J. MATER. CHEM., 1994, 4(10), 1603-1610 New Routes to Alkali-metal-Rare-earth-metal Sulfides John P.Cotter, Jonathan C. Fitzmaurice and Ivan P. Parkin* Department of Chemistry, Christopher lngold Laboratory, University College London, 20 Gordon Street, London, UK WCIH OAJ Ternary alkali-metal-rare-earth-metal sulfides, MRES, (M =Li, Na, K; RE =Y, La-Yb) were prepared by heating RECI, and M,S under an H,S-N, atmosphere at 750-800 "C for 10 min. The ternary sulfides form thin, mostly colourcrd plates. The structures were characterised by X-ray powder diffraction, which revealed high-temperature cubic modifications for LiHoS, and LiErS,. The disordered cubic form (NaCI type) was observed for LiLnS, (Ln =Pr-Er) and NaLnS, (Ln = La-Nd), while the ordered rhombohedra1 form (a-NaFeO, type) was observed for LiYbS,, NaLnS, (Sm-Yb) arid KLnS, (La-Yb).Alkali-metal-rare-earth-metal sulfides may also be synthesized from the rare-earth-metal chloride and alkali- metal halide under comparable conditions. Similarly, thermolysis of the rare-earth-metal sesquisulfide (y-RE,S3). oxysulf-ide (RE202S)or oxychloride (REOCI) with M2S or MCI in an atmosphere of H2S-N2 gives MRES,. The materials were characterised by powder X-ray diffraction, scanning electron microscopy (SEM), energy dispersive analysis by X-rays (EDXA), Raman, infrared, magnetic moments and microanalysis. Rare-earth-metal chalcogenides (e.g. RE2E3; RE =Y, La-Yb; E =S, Se) and ternary metal rare-earth-metal sulfides MRES, (M: Groups 1-3) are refractory materials that have a number of potential and actual technological uses, ranging from infrared lenses' (transmission range ca.1-14 pm), catalysts,2 colour phosphors3 (e.g. television sets utilize Eu-doped Y oxysulfides for the red pixels), lasers (neodymium ~ulfide),~ semiconductor dopants to ionic conductor^.^ Alkali-metal-rare-earth-metal sulfides MRES, are tra-ditionally prepared by high-temperature (900-1 100 "C) pyrol- ysis of an alkali-metal source and rare-earth-metal oxide under an atmosphere of H,S. Sodium lanthanide sulfides were prepared by direct combination of the elements at 800°C for 7 days in the presence of an excess of NaCl (Na,LnCl, also observed, Ln =lanthanide); lower ratios of NaCl resulted in the sesquisulfide.6 Lithium lanthanide sulfides (Ln =Pr-Dy) have been prepared in the cubic modification, from the reaction of a mixture of lithium carbonate and lanthanide oxide with H2S, at 900 "C for 24 h.7 This method was also utilised for the preparation of MLnS2 (Ln =lanthanide, M = Rb, Cs').Two solution-based coprecipitation methods have also been utilised: reaction of lanthanide nitrates in nitric acid with alkali-metal hydroxides' and lanthanide oxides in chloric acid with sodium chloride." Thermolysis of the precipitates at 900--1200"C in an H2S atmosphere for 2-24 h produced MLnS,. In the reactions that utilised chloric acid, a contami- nant of LnOCl was found in the product that could be removed by pretreatment with H2 and ammonium chloride at 450 'C.Lanthanum chalcogenides may also be synthesized via molecular precursor routes, where the reaction of lanthanum trisdialkylamides with H2Sin benzene followed by thermolysis in H2S produced La2S3.11 Kaner has shown that solid-state metathesis can offer an alternative method of synthesizing metal sulfide^.'^,^^ MoC1, and Na2S undergo a spontaneous exothermic reaction to produce NaCl and MoS,, with temperatures in excess of 1000"C being generated. The coproduced sodium chloride is readily removed by trituration to leave phase-pure metal sulfide. Extensions to this reaction have yielded a range of transition-metal chalcogenides and main-group s~lfides.'~ We have utilised metathesis and extended its usage to the synthesis of lanthanide and transition-metal nitrides,15 phosphides,16 arsenides and stibides." In most cases the reaction is initiated by minimal energy input such as a hot wire, grinding the reagents or by brief heating in a microwave or conventional oven.The reaction is self-sustaining once initiated and is often complete in 2-3 s, as shown by the accompanying thermal flash. We have also investigated the reaction of alkali-metal sulfides with rare-earth-metal halides in an attempt to broaden the range of materials available from the self-propagating route. Compared to the formation of lanthanide pnictides," reaction of rare-earth-metal halides with M2E (M ==Li, Na, K) are slow and not accompanied by a thermal flash. Indeed, thermolysis of the reagents at 550-700 "C in quartz ampoules was required and the products from the reactions were often contaminated with oxygen either in the form of REOCl or RE,02S (even after carbon coating of the ampoule).In this paper we report on an extension to these expcriments, where the metathesis reactions were carried out under a stream of H,S-N,, to produce rapidly a range of single phase- pure MRES, materials (depending on the stoichiometric ratio of REC1, to M2S). Experimental All reactions and reagent preparations were carried out under anaerobic or vacuum conditions. Glassware and quai tz/Pyrex ampoules were flame-dried before use. All solveiits were degassed with nitrogen. Tetrahydrofuran (thf) yas distilled from sodium-benzophenone and stored over 4 A niolecular sieves; methanol was disdilled from magnesium (activated with I,) and stored over 3 A molecular sieves.Rare-ear th-metal trichlorides, sulfur and alkali-metal halides were purchased from Aldrich and Strem Chemicals; lithium, sodium and potassium metal from BDH. All chemicals were used as received. Hydrogen sulfide (99.9%) was purchascd from Matheson Ltd. Alkali-metal chalcogenides M,E (M := Li, Na, K; E =S, Se), (Mo,5M'o.5)2E (M, M' =Li, Na and Li K) and Li2(Eo,5E'o.5)(E, E'= S, Se) were prepared from stoichiometric amounts of the elements in liquid ammonia under an inert atmosphere. Rare-earth-metal sesquisulfides (RE2&) were made by reaction of rare-earth-metal halides with H2S at 800 "C. Rare-earth-metal oxysulfides (RE202S) and oxychlor- ides (REOCl) were obtained from thermolysis of REC1, and Na2S in evacuated quartz ampoules at 550-700 "C;the phases were identified by powder XRD.X-Ray powder diffraction measurements were recorded on a Siemens D5000 difTactometer using nickel-filteretl Cu-Ka radiation (%=1.5406 A) as finely dispersed powders on a silicon mirror. An external NBS 6100 silicon standard was used for calibration. Diffraction patterns for single phases were indexed using the exhaustive (TREOR) approach. Crystallite sizes were determined using the Scherrer equation" and referenced to a standard KC1 pattern. SEM profiles and EDXA were performed on a JEOL JSM 820 instrument using the Kevex detection system (Quantum detector Kevex delta 4 and Quantex 6.2 software).Infrared spectra were recorded on a Nicolet 205 (CsI) using CsT pellets (100-600 cm-I). Raman spectra were recorded on a Dilor XY spectrometer. The samples were measured as powders dispersed at the focal point of the microscope attachment. The 514.53 nm line of an argon laser (50 mW) was the excitation source; the slit width was 300 pm. Magnetic moment measurements were measured on a Johnson Matthey balance (by the Evans method). Sulfur microanalyses were determined by the departmental service at University College. Thermolysis studies were accomplished in a brick tube furnace and the temperature was monitored by an external thermocouple; the apparatus is represented schematically in Fig.1. The thermolysis system, comprising a quartz tube and alumina boat were heated to 750 "C under vacuum and cooled to room temperature under a flow of nitrogen prior to the addition of reagents. Reactions were initiated in this apparatus at 750-800 "C under a flow of H2S-N, (30 :70) for a period of 10min. The products were allowed to cool under the H,S-N, flow. Preparation of MRES2from RECl, and M2Sin H2S(M =Li, Na, K; RE =Y, La-Yb) Alkali-metal sulfide M,S (1.5, 3.0 and 5.0mmol for M=Li, Na and K, respectively) and anhydrous rare-earth-metal tri- drying agent p205 I1I uu MI bubbler NaOH PbfOAc), Fig. 1 Schematic diagram of the thermolysis apparatus J. MATER. CHEM., 1994, VOL. 4 chloride (1.0 mmol) were ground together in an agate pestle and mortar.The powder was placed in an alumina boat inside a quartz tube that was sealed by Youngs taps. The tube was connected to the H,S-N, system as shown in Fig. 1. The apparatus was heated to 800 "C over 30 min under a flow of H,S-N, and maintained at that temperature for 10 min before it was cooled to room temperature. The resulting fused mass was triturated with thf (50ml) (for LiRES,), methanol (NaRES,) or water (KRES,) to remove the coproduced alkali- metal halide and any remaining M2S. The (mostly coloured) powders were then collected by filtration, washed with acetone (20 ml) and diethyl ether (20 ml) and dried in L'CICUO. The alkali-metal lanthanide sulfides were characterised by X-ray powder diffraction and magnetic moments measurements (Tables 1-3), Raman, infrared, SEM/EDXA (Fig. 2) and sulfur microanalysis.Sulfur microanalyses [%(calculated value in parentheses)] are as follows: LiSmS,, 27.7 (28.9); NaCeS,, 26.5 (28.2); KSmS2, 23.6 (25.1); LiHoS,, 26.7 (27.2); NaSmS,, 24.6 (26.8); KYbS,, 22.1 (23.1). [The samples analysed by microanalysis were also examined by EDXA which showed no oxygen present (1-2% detection limit).] Preparation and isolation of mixed alkali-metal lantha- nide sulfides of the form (Li, -xNax IRES, and (Li, -,K,)RES, (Table 4) were as described above using RECl, and Lio.5Nao.5S2 as starting materials. and Lio,5Nao,5S2 W1Oprn Fig. 2 SEM profile of KPrS, after washing Table 1 X-Ray powder diffraction data and magnetic moment measurements for LiRES, obtained from thermolysis of a mixture of REC1, and Li,S (2 :3) in an H,S-N, atmosphere at 750-800 "C phase colour of detected material LiYS, white La33 yellow Ce,S, red Pr2S3' lime green LiNdS, light green LiSmS, pale yellow LiEuS,' metallic grey LiGdS, white LiTbS, pink LiDyS, white LiHoS, pink LiErS, white LiYbS, yellow 'An eTcess of Li,S @a.5 x) give! Pr,S, 5.930 A). c =18.68 A. c =18.63 A. e = lattice/space group measured (-t0.005) ref. 7 PclbS(* 0.08)/ PB PcadM3+I/PB cubic, Fm3m 5.461 5.473 0 0 cubic, 1*3 8.722 8.727 0 0 cubic, 143d 8.626 8.636 2.12 2.11 8.567 8.576 3.08 3.58 5.613 5.628 3.61 3.62 5.555 5.588 1.35 0.85 cubic Fm3m 5.553 5.514 5.497 5.606 5.530 5.505 4.80 7.28 9.13 0 (Eu2' 7.94) 7.94 9.72 5.462 5.474 9.93 10.63 5.429 3.898' 10.11 10.61 5.415 3.875d 8.93 9.59 hexagonal, R3m 3.808' 3.842f 3.89 4.54 and LiPrS, (cubic, a =5.677 A;ref.a =5.687 A). Moisture-sensitive; decomposes to EuS (cubic, a = 18.560 A. c=18.54 A. J. MATER. CHEM., 1994, VOL. 4 1605 Table 2 >;-Ray powder diffraction data for NaRES, obtained from thermolysis of a mixture of REC1, and Na,S (1 : 3) in an H,S-N, atmosphere at 750-800 ^C lattice parameter/A measured' ref. 7 phase (s) colour of lat tice/space detected material group a C (I C NaY S, white hexagonal, R3m 3.936 19.832 3.968 19.89 NaLaS, yellow 5.868 5.881 NaCeS, red cubic, Fm3m 5.819 5.832 NaNdS, pale green 5.767 5.768 NaSmS, white 4.018 19.846 4.056 19.87 NaEuS, orange 4.035 19.916 4.042 19.92 NaTbS, white hexagonal, Rjm 3.988 19.880 3.989 19.87 NaHoS, white 3.934 19.870 9.949 19.86 NaYbSZb yellow 3.904 19.869 3.902 19.91 NaLaS, + yellow cubic, Fm3m 5.868 5.881 La,% cubic, 1_*3 8.720 8.727 y-Pr,S, lime green cubic, Z43d 8.567 8.576 NaNdS, + green cubic, Fm3m 5.767 5.768 -;'-Nd,S, cubic, 143d 8.503 -8.525 NaSmS, + yellow hexagonal,_ R3m 4.01 8 19.846 4.056 19.87 y-Sm,S, cubic, Z43d 8.43 1 8.437 The last four entries list multiphase products obtained from LnCl,/Na,S (2 : 3)." k0.005 A. NaYbS, (u= 3.897 A,c= 19.762 A)obtaiaed from Yb,02S/'3Na,S/H,S. Table 3 X-Ray powder diffraction data for KRES, obtained from thermolysis of a mixture of RECl, and K,S (1 : 5) in an H,S-N, attnosphere at 750-800 'C lattice parameter/A measured" ref.7 -phase(s) colour of lattice/space -detected material group a C U C KYS, white 4.020 21.862 4.023 21.85 KLaS, yellow 4.257 21.874 4.264 21.89 KPrS, pale green 4.190 21.858 4.185 21.75 KSmS, white hexagonal, R3m 4.190 21.857 4.107 21.76 KEuSz red-brown 4.094 21.841 4.094 21.85 KDYS, brown 4.029 21.858 4.030 21.83 KHoS, white 4.005 21.817 4.010 21.80 KYbS, yellow J 3.968 21.841 3.964 21.82 -y-La,S3 yellow cubic, 143d 8.720 -8.727 + x-La,& orthorhombi_, Pnum 7.577h 4.150 7.584' 4.144 -;'-Sm,S, orange cubic, 143d 8.43 1 -8.437 + x-Sm,S, orthorhombic, Pnam 7.379d 3.981 7.382' 3.974 The last tyo entries list multiphase products obtained from LnCl,/K,S (2: 2)." f0.005 A. ' b= 15.870 A. ' b= 15.860 A. h= 15.381A. h = 15.378 A. Table 4 X-Ray powder diffraction data for MI-,M,LnS, obtained from thermolysis of a mixture of RECl, and M,-,M',S (1 : 3) in ari H,S-N, atmosphere at 750-800 'C measured latLice parameter/A" phase colour of lattice/space unit-celb detected material group a C v~,)lume/A~ (Li,Na)SmS, white 4.019 19.557 273.6 ( Li,Na)ErS, hexagonal, Rjm 3.888 19.155 253.9 (Li,Na)YbS, green 3.874 19.223 249.8 (Li,K)SmS, green 3.883 19.632 256.4 (+ ;I-Sm2S3) (Li,K)TbS, grey hexagonal, R3m 3.894 19.048 250.1 ( Li,K)YbS, green 3.850 18.585 238.61 " f0.005 A. Alternative Methods of Preparing MRES, (i) Reaction of ground powders of lanthanide halides and alkali-metal hlaides ( 1:4) under comparable conditions (H,S, 8OO"C, 20min) was also found to produce MLnS,.The MLnS, materials were isolated and analysed, as previously described for reactions involving M,S and LnCl,. (ii) Reaction of rare-earth-metal sesquisulfides (yRE,S,), oxysulfides (RE20,S) or oxychlorides (REOC1) with an excess of M2S or MCl (ca. five-fold) in an atmosphere of H2S-N, (H,S, 800 "C, 20 min) gives MRES,. (iii) Reaction of rare-earth-metal halides with lithium selen- ide Li,Se and H2S under comparable conditions produces LiRES, (and no LiRESe,). For lithium sulfide selenide Li2(So.5, Se,,,), the products ranged from the sesquisulfide (y-Nd,S,) to a mixture of phases (y-Sm,S, and LiSmS,) to LiGdS,.Results and Discussion Reaction of rare-earth-metal halides with M,S (M =Li, Na, K) at 750-800 "C for 10 min under a flow of H,S produces crystalline alkali-me tal-rare-eart h-me tal sulfides, rare-earth- metal sesquisulfides RE,& or a mixture of the two phases which is dependent on the stoichiometry. Optimisation of the stoichiometry produces the single phase MRES, with yields being virtually quantitative. The idealised reactions may be summarised as follows (note: these equations are balanced for the initial ratios of metal halide to alkali-metal sulfide actually used; Ln stands for the cases where the reaction does not involve Y and RE stands for the cases which include Y and the lanthanides. The reactions may not be due totally to the direct exchange of the ions, uide infra). H,S/750 "C2LnC1, +3Li2S -y-Ln,S, +6LiC1 (1)(Ln =La, Ce, Pr) (In the case of PrC1,+5Li2S, a mixture of the sesquisulfide (y-Pr2S3) and LiPrS, was observed.) H,S:750 'C2REC1, +3Li,S +H,S -2LiRES, +4LiC1+ 2HC1 (RE=Y, Nd-Yb) (2) H,S/750 "C4LnC1, +6Na,S +H,S -2NaLnS, +yLn,S3 + lONaCl +2HC1 (3)(Ln =La, Nd, Sm) H Si750 CREC1, +3Na,S -NaRES, +3NaC1+ (3Na) (4)(RE=Y, La-Yb) H S'750'C2LnC1, +3K2S -Ln,S, +6KC1 (5)(Ln =La, Sm y and a phases) H,S/750 'CRECl, +3K,S -KRES, +3KClf (3K) (6)(RE=Y, La-Yb) SEM analysis of the alkali-metal-rare-earth-metal sulfides indicates that prior to trituration some MCl is evident.Washing removes the alkali-metal halide to reveal either uniform cubes or flat hexagonal crystallites of typical dimen- sion 10 pm (Fig.2). Analysis of the backscattered electrons indicates a homogeneous sample composition. Compositional EDXA studies of the triturated material, both as single spots and maps across the surface gave only the rare-earth metal, alkali metal (for sodium and potassium as lithium below the threshold detection limit) and sulfur and matched within experimental error (& 3% of stoichiometric ratio) the single J. MATER. CHE.M., 1994, VOL. 4 phases detected by powder X-ray diffraction. Chlorine was evident by EDXA in the unwashed material. In none of the samples was oxygen observed by EDXA (1 -2% detection limit). The X-ray powder diffraction patterns (Fig. 3) for MRES, showed two different phases: the disordered cubic (NaC1 type) and hexagonal (a-NaFeO,) lattices consistent with the litera- t~re.~The rare-earth-metal and alkali-metal cations occupy random (Na') sites in the NaCl lattice; the hexagonal phase is derived from a rhombohedral deformation of the cubic lattice with a cubic close-packed S2-arrangement with octa- hedral voids occupied by Na' and RE3+ and alternating cation layers perpendicular to the threefold a xis.The cubic lattice parameter decreases in a nearly linear fashion as the ionic radii of the rare-earth-metal decrease; a similar trend is observed for the hexagonal lattice u parameter. An exception is noted for europium, where it is possible that Eu3+ and Eu2+ coexist in both LiRES, and NaRES,.The hexagonal lattice parameter c remains essentially constant as a function of the lanthanide ionic radii. In contrast to the previously observed hexagonal phases for LiHoS, and L~EI-S,,~ high-temperature cubic modifications were observed. A comparison of the lattice constants for the cubic (a,) and hexagonal (ah) cells7 reveals that ah=a, J2/2 and Ch=uC2J3. This suggests that the cubic structure has a marked structural resemblance to the rhombohedral structure. Observed and calculated intensities for the hexagonal phase were qualitatively compatible, except for (001) lines, where prefered orientation of the crystallites as thin platelets parallel to the (001) direction leads to enhancement of these lines. The crystallite sizes for LiRES,, determined by the Scherrer equation" from the powder diffraction line uidths and com- pared too a standard KC1 pattern, were of the order of 520-600 A.CrystalliteQ sizes for NaRES, and KRES, were in the range 300-400A. Where the sesquisulfide was an additional phase [Fqn. (3) and (5)], the crystallite size was reduced (250-350 A). These crystallite sizes are comparable to those obtained by metathetical reaction^.'^.^^ A general reduction in crystallite size is observed on moving form LiRES, to NaRES, for comparable reaction conditions and may be due to the better diffusion of lithium in the melt in forming the material. The formation of mixed alkali-metal lant hanide sulfides offers the possibility of altering electronic properties.It was possible to prepare compounds of the type Li,-,Na,LnS, and Li, -,K,LnS, for Ln =Sm-Yb (Table 4). There is a trend towards the formation of LnS, (x =1.5-1.75) for the lighter lanthanides (eg. y-La,S,, y-Ce,S, and Nd,S,). The lattice types all belong to the hexagonal R3m phase and show a reduction in unit-cell volume from NaLnS, (5%) in the case of Li,-,Na,LnS, (Ln=Sm, Er, Yb) and from KLnS, (20%) for Li, -,K,LnS, (Ln =Sm, Tb, Yb). The individual phases, e.g. LiYbS, and KYbS,, are both yellow; the mixed phase is green and suggests the modification of the optical and elec- tronic properties from the individual compouFds. In both cases the crystallite size was reduced (100-250 A). The MRES2 materials and y-RE,& phase materials were also characterised by FTIR and Raman spectroscopy.The y-sesquisulfides, which belong to the cubic Th,P, type structure, all have similar Raman spectra with wavenumbers which decrease in a nearly linear fashion with lanthanide ionic radii. Disorder (i.e. randomly distributed vacancies) in the structure leads to broadening (20-40 cm-') of certain lines. The band in the region 230-250cm-' may be assigned to the A,, mode;,' it is somewhat narrower than other bands but is still very broad compared to other sesquisulfide phases. This band is also evident in the cubic NaCl type LiLnS, phases, with very broad band widths (ca. 80cm-'), consistent with a J. MATER. CHEM., 1994, VOL. 4 20 30 40 50 60 70 80 90 2Wdegrees Fig.3 X-Ray powder diffraction pattern for LiYbS, (a), NaYbS, (b)and KYbS, (c). Note only the first 20 lines were indexed on this pattern. disordered structure and an observed increase in the wave- number from Nd to Er (e.g. LiSmS,, 246cm-'; LiDyS,, 264 cm-l; LiHoS,, 268 cm-'). Typical Raman spectra for RE,S, and MRES, are shown in Fig. 4. For the hexagonal lattice, there are four infrared-active polar modes (2A,, +2Eu) and two Raman-active modes (A,,+E,).,' The band widths (ca. 10-20 cm-') in the Raman 1057 h u).-c C 20 1000 500 wavenumber/cm-' Fig. 4 Raman spectra of (a) Pr,S3 and (b)LiDyS, spectra are consistent with an ordered structure; the wave- number for the A,, mode (of weaker intensity compared to Eg) increases from La-Yb (e.g.NaNdS,, 236 cm-'; TcaSmS,, 263 cm-'; NaDyS,, 276 cm-'; NaYbS,, 284 cm- '). The stronger intensity E, line is located in the region 204-210 cm-'. A similar trend is observed for the KRES, derivatives, e.g. (Alg) KPrS,, 252 cm-'; NaSmS,, 262 cm-l; KYbS,, 281 cm-'. The stronger intensity E, line is also located in the region 204-210 cm-'. The mixed alkali-metal lanthanide phases Li, -,Na,LnS, also show the Raman-active (Algand E,) modes but the peaks are broadened, e.g. Yb 288 cm-' (21 cm-') and 206 cm-' (42 cm-') (where the figures in brackets represent peah widths at half height). Infrared spectra for the hexagonal lattice type were all similar. A broad band centred in the region 310 cm- and a sharper band located at ca.150cm-' may be assigned to E, modes. The A,, stretching mode may be correlated wirh peaks at 220 cm-' and 330 cm-'. The magnetic moment measurements for MRES2 were consistent with the rare-earth metal in the RE3+ ovidation state (Table 1). An exception was noted for LiEuS, (p==4.8pB) where the lattice parameter also indicates a mixture [of Eu2+ and Eu3+ oxidation states. The potassium analogue (p= 3.41 pB)agrees more closely with the higher oxidation state. Sulfur microanalyses were consistent with the formula MRES, and agreed with the phases observed by powder XRII. Reaction of rare-earth-metal(II1) chlorides with H2S at 750 "C for 10 min in the apparatus shown in Fig. 1 produced rare-earth-metal sesquisulfides (RE,S3) in quantative yield.A mixture of two phases (a-and y-Sm,S,) was observed for samarium, while the single-phase materials y-Ln,S3 ( Ln =Pr, Nd, Gd) and 6-Ho2S3 were obtained. Reaction of REC1, with M2S in evacuated quartz ampoules at 750°C produced a mixture of up to three phases, i.e. REOC1, RE2S3 and RE,O,S J. MATER. CHEM., 1994, VOL. 4 due to the oxophilic nature of the rare-earth-metal chloride (through oxygen removal from the quartz) at elevated reaction temperatures. In order to minimise the role of oxygen contamination, the reaction of REC1, with M,S was investigated in an H,S-N, flow stream at 800°C. (If the reaction was carried out with no H2S flow then some RE,02S was also observed.) The predominant phase MRES, was observed independent of the containment vessel (e.g.alumina boat); the formation of M LnS, corresponded with previous observations for Ln = Pr-Yb' [eqn. (2)] with the sesquisulfides formed for La, Ce and Pr [eqn. (l)].Increasing the stoichiometric ratio of Li2S to PrCl, to 5:1 resulted in a mixture of y-Pr,S, and LiPrS,. The europium analogue was found to be moisture-sensitive, forming EuS. It was necessary to optimise the formation of NaRES, and KRES, by increasing the stoichiometric ratio of M,S (M=Na, three-fold excess; M=K, five-fold) [eqn. (4) and (5)]to REC13. Lower ratios of the alkali-metal sulfide, in both cases, resulted in the incorporation of the sesquisulfides [eqn. (5)] for potassium and a mixture of the sodium lantha- nide sulfide and sesquisulfides [eqn.(3)], with a concomitant decrease in the crystallite size of the products (Scheme 1). The role of H,S is probably two-fold, both in controlling the oxygen contamination and being intimately involved in the reaction. It is apparent that the formation of RE,S, can be derived either directly from the reaction of REC1, with H,S, or by the metathesis of REC1, with M2S. It is likely that metathesis is in part responsible for this reaction as MC1 was always detected in the reactions and it is unlikely that it would be formed unless some exchange of ions has taken place. The incorporation of alkali-metal chloride in the fused material is indicative of a flux being formed. The differences of ionic mobility within the flux are probably the reason that an excess of reagent is required in the preparation of both the sodium and potassium lanthanide sulfides.The idealised equa- tions (4) and (6) with sodium and potassium being formed are incorrect in that the native metal was not observed after the reaction but some unreacted M,S; the additional M,S was required to insure sufficient Na or K was present so that a single phase MRES, was produced. Owing to the ambiguity of the reaction pathway, we studied the thermolysis of lanthanide chlorides with alkali-metal chlorides at 800°C in H,S. The reactions were also found to produce alkali-metal-rare-earth-metal sulfides MRES, in quantitative yield. These reactions are intrinsically quite simple in that they do not require any prior synthesis, unlike those involving M,S, and they may be applied to a range of materials, such as the synthesis of Group 2 and transition- metal mixed lanthanide sulfides.Hence the source of sulfur in the product comes directly from the H,S. It is impossible to RECl3 + MZS RECI, + Li2Se H2SI RE202S + M+t~lCl rule out H,S as the sulfur source for the reaction of REC1, and M,S (M=Li, Na, K) indicating that no metathesis reaction occurs. This would make the idealised equations (1)-(6) highly speculative. However, as stated above, the observation of significant amounts of MC1 (M=Li, Na, K) in the products from the reaction of RECl, and M,S indicate that a metathetical pathway is in part responsible for the reaction. Further reactions of M,S and RECI, in the apparatus shown in Fig.1 without an H2S flow did produce MRES, as the primary product but also a small amount of Ln,02S. A detailed EDXA analysis of the starting materials has shown that oxygen is not present (< 1-2%) and that any oxygen incorporation in these reactions comes from the containment vessels. The use of H2S in these reactions purges any oxygen from the product presumably by the elimination of water. The reaction of M2S and RECl, was not accompanied by a thermal flash, as previously seen for metathesis and the products were contained as a fused mass within the open alumina boats, rather than being spread over the reaction vessel walls. Hence there is no direct analogy between these reactions and the metathesis observed between alkali-metal pnictides with lanthanide halides17'18 and by Kaner between transition-metal chlorides and sodium sulfide. Hess's law calculations22 indicate that the reactions are less exothermic than reactions between high oxidation state t ransition-metal halides and sodium sulfide.It was also possible to make a range of alkali-metal-rare- earth-metal sulfides by the reaction of RE,S, and RE20zS with M,S or MC1 under an atmosphere of H,S-N, at 800 'C. The routes to alkali-metal lanthanide sulfides are also sum-marised in Scheme 1. It would appear that provided the stoichiometry is correct, almost any source of the rare-earth metal and alkali metal can be utilised in the formation of alkali-metal lanthanide sulfides and that this is the prefered thermodynamic product for a number of reactions.The ther- molyses proceeded rapidly at 800 "C, with the formation of a molten flux incorporating the lanthanide and the alkali-metal precursors, which allows for rapid diffusion and formation of MLnS,. Reaction of mixed alkali-metal sulfides [ie. (Li,,,Ko,5)2S and (Lio.5Nao,5)2S] with LnC1, (Ln=Sm, Er, Tb, Yb) and H,S at 800 "C produced the mixed alkali-metal lanthanide sulfides MI-,M',LnS, (M, M'=Li, K; Li, Na) in good yield (Table 4), as assessed by X-ray powder diffraction and EDXA. Conclusions The reactions of rare-earth-metal(rI1) chlorides with alkali- metal sulfides at 750-800 "Coffer a rapid route to pure single- phase crystalline alkali-metal-rare-earth-metal sulfides, or in RECS RECI, + MCI lH* RE2S3+MCI LnCI, + Li,S M = Li.Na, K, RE = Y or lanthanide Scheme 1 Routes to alkali-metal rare-earth sulfides. (Note: Reactions involving lithium selenide or lithium sulfide selenide and RECl, and H,S at 750 C formed only LiRES,. Reaction of REC1, and M,S at 750°C without H,S flow primarily produced MRES2, but invariably formed a minor phase of RE,O,S.) J. MATER. CHEM., 1994, VOL. 4 1609 some cases rare-earth-metal sesquisulfides. The stoichiometry of the reaction has to be closely monitored to insure that the required phase is produced. The reaction probably proceeds by combination of a metathetical ionic exchange and a heterogeneous reaction with H2S. The reaction occurs at a temperature that is 100 "C lower than conventional prep- 8 9 R.Ballestracci and E. F. Bertaut, Bull. Soc. Fr. Mineral. Cristallogr., 1964. 87, 512; P. N. Kumta and S. H. Risbitd, Prog. Crystal Growth Charac., 1991,22, 321. W. Bronger, W. Bruggemann. M. von der Ahe and D. Schmitz, J. Alloys Compounds, 1993,200,205. R. Ballestracci and E. F. Bertaut, Colloq. Int. Centre. N,it. Rech. Sci., 1967, 157,41. arations and also shows versatility in the preparation of M, -,M,'LnS, materials. 10 11 M. Sato, G. Adachi and J. Shiokawa, Muter. Rex Bull., 1984, 19, 1215. G. C. Allen, M. Paul and M. Dunleavy, Adc. Muter., 199.', 4,424. The Leverhulme Trust (fellowship to J.C.F.), the Royal Society (glove box purchase) and the University of London Central Research Fund are acknowledged for financial support.12 13 P. R. Bonneau, R. F. Jarvis and R. B. Kaner, Nuture (,',ondon), 1991,349,510. P. R. Bonneau, R. F. Jarvis and R. B. Kaner, horg. Chew, 1992, 32, 2127. N. Williams of the Open University is thanked for assistance with the SEMiEDXA studies and D. Ciomartan of UCL for help with the Raman spectra. 14 15 16 I. P. Parkin and A. T. Rowley, Polyhedron, 1993,12,2961. I. P. Parkin, A. Hector and J. C. Fitzmaurice, J. Ch(1m.Soc., Dalton. Trans., 1993, 2435. 1. P. Parkin and A. Rowley, J. Muter. Chem., 1993,3,689 17 J. C. Fitzmaurice, A. Hector, A. Rowley and I. P. Parkin, 1. Muter. References 18 Chem., 1994,4,289. J. C. Fitzmaurice, A. Hector, A. Rowley and I. P Parkin, 1 2 3 4 5 6 K. L. Lewis, J. A. Savage, K. J. Marsh and A. P. C. Jones, New Optical Materials, Proc. SPIE-Int. Soc. Opt. Eng., 1983,400, 21. C. Y. Yeh and C. Sarini, US Pat. 4,560,804. B. T. Kilbourne, A Lanthanide Lanthology, Molycorp Inc, White Plains. NY, 1993, Part 1. A. Bornstein and R. Reisfeld, J. Non. Crystalline Solids, 1982, 50, 93. T. Ogura, A. Mikami, K. Tamaka, T. Taniguchi, M. Yoshida and S. Nakajima, Appl. Phys. Lett., 1986,48, 1570. T. Schleid and F. Lissner, Eur. J. Solid. State. Inorg. Chem., 1993, 30, 829. 19 20 21 22 Poljlhedron, 1993, 13, 235. H. P. Klug and L. E. Alexander, X-Ray Difraction Proc edure fur Poljcrystalline and Amorphous Muteriuls, Wiley, New J ork, 2nd edn., 1974. D. S. Knight and W. B. White, Spectrochim. Actu, Purt A, 1990, 46,381. P. Bruesch and C. Schuler, J. Phjx Chem. Solids. 1971,32, 1025. D. D. Waymann, W. H. Evans, V. B. Parker, R. H. $chumm, I. Halow, S. M. Bailey, K. Chumey and R. L. Nuttal, The NBS Tables of Chemical Thermodynamic Properties, ,Zmerican Chemical Society, Washington DC, 1982. 7 R. Ballestracci, Bull. Soc. Fr. Mineral. Cristallogr., 1965, 88, 207; M. Tromme, C.R. Acad. Sci. Paris., Ser. C, 1971, 273, 849; Paper 4/02815E; Receiued 12th MJ~,1994