhombic oxynitride phase (probably of the orthorhombic ZrO, type) is observed in preparations at 700 "Cfor compositions with 0.26 < x < 0.68 for ZrxTal-xOl+xN1-x. (iii) A cubic phase with a fluorite- type subcell is present in preparations with 0.26 < x< 0.68 at 900 "C. The observed weight increase upon oxidation corre- sponds to compositions MX, ,85-1 .s7.Electron microdiffraction patterns indicate a monoclinic unit cell with a = J6/2 af,b = J2af, c = 2,/2af and = 125.3", with a, = equivalent cubic fluorite cell edge. (iv) A monophasic Ta,N, type phase, which contains up to 18 atom% Zr, is obtained at 900 and 1000°C.The phase compositions obtained by ammonolysis are found to be dependent on the hydrolysis conditions for the Zr-Ta gels. Preparations made with gels hqdrolysed under acidic conditions were biased towards phases that form at lower temperatures. J. MATER. CHEM., 1994, VOL. 4 1301 Our results for x =0 and x =1 generally agree with previous studies of compounds formed by ammonolysis of pure Zr and Ta oxides. 16 17 18 R. Juza, A. Rabenau and I. Nitschke, Z. Anorg. Allg. Chem., 1964, 332, 1. J-C. Gilles, Bull. SOC. Chim. Fr., 1962,2118. R. Collongues, J.-C. Gilles, A. M. Lejus, M. Perezy Jorba and D. Michel, Muter. Res. Bull., 1967,2, 837. The authors thank Prof. M. Nygren for support and valuable 19 20 R. Norrestam, Ark.Kemi, 1968,29,343. M. Ohashi, H. Yanamoto, S. Yamanaka and M. Mattoti, Mater. discussions and Mr. G. Westin for advice concerning the preparation of the Zr-Ta gels. 21 Res. Bull., 1993, 28, 513. M. R. Thornber, D. J. M. Bevan and J. Graham, Acta Crj~stullogr., Sect. B, 1968,24, 1183. 22 K. E. Johansson, T. Palm and P-E. Werner, J. Phys. E, 1980, 13, 1289. References 23 24 R. D. Shannon, Acta Crystallogr., Sect. A, 1976,32, 751. D. B. Wiles, A. Sakthivel and R. A. Young, Users Guide to 1 2 3 4 5 6 7 8 9 10 11 R. Marchand, F. Pors and Y. Laurent, Rev. Int. Hautes Temp. Rkfruct., Fr., 1986, 23, 11. P. Antoine, R. Marchand, Y. Laurent, C. Michel and B. Raveau, Muter. Res. Bull., 1988,23, 953. P. Antoine, R. Marchand and Y. Laurent, Rev.Int. Hautes Temp. Rifract., Fr., 1987, 24,43. R. Marchand, C.R. Acud. Sci. Paris, Ser. C., 1976,282, 329. X. H. Wang, A. M. Lejus, D. Vivien and R. Collongues, Muter. Res. Bull., 1988,23,43. G. Liu and H. A. Eick, J.Solid State Chem., 1990,89,366. A. Joly, C.R. Hehd. Siances Acad. Sci., 1876,82,1195. G. Brauer, J. Weidlein and J. Strahle, Z. Anorg. Allg. Chem., 1966, 348,298. J. Strahle, Z. Anorg. Allg. Chem., 1973,402,47. N. E. Brese, M. O’Keeffe, P. Rauch and F. J. DiSalvo, Acta Crystallogr., Sect. C, 1991,47,2291. P. Tiedemann and H. K. Muller-Buschbaum, Z. Anorg. Allg. Chem., 1982,494,98. 25 26 27 28 29 30 31 32 33 Program DB W3.2S for Rietveld Analysis of X-ray and Neutron Powder Diffraction Data Patterns (Version 8804), School of Physics, Georgia Institute of Technology, Atlanta. J-F. BCrar and P. Lelann, J. Appl. Crystallogr., 1991,24, 1. C. J. Howard, E. H. Kisi, R. B. Roberts and R. J. Hiil, J. Am. Ceram. SOC., 1990,73,2828. E. H. Kisi, C. J. Howard and R. J. Hill, J. Am. Ceram. SIX, 1989, 72, 1757. B. G. Hyde and S. Andersson, lnorganic Crystal Structurta, Wiley, New York, 1989, p. 69. B. G. Hyde, L. A. Bursill, M. O’Keeffe and S. Andersson, Nature (London), 1972,237,35. I. E. Grey, C. Li, I. C. Madsen and G. Braunshausen, Mder. Res. Bull., 1988,23,743. G. Pourroy, E. Lutanie and P. Poix, J. Solid State Chem., 1990, 86,41. A. Willgallis and H. Hartl, Z. Kristallogr., 1983, 164, 59. A. Harari, J-P. Bocquet, M. Huber and R. Collongues, C R. Acad. Sci. Paris, 1968, 267, 1316. 12 D. Armytage and B. E. F. Fender, Acta Crystallogr., Sect. B, 1974, 34 P. Bordet, A. McHale, A. Santoro and R. S. Roth, J. Solid State 30,809. Chem., 1986,64,30. 13 M. Weishaupt and J. Strahle, Z. Anorg. Allg. Chem., 1977, 429, 261. 35 36 A. Siggel and M. Jansen, Z. Anorg. Allg. Chem., 1990,582,93. R. T. Tuenge and L. Eyring, J. Solid State Chem., 1982,41, 75. 14 C. J. Howard, R. J. Hill and B. E. Reichert, Acta Crystallogr., Sect. B, 1988,44,116. 37 E. Schweda, D. J. M. Bevan and L. Eyring, J. Solid State Chem., 1991,90, 109. 15 A. Fontbonne and J-C. Gilles, Rev. lnt. Hautes Temp. Rkfract., Fr., 1969,6, 181. Paper 3/07020D; Received 25th November, 1993