Solid Compounds of Transition Elements

IWIS-mf— 10/rl

VIII. International Conference of Solid Compounds of Transition Elements

ABSTRACTS

April 9-13 1985 Vienna /Austria PREFACE

This volune contains the collected extended abstracts of the invited plenary lectures and of the contributed papers (short oral presentations or posters) accepted for presentation at the VIII International Conference on Solid Compounds of Transition Elements to be held in Vienna, April 9^-14^, 1985. The contributions are numbered in the order of the conference program (PL... plenary lecture, 0... oral contribution, P... poster). The posters are grouped to poster sessions as indicated by the first number following "P". The subdivision into A and B has rather technical reasons.

For various reasons it was not possible to assign a single topic to each session. It was attempted, however, to assemble related papers in groups.

The extended abstracts presented in this volune are photo-copied from the originals sent by the authors.

We hope that this volune may help the participants as a guide for a most profitable scientific impact of the conference and as a valuable documen- tation. We anticipate, however, that the authors of the contributions will publish their papers in more details in appropriate scientific journals.

K.Komarek

H.Boller A.Neckel

Vienna, March 1985 INTERNATIONAL ADVISORÏ COfUTTEE E.F. Bertaut (Grenoble) H. Nowotny (Vienna) R. Fruchart (Grenoble) E. Parthé (Geneva) J.B. Goodenough (Oxford) A. Rabenau (Stuttgart) F. Grjfovold (Oslo) S» Rundquist (Uppsala) F. Jellinek (Groningen) A. Hold (Providence)

LOCAL ORGANIZING COttOTIEE Chairman: V.L. Komarek Vice-Chairmen: H. Boiler and A. Neckel Members: R. Eibler, P. Herzig, K. Hiebl, H. Ipser, P. Rogl, W. Schnöll, C.J. Schuster, K. Schwarz, P. Terzieff, P. Weinberger, E. Wimner

PATRONAGE Vizekanzler und Bundesminister für Handel, Gewerbe und Industrie Dr. Norbert Steger Bundesminister für Wissenschaft und Forschung Univ.-Doz.Dr. Heinz Fischer

Grateful acknowledgement is made for financial support of the Conference: Bundesministerium für Wissenschaft und Forschung Österreichische Forschungsgemsinschaf t Bundesministeriun für Handel, Gewerbe und Industrie Fachverband der chemischen Industrie Österreichs

Our special thanks are due to Mr. Theodor Kery, Landeshauptmann (Governor) of Burgenland for his invitation to the reception at the Esterhazy Castle in Eisenstadt und to Dr. Helmut Zilk, Lord Mayor of the City of Vienna, for his invitation to the reception in the Town Hall.

SPONSORS Sponsorship has been granted by the international organisations IUPAC and IUPAP. CONTENTS

PLENARY LECTURES

PL 1 TERNARY TRANSITION METAL CHALC06ENIDES WITH FRAMEWORK STRUCTURES AND THE CHARACTERIZATION OF THEIR BONDING BY MAGNETIC PROPERTIES W. BRONGER PL 2 INTERCALATION REACTIONS OF TRANSITION METAL COMPOUNDS VIA ELECTRON/ION/TRANSFER. R. SCHOELLHORN PL 3 ELECTRON MICROSCOPY AND DIFFRACTION OF MODULATED STRUC- TURES J. VAN LANDUYT PL 4 ALL ELECTRON LOCAL DENSITY THEORY OF TRANSITION METAL COMPOUNDS AND THEIR SURFACES A.J. FREEMAN PL 5 ORDER AND DISORDER IN TRANSITION METAL CARBIDES AND NI- TRIDES: EXPERIMENTAL AND THEORETICAL ASPECTS C.H. DE NOVION PL 6 SYNTHESIS AND THERMODYNAMICS OF NONSTOICHIOMETRIC RARE EARTH COMPOUNDS UNDER EXTREME CONDITIONS E. KALDIS PL 7 RECENT DEVELOPMENTS IN SOLID MAGNETISM WITH SPECIAL REFLRLNCE TO TRANSITION METAL COMPOUNDS E.P. WOHLf-ARTH PL 8 CRYSTAL CHEMISTRY OF TRANSITION METAL BORIDES YU.B. KU2MA PL 9 STRUCTURE, DEFECTS AND PROPERTIES OF SOME REFRACTORY BORIDES T. LUNDSTROEM PL10 NEW EXAMPLES OF CLUSTERS, EXTENDED METAL-METAL BONDING AND INTERSTITIAL DERIVATIVES JOHN D. CORBETT CONTRIBUTED PAPERS **

0 1 THE «_ TRANSITION IN PURE AND DOPED FeS : PHYSICAL AND STRUCTURAL ASPECTS

** Names of the author(s) who 1nte/id(s) to present the paper at the conference are underlined. 012 ON THE CHROMIUM MAGNETIC MOMENTS X. OUOET 013 MAGNETIC ORDERING IN INTERMETALLICS OF THE ThCr2Si2 TYPE J. LECIEJEWICZ, AND A,. SZYTULA 014 MAGNETIC AND ELECTRONIC PROPERTIES OF THE tu(l-x)La(x}S SOLID SOLUTION FROM (151}Eu MOESSBAUER SPECTROSCOPY JL!L SANCHEZ, J.M. FRIEDT, K. WESTERHOLT, AND H. BACH 015 ON THE PHYSICAL PROPERTIES OF SEVERAL M2Mo6X5 COMPOUNDS (H = GROUP IA, GROUP IIIA METAL; X = Se, Te)

O.M. TARASCON, Ft 3^ DI SALVO, AND J. V. WASZCZAK 016 TRANSMISSION ELECTRON MICROSCOPIC INVESTIGATION OF V5Si 3 PRECIPITATES IN V3Sf (A15 STRUCTURE)

A.BEN LAMINE, FA REYNAUD, AND J.P. SENATEUR 017 PREDICTION OF THE IDEALIZED COMPOSITION OF COMPOUNDS WITH CENTRED TRIGONAL PRISMS IN RE - Ni - Si SYSTEMS ii. yyI, AND E. PARTHE 018 GROWTH SIMULATION OF BORIDE COMPOUNDS S.i HAMAR-JHIBAULT, AND R. HAMAR 019 CRYSTAL CHEMISTRY, STABILITY AND ORDER IN TERNARY METALLIC PNICTIDES R. MADAR, E. DHAHRI, P. CHAUDOUET, J.P. SENATEUk, R7 FKUCHART, AND B. LAMBERT 026 INVESTIGATIONS INTO CRYSTALLINE AND AMORPHOUS COPPER - ARSENIC - CHALCOGENIDES R. BLACHNIK, AND G. KURZ 021 STRUCTURAL AND MAGNETIC PHASE DIAGRAM FOR THE MnAs - CrAs SYSTEM H. fJELLVAG, AND A. KJEKSHUS 022 STRUCTURE AND BONDING OF TRANSITION METAL CARBIDES AND HYDRIDES J. HAUCK 023 INTERBAND DIELECTRIC RESPONSE IN TRANSITION METAL HYDRIDES M. FLIYOU, R. RIEDINGER, AND M..A.. KHAN PI A I STRUCTURAL AND MAGNETIC PROPERTIES OF COBALT QALLIUM SULFIDES: ot - CoGa^ and V-CoGa^ E. AGOSTINELLI, L. GASTALDI, H.6. SIMEONE, AND S. VITICOLI PI A 2 PHASE RELATION AND AGING EFFECTS IN Fe(l-x)Co(x)S SYSTEM L. BARTHELEMY, AND C. CARCALY PI A 3 PHYSICAL PROPERTIES OF THE SOLID SOLUTIONS TKu(2- x)Fe(x), 8

R. BERGER, C.F VAN BRUGGEN, ANO Ft JELLINEK PI A 4 STRUCTURAL CHARACTERIZATION OF BULK AND THIN FILM PHASES OF Ni(l+x)Te2 (0<=x<=l) SURAJ BHAN, ANO MRITYUNJAYA SINGH PI A 5 THREE-NUCLtAR Ta-CLUSTERS IN THE COMMENSURATE CDW-STATES OF 2H-TaS2 AND 2H-TaSe2 L.Byj.2i A- LERF, ANO S. SAIBENE PI A 6 ELECTROOEPOSITED TUNGSTEN SELENIDf FILMS: STRUCTURAL, OPTICAL AND ELECTRICAL CHARACTERIZATION. . S.CHANDRA, AND S.N. SAHU PI A 7 ELECTRIC CONDUCTIVITY OF AgCrTiS4 AND AgCrZrS4 Z. CYBULSKI PI A 8 STRUCTURAL ANO MAGNETIC STUDY OF THE Cr2S3-Ga2S3 SYSTEM. L. GASTALUI, 5. VITICOLI, J. FLAHAUT, M. GUITTARD, A. TOMAS, AND M. WINTENBERGER PI A 9 PHYSICAL PROPERTY CHANGES OF (T,M)(l+x)Nb(3-x)Sel8 WITH T=Fe, Cr ANO M=Nb, Ti H. GRUBER PI A18 MAGNETIC PROPERTIES OF NON-STOICHOMETRIC T]Fe2Se2 . HAEGGSTROEM, H.R. VERMA, S. BJARMAN, AND 7 BER6ER PI All NEW TERNARY CHALCOGENIDES OF THE COINAGE METALS WITH THALLIUM(I) OR ALKALI METALS KURJ Qi. KLEPP PI A12 FAR-INFRARED AND X-RAY INVESTIGATIONS ON MIXED TRANSI- TION METAL OICHALCOGENIDES WITH SEMICONDUCTING/METALLIC BEHAVIOUR G. KLICHE PI A13 FAR-INFRARED REFLECTION SPECTRA OF PYRITE AND MARCASITE TYPE MANGANESE, IRON, AND PLATINUM GROUP CHALCIDES H.D. LUTZ, 6. SCHNEIDER, G. WAESCHENBACH, AND G. KLICHE PI A14 POLYSULFIDOAURATE(I) AND THIOAURATE(I): SYNTHESIS AND STRUCTURE OF AuS9(-), Au2S8(2-) AND Aul2S8{4-)

M.P. PARDO, AND Jt FLAHAUT. PI A16 HYURATED LAYERED PHASES M(x)H20(y)CrS2 .(M...ALKALI, ALKALINE EARTH METAL) DERIVED FROM 2H- AND 3R- K(x)H28(y)CrS2 BY TOPOTACTIC ION EXCHANGE R. QUINT, H. 6OLLE.R, ANO H. BLAHA PI A17 HYPERFINE SPECTROSCOPIC STUDY OF POLYMORPHIC PHASE TRANSITIONS IN TaS2 AND TaSe2

S. SAIBENE, At LEKF, AND T. BUTZ PI B 1 INELASTIC NEUTRON SCATTERING AND LATTICE DYNAMICS OF HfSe2, Sn$e2 AND TiS2 Qt SCHAERLI, W. BUEHRER, AND F. LEVY PI B 2 MAGNETIC AND THERMOELECTRIC PROPERTIES OF SINGLE CRYSTAL CoTe AND NiTe tL SCHICKEJANZ,. P.. J.ERZIEFF, ANO K.L. KOMAREK PI B 3 WSe2 HOMO- ANO HETEROJUNCTIONS EL SPAEH, M.CH. LUX-STEINER, M. OBERGFELL, AND E. BUCHER PI B 4 LATTICE INSTABILITIES IN THE PSEUOOBINARY SYSTEM InMo6S(8-x)Se(x) JI.53^ JARASCON, F. J. DI SALVO, AND J.V. WASZCZAK PI B 5 FIRST EXAMPLE OF (MO3SE3) CLUSTERS IN SOLUTION J.M. TARASCON, F.J. 01 SALVO, C.H. CHEN, P.J. CARROLL, M7"WALSH7"AND"L. RUPP PI B B ON THE VANADIUM - TELLURIUM PHASE DIAGRAM »I IP.SER.. AND E. WACHTEL PI B 7 CRYSTALLOGRAPHIC AND MAGNETIC STRUCTURE OF Tl-Fe-S COMPOUNDS D. WELZ, P. DEPPE, M. ROSENBERG, H. SABROWSKY, AND W. SCHAEFER PI B 8 CHANGES IN RAMAN SPECTRA OF (TaSe4)2I INDUCED BY THE PEIERLS TRANSITION A,. ZWIÇK,. ÇLA.. RENUÇÇI, P. GRESSIER, AND A. MEERSCHAUT

PI B 9 THE PYROCHLORES Pb2M0.5Sbl.506.5 (M: Al, Se, Crr Rh) C. CASCALES, I. RASINES, P. GARCIA CASADO, AND J. VEGA

PI 818 THE OXIDES M3Sb5012 (M= Y,Pr,Nd,Sm,Eu,Gd,TbtDy,Ho,EP, Tm.Yb.Lu) CM. MARCANO, AND I. RASINES PI Bll A MOESSBAUER- AND X-RAY INVESTIGATION OF Zn-CONTAINING FAYALITE AND Fe-CONTAINING WILLEMITE li ERICSSON, AND A. FILIPPIDIS PI B12 SOLID STATE CHEMICAL MODEL FOR THE SOLUBILITY BEHAVTOUR OF Co- Mn-CARBONATE SOLID SOLUTIONS di QA-SSJAEGER, A. FLUCH, ANO W. ENGELMANN

?2 A i THE TERNARY SYSTEM ERBIUM-BORON-CARBON. ISOTHERMAL SECTION AT 1500 C. -L BAUER P2 A 2 SIMPLE CHEMICAL TWINNING: A MODEL TO EXPLAIN THF. MODULATED STRUCTURES WHICH APPEAR DURING THE CRYSTALLISATION OF SPUTTERED IRON-CARBON AMORPHOUS ALLOYS I*. 5éy££l5EQSSE, G. LE CAER, AND C. FRANTZ P2 A 3 TERNARY LANTHANOID MANGANESE CARBIDES WITH FILLED BaCdll and Th2Znl7 TYPE STRUCTURES Si. BLQÇK. AND W. JEITSCHKO P2 A 4 NEW SPUTTERED Nb-N(x) FILMS WITH HIGH NITROGEN CONCENTRATIONS: ELABORATION AND PROPERTIES !L ÇABANEL, J.C. JOUBERT, J. CHAUSSY, AND J.MAZUER P2 A 5 UNiySUAL ELECTRICAL BEHAVIOUR ASSOCIATED WITH THE SEMICONDUCTOR-METAL PHASE TRANSITION IN MIXED PHASE VANADIUM OXIDES A. CLARK, n^ LOVEL.L, ANO D^L^ IUNNIÇLIFFE P2 A 6 THE INFLUENCE OF STOICHOHETRIC OEVIATIONS ON ELECTRICAL PROPERTIES OF TITANIUM OXIDES H. GRUBER, AND E,. KRAUJZ P2 A 7 FCC OXYCARBIDE PHASE OF SCANDIUM AND OF YTTRIUM: COMPOSITION LIMITS £2. KAREN, V. BROZEK, AND B. HAJEK P2 A 8 HYDROLYSIS OF CARBIDES OF MANGANESE Mn5C2 AND Mn23C6 £i KAREN, AND B. HAJEK P2 A 9 THERMAL EXPANSION STUDIES ON THE GROUP V-VI TRANSITION METALS

P.? A10 PREPARATION AND CHARACTERIZATION OF NEW OXYNITRIDES WITH PEROWSKITE STRUCTURE !L MARCHAND, F. PORS, AND Y^ LAURENT P2 All X-RAY DIFFRACTION AND DIFFUSE SCATTERING OF M7C3 SINGLE CRYSTAL CARBIDES H-p-- MORNTROLI, M. KHACHFI, M. GANTOIS, AND A. COURTO!" P2 A12 DIFFRACTION STUDIES OF ELECTRON DENSITY, LATTICE VIBRATIONS AND INTERDIFFUSION IN SOME TRANSITION METAL CARBIDES V. VALVODA ?Z A13 CRYSTAL STRUCTURES AND MAGNETIC PROPERTIES OF LOW- niMENSIONALLY BRIDGED TRANSITION METAL HEXACYANO COMPLEXES (L BABEL, M. WITZEL, AND J. PEBLER P2 A14 X-RAY EMISSION SPECTROSCOPIC STUDY OF SOME TRANSITION ELEMENTS IN CALCIUM GERHANATE GARNETS

E. AGOSTINELLI, AND D. FIORIANI P3 A 2 ASYMMETRY OF THE HELICAL-FERROMAGNETIC TRANSITIONS IN MnAs(l-x)P('x)-CRYSTALS !L BAERNER, CH. KUHRT, AND A.F. ANDRESEN P3 A 3 MAGNETIC PROPERTIES OF VITREOUS PHASES OF THE SYSTEM MnS-Ga233-La2O3. S. BARNIER, M. GUITTARD, M. WINTENBERGER, AND J.. FLAHAUT P3 A 4 MAGNETIC PROPERTIES OF SOLID SOLUTION PHASES IN THE SYSTEM CrP - CrAs - MnP - MnAs H. FJELLVAG, A. KJEKSHUS, S. STOLEN, AND A.F. ANDRESEN P3 A 5 HEAT CAPACITY AND MAGNETIC A.C. SUSCEPTIBILITY OF MnRhAs. J. GARCIA, C. RILLO, J. BARTOLOME, D. GONZALEZ, R. NAVARRO, B. CHENEVIER,"P. CHAUDENET, AND D. FRUCHART P3 A 6 MORIN-TRANSITION IN Ti-SUBSTITUTED HEMATITE: A M0E5S- BAUER STUDY T. ERICSSON, A^ KEISHNAMUglHY, AND B.K. SRIVASTAVA P3 A 7 MAGNETIC STRUCTURES OF Fe6e :L BERNHARD, B. LEBECH, 0. BECKMAN, AND T. FRELTOFT P3 A 8 MAGNETIC STRUCTURES OF FeSn2 G. VENTURINI, B. MALAMAN, B. ROQUES, 0. FRUCHART, SJ3D"5"LrCSER P3 A 9 PRESSURE DEPENDENCES OF THE CURIE TEMPERATURE FOR CoMnSi(x)Ge(l-x) SYSTEM OF SOLID SOLUTIONS sL NIZIQL, R. ZACH, AND J.P. SENATEUR P3 A10 MAGNETIC AND STRUCTURAL PHASE DIAGRAM OF Mn(4-x)Ni(x)N C. RILLO, J. GARCIA, J. BARTOLOME, J.A. PUERTOLAS, !L NAVARRO, AND D.- FRUCHART P3 All MAGNETIC ORDERING IN THE SPINEL SOLID SOLUTION ZnCr(2x)Al(2-2x)S4 A. WIEDENMANN, J. ROSSAT-MIGNOO, M. HAMEDOUN, J7L."OORMANN7~AND M, NOGUES P3 A12 HALF-METALLIC FERROMAGNETS MAGNETIZATION AND ELECTRICAL TRANSPORT PROPERTIES OF SOME HEUSLER-TYPE ALLOYS ?L:L OTTO, C. HAAS, AND C.F. VAN BRUGGEN P3 A13 MAGNETIC PROPERTIES OF MIXED HEUSLER ALLOYS (M, M' )I:MNSN (M,M\..Co.Ni,Cu.Rh.Pd) E. UHL, AND H. BOLLER P3 A14 SOFT MAGNETIC AMORPHOUS RIBBONS G. BADUREK, R^ GROESSINGER, H._ SASSIK, AND A. VEIOER P3 A15 RE-FE-B A NEW FAMILY OF MATERIALS FOR PERMANENT MAGNETS R. EIBLER, R. GROESSINGER, G. HILSCHER, H. KIRCHMAYR; H7 §SSSIK7"AND"G7"WTESING P3 A16 MAGNETIC STRUCTURES DETERMINED BY NEUTRON DIFFRACTION - DESCRIPTION AND SYMMETRY ANALYSIS 4i QLES, W. SIKORA, A. BOMBIK, AND M. KONOPKA P3 A17 STABILIZATION AND CHARACTERIZATION OF UNUSUAL OXIDATION STATES OR UNUSUAL ELECTRONIC CONFIGURATIONS OF TRANSITION ELEMENTS. G. DEMAZEAU, M. POUCHARD, B. BUFFAT, AND P. HAGENMULLER P3 A18 CRYSTAL CHEMISTRY AND MAGNETISM IN SILICIDES LaFe(2- x)Rh(x)Si2 (ThCr2Si'2-TYPE) P. ROGL, K. HIE8L, AND fi. WIESIN6ER P3 B 1 TRANSMISSION ELECTRON MICROSCOPIC STUDY OF V3Si SINGLE CRYSTALS DEFORMED BY COMPRESSION AT 1650 C. A. BEN LAMINE, f\ REYNARD, AND J.P.SENATEUR P3 B 2 TERNARY TRANSITION METAL SILICIDES (OR QERMANIDES) BUILT UP OF INFINITE COLUMNS OF Si(Qe) - CENTERED SQUARE ANTIPRISMS AND TRANSITION METAL - CENTERED OCTAHEDRA. !L CHABOT, E. PARTHE, AND K.. CENZUAL P3 B 3 STRUCTURAL AND MAGNETIC PROPERTIES UPON HYDRIDATION OF SOME BORIDES OF RE2Fel4B TYPE. Si EByCHARJ, AND P. WOLFERS P3 B 4 THE USE OF THE "INHOMOGENOUS LINEAR STRUCTURE SERIES" ON THE STRUCTURAL DESCRIPTION OF SOME TRANSITION METALS COMPOUNDS YLL GRIN P3 B 5 MOESSBAUER STUDY OF THE TERNARY SYSTEM (Fe(1-x)V(x))3Ge L. HAEGGSTROEM, J. SJOESTROEM, A. NARAYANASSAMY, ~R7R7~EM P3 B 6 CRYSTAL STRUCTURE OF Nd2Col4B. MAGNETIC PROPERTIES OF Nd2Col4B AND Y2Col4B D. Lf ROUX, H. VINCENT, D. FRUCHART, P. L'HERITIER, AND'R.'FEUCHART P3 B 7 ANALOGY OF INSERTION AND STACKING WAYS IN Nd2Fel4B, Mn5SiC AND Fe5SiB2 PHASES. PH. L'HERITIER, P. CHAUDOUET, R. FRUCHART, C.B."SHOEMAKER, AND D.P. SHOEMAKER P3 B 8 STRUCTURAL AND MAGNETIC PROPERTIES OF THE NEW REMnSi2 SILICIDES (RE = La-Sro) B. MALAMAN, G. VENTURINI, M. MEOT, B. ROQUES, AND~D7~FR"UCHART P3 B 9 NEW TERNARY RARE EARTH - TRANSITION METAL GERMANIDES WITH TiMnSi2, Sc5Co4SU0 OR U2Co3Si5 TYPE STRUCURES. SUPERCONDUCTIVITY IN THESE COMPOUNDS M.MEOT, G. VENTURINI, E. MC RAE, J.F. MARECHE, B7"MALAMAN, AND B. ROQUES P3 Bie ORIGIN OF THE STRUCTURAL DISTORTION BETWEEN PHASE I1 AND I IN SnM3Rh4Snl2 COMPOUNDS S. MIRAGLIA, J.L. HODEAU, J. CHENAVAS, M. MAREZIO, £7 LAVTRON7 M7~GREDTRA7"AND J.P. REMEIKA P3 Bll INVERSION BOUNDARIES IN M783 BORIDES 2*2*. flfiSSISSkliM - KHACHFI, AND W. QANTOIS P3 812 CRYSTAL GROWTH AND TRANSPORT PROPERTIES OF THE REFRACTORY METAL DISILICIDES MoSf2 ANO WSf2. J.P. SENATEUR, 0. THOMAS, R. MAOAR, 0. LABORDE, AHD-E:~!?55EKCHER P3 B13 NEW TERNARY RARE EARTH - TRANSITION METAL GERMANIOES WITH Yb3Rh4Snl3, 8aNfSn3, U4Re7Sf6, ThCr2S?2 OR Ca8e2Ge2 - TYPE STRUCTURES. SUPERCONDUCTIVITY IN THESE COMPOUNDS G. VENTURINI, M. MEOT, M. FRANCOIS, B. MALAMAN, J.F. MARECHE, E.. MC RAE, AND B. ROQUES P3 B14 THE HIGH-TEMPERATURE MICROHARONESS OF SOME REFRACTORY MATERIALS AS MEASURED IN SINGLE CRYSTALS I. WESTMAN, T. LUNDSTROEM, M.M. KORSUKOVA, V.N. GURIN, AND~S7p7~NIKANOROV P3 815 TEMPERATURE DEPENDENCE OF THE LATTICE CONSTANTS Of TRANSITION METAL DISILICIDES AND 8. LOENNBERG

P4 A 1 STRUCTURAL INVESTIGATIONS OF BISMUTHIDES OF TITANIUM, ZIRKONIUM AND HAFNIUM. HELGA BLOCK, AND WOLFGANG JEITSCHKO P4 A 2 MnRuAs, MnRhP : PHYSICAL PROPERTIES ANO MAGNETIC STRUCTURES B. CHENEVIER, M. ANNE, M. BACMANN, 0. FRUCHART, AND"P7~CHAUOOUET P4 A 3 MAGNETIC BEHAVIOUR OF MNRUP !L CHENEVIER, M^ BACMANN, 0. FRUCHAFIT, AND P.CHAUDOUET P4 A 4 SOME UNUSUAL FEATURES ABOUT THE MAGNETIC BEHAVIOUR OF MnRhAs !L CHENEVIiS. Mi. BACMANN, AND D. FRUCHART P4 A 5 NEW CHANNEL AND CLUSTER STRUCTURES FORMED IN TERNARY ALKALINE EARTH AND LANTHAN COMPOUNDS

^1^ P.IVAN,. Rt GUERIN, M. SERGENT P4 A17 STRUCTURAL CHEMISTRY OF TERNARY SCANDIUM COBALT PHOSPHIDES i^ EIINBQLD, AND W. JEITSCHKO P4 B 1 PREPARATION AND STRUCTURE REFINEMENTS OF TERNARY AND QUATERNARY RHENIUM PNICTIDES WITH Re6P13 STRUCTURE AND W. JEITSCHKO P4 B 2 PREPARATION AND CRYSTAL STRUCTURES OF THE POLYPHOSPHIDES TiMn2P12 AND Nbre2P12 IL!L SCHOkZ, AND W. JEITSCHKO PI 8 3 EFFECT OF EXTERNAL PRESSURE ANO CHEMICAL SUBSTITUTION ON THE PHASE TRANSITIONS IN MnAs A^ ZIEBA, R. ZACH, H. FJELLVAG, AND A. KJEKSHUS P4 B 4 NONSTOICHIOMETRY IN 88-TYPE NiSb Si k£y§Sk!t H- IPSER, P. TERZIEFF, AND K.L. KOMAREK P4 8 5 MAGNETIC TRANSITIONS, SITE OCCUPANCIES AND STRUCTURAL VARIATIONS IN (Fe(l-x)Mn(x) J2P (8. 81<--K< = 8. 85): A COMBINED MOESSBAUER AND X-RAY DIFFRACTION STUDY. B.K.SRIVASTAVA, T. ERICSSON, L. HAEGGSTROEK, H.R. VERMA, AND Y.. ANDERSSON P4 B 6 A STUDY OF THE HOMOGENEITY RANGES OF THE KAPPA-PHASES IN THE Hf-Mo-(Si,P,S,GetAs,Se) SYSTEMS Ai HARSTA P4 B 7 K2PtH4, SYNTHESIS AND STRUCTURE W. BRONGER, G. AUFFERMANN, K. HOFMANN, P. MUELLER, AND H.F. FRANZEN P4 B 8 QUENCHING ANO ELECTRON IRRADIATION EFFECTS IN ORDERED BETA-PdH(D) AROUNO THE RESISTIVITY ANOMALY NEAR 50 K. J.P. BURGER, J.N. DAOU, A. LUCASSON, ANO P.. VAJDA P4 B 9 SUPERCONDUCTING PHASES OF PDH(x). S.O. OANZ, A.J. PINDOR, AND F.D. MANCHESTER P4 810 TRANSITION METAL-HYDRIDES : ELECTRONIC STRUCTURE, STRAIN EFFECT ANO H-H BINDING ENERGY J. KHALIFEH, G. MORAITIS, CH. MINOT, M.A. KHAN, AND C. OEMANGEAT P4 Bll ON THE HYDRIDING KINETICS OF La2Mgl7 AND La(2- x)Ca(x)Mgl7 M. KHRUSSANOVA, M. TERZIEVA, AND P. PESHEV P4 B12 INFLUENCE OF INERT GASES UPON HYDROGEN ISOTOPES ABSORPTION AND DESORPTION BY MraN14.5A18. 5 £«. lySCHER, P. WEINZIERL, O.J. EDER, AND E. LANZEL P4 B13 LOW-TEMPERATURE STRUCTURE OF Mg2N1H4: EVIDENCE FOR HICROTWINNING !L ZQLLIKER, K. WON, AND CH. BAERLOCHER PA B14 PREPARATION, STRUCTURE AND PROPERTIES OF Mg2CoH5 CONTAINING SQUARE-PYRAMIDAL (CoH5)(4-) ANIONS E«. 2QLLIKER, K. WON, P. FISCHER, AND J. SCHEFER PL I

TERNARY TRANSITION METAL CHALCOGENIDES WITH FRAMEWORK STRUCTURES AND THE CHARACTERIZATION OF THEIR BONDING BY MAGNETIC PROPERTIES

W. Bronger

Institut fiir Anorganische Chemie der RWTH Aachen, Professor-Pir- let-StraBe 1, D-5TOO Aachan

Zintl first pointed out, that characteristic atomic arrangements are found, when elements on the borderline between metals and nonmetals are combined with a more electropositive partner, for instance an alkali metal, to form a binary compound. This so-called Zintl principle reveals a method of synthesis by which compounds with, a decreasing linkage of structural units can be obtained: by transferring electrons from an alkali metal to the partner the structure of tha latter one is transformed into lower dimensional units in accordance with the basic idea of the (8-N)-rule.

Regarding the metal chalcogenides under discussion- the extension of the above mentioned principle to ternary and higher compounds can be illustrated. Anionic parts of a structure - in this case transition metal chalcogenide frameworks [MX ]x~ - can be obtained, if as the cationic partner an alkali metal is chosen again, transferring x electrons to the anionic framework. Thus the ternary compound has the composition A MX where A means an alkali metal, M a transition metal and X sulphur, selenium or tellurium.

In the following some examples of new ternary chalcogenides are discussed, all having tetrahedrally coordinated transition metal atoms - manganese, iron or cobalt -; the tetrahedra are linked by edges to build up [MX]-frameworks.

The insertion of paramagnetic atoms into the framework structures makes possible to get information on the nature of bonding characteristics by measuring magnetic moments. Because the [MX]-frameworks are of low dimensionality it is possible to interpretate the measurements by simple model calculations.

Measurements of magnetic susceptibilities in the temperature range from 3.7 to 295 K by the Faraday method yielded antiferromagnetism for all compounds. Neutron diffraction experiments al- PL I

lowed the direct determination of magnetic moments at temperatures helow the threedimensional ordering.

To discuss the bonding in the [MX]-frameworks, the influence of the crystal field on the electronic structure of the transition metal atom was calculated. According to the method of the strong field interelectronic interaction and spin-orbit-coupling were neglected. The ligands were replaced by point charges. The results show that the magnitude of the magnetic moments can be correlated with the crystal field splittings obtained by the model calculations. For small splittings moments are found which are near the high spin state and for large splittings those belonging to a low spin state of the transition metal atoms, which are tetrahedrally surrounded by chalcogen ligands and influenced by the next nearest metal atoms in the framework.

According to the concept of an extended Zintl principle, it was also possible to apply the synthesis of framework structures to ternary phosphides, arsenides and hydrides. Thus, for example, the

compounds Na^PtH. and K2PtH. could be synthesized and their structures were determined by neutron diffraction.

References:

W. Bronger, Angew. Chem. 92_ (1983), 12. W. Bronger and P. Miiller, J. Less-Common Met., TOO (1984), 241. PL II

INTERCALATION REaCHQNS OF TRANSITION METAL COMPOUNDS VIS ELECTRON/ ION/TRANSFER- R. SchSllhorn Institut fiir Anorganische und Analytische Chemie der TU Berlin StraBe des 17. Juni 135, D-1000 Berlin 12

The insertion of atomic or molecular species into solid materials ("host lattices") is a most interesting example of a low temperature solid reaction with particularly strong correlation between structural and physical properties of a solid and its chemical reactivity. Basically we can distinguish between three different types of reactions: (i) intercalation of neutral species, (ii) intercalation via exchange reactions and (iii) intercalation by redox reaction via electron/ion transfer (1). The focus here will be on the last mentioned type of intercalation process which has received particular attention in recent years. The fundamental reaction can be described as the simultaneous insertion of guest cations A and a corresponding amount of electrons into a solid host lattice Z: A+ + ne" + Z —- A* [ZJX~ It is obvious that essential requirements for the process are (i) the existence of an interconnected system of vacant lattice sites to allow for accommodation and diffusion of guest ions and (ii) appropiate electronic conductivity and band structure to account for election uptake and transport. The specific characteristics of a host compound will moreover be strongly affected by its structural dimensionality. Framework structures are rather rigid and allow uptake of guest cations only up to a critical size corresponding in a first approximation to the minimum diffusion channel diameter, while layer and chain type host lattices are able to adapt flexibly to the size of guest cations. The number of host lattices known to undergo electron/ion transfer reactions has increased substantially in recent years, the great majority being represented by transition metal/nonmetal compounds MX . In chemical terms this is quite naturally a consequence of the ability of transition metal ions to undergo easily changes in oxidation state. While the metal component M can be almost any transisiton metal, the number of nonmetal elements X PL II

being constitutents of host lattices is relatively small and presently essentially limited to group VIA elements. Some host lattices containing group VIA and VA elements are known as ternary phases e.g. Ta.S-C, FePS,; among the transisiton metal halides RuCl, with layer structure and RuBr, with chain structure have shown to be versatile host lattices. As a consequence of predominantly ionic bonding most binary oxides have framework structure, i.e. the number of potential guest cations is strongly limited by size restrictions. Transition metal chalcogenides, on tlie contrary,are characterized by highly polarizable anions and significant covalent M-X bonding leading frequently to layer and chain type structures,which allow the uptake of guest species with widely varying size. Also, the electronic conductivity of chalcogenides is qenerally higher as compared to oxides. For these reasons particularly sulfides and selenides of transition elements have been favorite objects of studies on intercalation reactions by redox processes in the past (1-6) and much of our present knowledge is in fact derived from investigations of these compounds. Several lines on which interest in this field has focused lately will be discussed in more detail: - the synthesis of new metastable materials via deintercalation of ternary phases according to AxMyXZ = MyXZ + XA+ + XS~ which is possible if A is a selectively mobile monovalent cation; examples are the formation of Ti,S., cubic TiS~,

VS_, CrSe«, Mo.Se3, MogSg which are not accessible by other methods of preparation. - the investigation of reaction mechanisms {in particular stacking and staging phenomena in layered chalcogenides) - the systematic study of changes in physical properties correlated with intercalation reactions - potential application of these systems e.g. as reversible electrodes in secondary batteries, electrochxomic systems for passive displays.

(1) R.Schfillhorn, Angew. Chem. 92,1015(1980);Angew.Chem.Int.Ed. Engl. J_9,983(1980); R. Schdllhorn in Inclusion Compounds, PL II

Vol.1, p.249, Eds. J.L. Atwood, J.E.D. Oavies and D.D.MacNicol, Academic Press, New Nork, 1984. (2) Intercalated Layered Materials, Physics and Chemistry of Materials with Layered Structures, Vol. VI, Ed. F. Levy, D. Reidel/ Dordrecht, Boston, 1979. (3) Physica 99B, Proc.Int.Conf.on Layered Materials and Inter- calates, Nijmegen, Netherlands, 1979, Eds. C.F. von Bruggen, C. Haas and J.H. Myron, North Holland, Amsterdam, 1980. (4) M.S. Whittingham, Prog.Solid.State Chem. 12, 41 (1978); Surv.Progr.Chem. 9_j_ 55 (1980)- (5) J. Xouxel, Rev.Inorg.Chem., 1_j_ 245 (1979). (6) Intercalation Chemistry, Eds. M.S. Whittingham and A.J. Jacob- son, Academic Press, New York, 1982. PL III

ELECTRON MICROSCOPY AND DIFFRACTION OF MODULATED STRUCTURES J. Van Landuyt University of Antwerp, RUCA Faculty of Science, Groansnborgsrlaan 171. B-2020 Antwerpen, Belgium

The interest expressed from various fields of materials reasaarch in the application of electron microscopy and electron diffraction for the acquisition of detailed structure data largely prooves its utility for the characterization of solids. Successively physicist3, metallurgists, structural chemists, ceramists and mineralogists have shown affinity and interest in this technique by which a whealth of information can be obtained from the combined use of the electron optical facilities available in a modern electron microscope. It is the purpose of this review to illustrate in particular the applications to the characterisation of modulated structures. Modulated structures can be of various origins and in many cases they can be considered as resulting from the more or less periodic occurrence of interfaces or a composition driven modulation. A classification based on the types of interfaces and their identification criteria will be outlined.

The obssrvables of modulated structures here discussed will mainly concern two aspects : electron diffraction evidence and direct space images.

The diffraction evidence on modulated structures in electron diffraction is mainly revealed by the appearance of additional spots or satellites at commensurate or incommensurate positions. These are in direct relation with a super- period in direct space. Spacing and or orientation anomalies in the position of the satellites with respect to the basic reflections can directly be interpreted in terms of the basic structure. In some cases the relative shift of satellites and matrix reflections enables the determination of the displacement vector of the periodic interfaces. The simple interpretation of electron diffraction patterns furthermore anablas a straightforward reconstruction of the reciprocal lattice of the superstructure and its relation with that of the basic structure.

The direct space evidence as revealed in the electron optical images results from the electron beam-specimen interaction transferred by the imaging lens system of the electron microscope. Since most of the interfaces which consti- tute the modulated structures will interfere with the diffraction behaviour of the electrons a "diffraction contrast image" will be observed at the appropriate magnification. Three observation modes are hereby used which yield PL III

complementary elements af information. a) the bright field diffraction contrast mods) b) the dark field mode whereby useful information is obtained by imaging in a single or mora diffracted beamsj c) the multiple beam structure imaging mode which has by nuw proven to be very powerful in revealing the ndcrostructure down to an atomic seals of structure defects and their periodic occurrence* Aa a supplementary technique optical diffraction of these high resolution images oftan yields complementary information e.g. on the occurrence of microdomains of orientation- and or translation variants of sizes below the selection capabilities of the electron diffraction. The direct space images yield information on the physical nature of the modulation on its periodicity and structural units of which it is composed.From an analysis of the defect images in combination with the electron diffraction data, structural models can be proposed relating the modulated structure with the basic structure from which it is derived. Discommensurations if present are directly revealed in the images. A few case studies will be used for illustrating the capabilities of the technique : Intra- and interpolytypic phase transitions in tantalum disulfide provide a wide variety of examples [see table]. Periodic distortions associated with charge density waves (COW) are revealed in image a3 well as diffraction observations. A movia will be projected illustrating dynamical observations of the transition between the various phases TaS2- Other examples will be used illustrating the power ana limitations of the technique for the analysis of the various types of modulated structure as there are : periodic distortions, periodic antiphase boundaries, stacking faults, twins, periodic repeats, subunit cell size structure units etc.

TaS2 LAYER STRUCTURE °K. 190° 350° 570° - T

1T 1T6 "y e shear tf. IT. - Phase

semiconductor | metal - Property superlattice| incommens.Y | incommens.B | diffusa scatt. - diffraction PL IV

ALL ELECTRON LOCAL DENSITT THEORY OF TRANSITION METAL COMPOUNDS AND IEEIR. SURFACES* A.J. Freeman, Physics Department, Northwestern University, Evanston, Illinois 6020!, U.S.A.

Recent developments in local density functional theory as applied to the properties of bulk solids and their surfaces are reviewed. Particular attention is paid to the recently developed full potential linearized augmented plane wave (FLAPW) method (1) which carries out solutions of the local density equations without shape approximations and which has recently teen demonstrated to yield highly precise and stable total energies (2). To illustrate this approach, specific examples will be presented of recent work on transition metal compounds and their surfaces with emphasis on both their structural properties and their electronic and magnetic properties.

Supported by the U.S. National Science Foundation

(!) E. Wimmer, H. Krakauer, M. Weinert, and A.J. Freeman, Phys. Rev. B. 2£, 864 (1981). (2) M. Weinert, E. Wimmer, and A.J. Freeman, Phys. Rev. B. 26_, 4571 (1982). PL V

ORDER AND DISORDER IN TRANSITION METAL CARBIDES AND NITRIDES : EXPERIMENTAL AND THEORETICAL ASPECTS C.H. de Novion and J.P. Landesman Section d'Etude des Solides Irradies, Centre d*Etudes Nucleaires, B.P. n" 6, 9226O Fontenay-aux-Roses, France

Many carbides and nitrides of transition metals can be described as a close-packed metal lattice (f.c.c. or h.c.) with the carbon or nitrogen atom in the center of the octahedral interstices. If these octahedal sites are not all occupied, the unoccupied sites may be considered as "vacancies" in the metalloid sublattice. Depending c-. the composition and the thermal treatment, these "vacancies" may be found ordered (forming a sublattice) or disordered. "Vacancies" and metalloid atoms form a pseudo "solid-solution", with order-disorder phenomena analogous to those encountered in metallic solid solutions Al-xBx- Two main such classes of compounds have been studied :

1)~ The rocksalt carbides and nitrides MCi_x and MNj_x, where the metal sublattice is f.c.c.; M is a transition metal of column in (Sc, Y), IV (Ti, Zr, Hf, Th), V (V, Nb, Ta) or VT (Mo, W). When large concentrations x of metalloid vacancies occur, several ordered structures such as described above have been discovered and crystallogrsphically described, the most important of these being : Sc^C, TijCj+jf, Z.r~iC\+v VgCf, V5C5,

NbfcCs, TijN (see review in ref. (1)), Nb4N3 (2,3), V32N26 <4). Rare-earth carbides RjC (R = Gd, Tb, Dy, Ho, Er) belong also to this category. In many samples, only short-range order of vacancies is obtained : this was first observed as a surface of diffuse intensity in the electron diffraction diagrams of TiCo.6r TiNO.6' vc O.75» NbCa8 (5). 2) The hemicarbides and heminitrides MjC and MjN, where the metal lattice is hexagonal dose-packed; M is a transition metal of column V or VI. The metalloid atoms occupy half of the available octahedral sites, in either an ordered or disordered distribution. For V;>C, Nb2C, Ta2C and their solid solutions, carbon atoms occupy alternate octahedral interstices aligned along the c axis; the existence of these structures has been recently theoretically discussed by a pair interaction model deduced from diffuse electron scattering profiles observed on quenched specimens (6).

A review of the experimental aspects of order and disorder in these systems has been published in 1977 (1). Since then, many new experimental data have been obtained, and a preliminary theoretical understanding of the origin of the long-range ordered structures has been performed (7,8). We mention some of these new results below. PLV

a) Recent neutron scattering results. The order-disorder transitions and kinetics of ordering have been recently studied by high temperature neutron diffraction for Tij Cj+X (9), Nb^Cs (1O), Ti^N (11), M02C and W2C (1Z). The Nb^Cs order-disorder transition (at Tc ~ 1O25° C) has also been detected by electrical resistivity (13). Cowley-Warren short-range order coefficients ttj were obtained for TiCo.76» TiCo.79» and NbCQ_73 from diffuse neutron scattering experiments on rapidly cooled specimens (14,16). In the studied titanium carbides, the short-range ordering is weak, when it is strong in the niobium carbide; but in all cases the a 2 coefficient is strong and negative: vacancies avoid to be second neighbours of the metalloid f.c.c. sublattice. These diffuse neutron scattering experiments show also that the metal atoms are displaced (by a few 1O~Z A) away from their vacancies first neighbours : a structural distortion occurs around the vacancies; this result was confirmed (concerning the absolute value but not the direction of displacement) in particular by channeling (15) and EXAFS (16) experiments. b) Theoretical understanding of the ordered vacancy structures in rocksalt carbides and nitrides. A first classification of these structures was made by Sauvage and Parthe (1?) in terms of various stackings of octahedra (surrounding a metal atom), the six corners of which are occupied by either a metalloid atom or a vacancy. Landesman et al (7,8) remarked that most of the experimentally observed vacancy structures are analogous to those predicted for f.c.c. metallic solid solutions by an Ising model with pair interaction energies Vj and V2 limited to first and second neighbours; carbides are found stable in the region V2 > O, Vj » Vj, and nitrides in the region v l >V2, V!>O. The pair interaction energies have been directly calculated from the electronic structure of the disordered state (8), and found in qualitative agreement with the above description and with the values deduced by the extended cluster variation method (18) from diffuse neutron scattering data (14). Calculation by recursion method of band densities of states aad energies of several

M2C ordered superstructures led to predict that in a large electron concentration range of the p.d band, the Cu-Pt type superstructure is the stablest (at OK) as experimentally observed for TijC, Zr^C and (rare-earth^C (7). PL V

(1) OH. de Novion, V. Maurice, J. de Phys. Colloq., 38, C7 - 211 (1977). (2) A.N. Cbristensen, R.G. Hazell, M.S. Lehmann, Act* Cbem. Scand. A 35, 111 (1981). (3) G. Heger, O. Baiungartaer, J. Phy«. C. : Solid State Phy». ^3, 5833 (198O). (4) T. Oaozuk*, J. AppL Cryst. 11., 132 (1978). (5) J. BUlingham, P.S. Bell, M.H. Lewis, A et a Cnrst» A 28, 6O2 (1972). (6) K. Hiraga, M. Hirabayashi, J. AppL Cryst.13, 17 (198O). (7) J.P. Landesman, P. Turchi, F. Ducastelle, G. Tregua, Mat. Res. Soc. Symp. Proc, •ol. 21 (1984), p. 363, Elsevier (New-York). (8) J.P. Landesman, G. Tregua, P. Turchi, F. Ducastelle, this conference, and to be published. (9) V. Moisy-Maurice, N. Lorenzelli, C.H. de Novion, P. Convert, Acta Met. 3O, 1769 (1982). (10) J.P. Landesman, A.N. Cbristensen, C.H. de Novion, N. Lorenzelli, P. Convert, accepted for publication in J.Phys. C : Solid State Phys. (11) A. Alamo, C.H. de Novion, 7th Int. Conf. on Transition Element Compounds, Grenoble, June 1982, p. II-A 1, and to be published (12) J. Dubois, T. Epicier, C. Esnouf, G. Fantozzi, unpublished. (13) L.C. Dy, W.S. Williams, J. Appi. Phys. 53, 8915 (1982). (14) V. Moisy-Maurice, C.H. de Novion, A.N. Christensen, W. Just, Solid State Commun. 39, 661 (1981). (15) R. Kaufmann, O. Meyer, Solid State Commun. ££_, 539 (1984). (16) V. Moisy-Maurice, Report CEA - R - 5127 (1981). (17) M. Sauvage, E. Parthé, Acta Cryst. A 28, 6O7 (1972) and A 3O, 239 (1974) (18) D. Gratias, private communication.

C.H. de NOVION SESI, Bâtiment 31, C.E.N., BP n* 6, F - 9226O FONTENAY-AUX-ROSES PL VI SYNTHESIS ANO THERMODYNAMICS OF NONSTOICHIOMETRIC RARE EARTH COMPOUNDS UNDER EXTREME CONDITIONS

E. Kaldis Laboratorium fur Festkorperphysik ETH Zurich, 8093 Zurich, Switzerland

One of the most tempting aspects of the synthesis of solid chemical compounds with interesting physical properties, is the possibility to control and optimize these properties by varying the chemical synthesis parameters. A very old method introduced by the metallurgy in order to vary the mechanical properties, the mixed crystal formation (alloying) is often used in the solid state research. In addition to that, nonstoichiometry can be used to vary (optimize) the physical properties. Three main problems appear ir. this kind of work: (a) purity, (b) crystallinity and (c) homogeneity. In most cases only a partial solution of these problems is possible. Among the various reasons for this calam- ity, a very trivial one is of tremendous importance: the lack of nonreactive container materials. This becomes particularly important as we leave the common temperature field of the synthesis of solids (T < 110CPC, due to the thermal stability of quartzglass) and move to the very interesting high melting point (m.p.) materials. In the case of the rare earth (RE) compounds, thermochemical reasons make W (and partly Mo or Ta) a suitable container material which allows reasonable solutions of the problems (a)-(c).

(a) The problem of purity. In W-columns UHV-distillation of several very reactive metals like Eu (1), Yb (and also Ca, Sr, Ba, Mg) (2) is possible, giving particularly low concentrations of the always present light elements (0, C, N, H). The purified fractions can be used for the synthesis of the corresponding compounds. To avoid contamination of the starting metals and the reaction products, gloveboxes with ultra-pure atmosphere are necessary, built by stainless steel and glass only, and gettered by hot cerium turnings, (residual atmosphere

< lppat H20; < 3ppm 02).

(b) The crystal growth problem. Crystal growth techniques for mono- to ses- quichalcogenides, -pnictides and -hydrides must be compatible with high m.p. or sublimation temperatures (up to 240CPC). At these temperatures the vapor phase is very important, so that growth under (vapor-liquid-solid) equilibrium conditions is necessary, with partial pressures of the reactive component up to se- PL VI veral tens of atmospheres. Such conditions are very well fullfilied in Targe tounqsten crucibles sealed by electron bombardment (3,4).

(c) The problem of homogeneity. Characterization of the degree of homogeneity (second phases and/or defects) is in most cases exceeding the capabilities of any single academic laboratory. As a result of it, we investigate most of the time only a few facets of the homogeneity problem. In principle, complete structural (crystal!ographic structure and lattice defects) and thermodynamic analysis are necessary for the characterization of the homogeneity of a given phase. Also here, the wall-material problem may become prohibitive. Structural characterization of most RE-compounds at high temperatures (T > 1000 °C) under chemical equilibrium conditions is not possible due to lack of x-ray permeable container or window material. Only quenching methods to room temperature, with all their disadvantages are possible. Thermodynamic studies which can be performed for many RE-compounds are

1) High Temperature Differential Thermal Analysis (HTDTA) up to 2400°C, in electron beam sealed toungsten crucibles. This allows the construction of the high temperature T-x phasediagrams e.g. Tm-Se (5), Sm-S (6).

2) Evaporation studies in UHV, using mass spectrometric investigations up to 2200^0 of a chopped molecular beam. This allows the measurement of high temperature P-T phase diagrams e.g. Yb-Te (7).

3) P-x diagrams are very important for the construction of the three-dimensional P-T-x phase diagram. By sensitive measurement of the vapor pressure they give evidence for the existence of homogeneous phases and two-phase regions at high temperatures. Unfortunately, such diagrams can be measured only for compounds with non-condensing volatile components like the RE-hy-

drides and -nitrides e.g. La-H2 (8), Ce-H2 (9), Yb-H2 (10).

For such systems the possibility appears to extend not only the temperature in the 200CPC range but also the pressure of the vapor phase in equilibrium with the solid in the kbar range. For such high-temperature, high-pressure P-T-x phase diagrams a main problem is the contamination. We hope to overcome this problem by using a differential pressure method which can be sustained by the toungsten crucible walls. PL VI (4) Enthalpies of formation are certainly necessary for the thermodynamic analysis mentioned above, as they give an important criterion for the thermodynamic stability. AH data are existing in thermodynamic tables for compounds which are readily dissolved (solution calorimetry) (5,6,11) or quantitatively oxidized (oxygen combustion calorimetry); have suitable vapor pressure equilibria or are suitable for emf measurements. However, most materials of interest for the solid state research (selenides, tellurides, phosphides, arsenides etc) are not included in these categories and therefore no AH data are existing for them. On the other hand, theoretical calculations up to now canot yield accurate AH values. It seems, therefore, necessary to expand the number of AH data by using the very difficult fluorine combustion calorimetry, as the most reactive method. Recently, we have applied this method in order to obtain information about the thermodynamic stability as a function

of nonstoichiometry for TmxSe a material with valence fluctuation and also for TmTe.

Due to lack of time, the above will be only briefly mentioned during the talk. However, several of these problems wil,l be discussed:

(a) in the frame of materials with valence instabilities using as examples our work about TmSe (5) and TmTe as well as about CeN (11). (b) in the frame of the composition dependend metal-semiconductor phase transitions in RE-hydrides. New data for the P-T-x phase diagrams of the light rare earths will be presented.

(1) G. Busch, E. Kaldis, J. Muheim and R. Bischof, J. Less. Common Met. 24, 453 (1971). — (2) J. Evers, E. Kaldis, J. Muheim, A. Weiss, J. Less. Common Met. 30, 83 (1973); 31_, 169 (1973). ~~ (3) E. Kaldis in Crystal Growth, Theory and Techniques Vol. I, p. 99, C.H.L. Goodman ed., Plenum 1974. (4) E. Kaldis, J. Crystal Growth 24-25, 53 (1974). (5) E. Kaldis, B. Fritzler, ProgrTlrTSolid State Chem. 14, 95 (1982). (6) H. Spychiger, E. Kaldis, E. Jilek in Valence Instabilities, p. 583, P. Wachter, H. Boppart eds., North-Holland (1982). (7) E. Kaldis, W. Peteler, J. Cryst. Growth 52, 125 (1981). (8) R. Bischof, E. Kaldis, M. Tellefsen, J. "Cryst. Growth (in print). (9) M. Tellefsen, R. Bischof, E. Kaldis, Proceed. 3rd Europ. Sympos. on Therm. Analysis and Calorimetry, Interlaken, Sept. 1984 (in print). (10) R. Bischof, E. Kaldis, I. Lacis, J. Less. Common Met. 94, 117 (1983). (11) E. Kaldis, B. Steinmann, B. Fritzler, E. Jilek, A. Wisard, in Rare Earths in Modern Science and Technology Vol. 3, p. 227, Me Carthy et al. eds. Plenum 1982. PL VII

RECENT DEVELOPMENTS IN SOLID STATE MAGNETISM WITH SPECIAL HEFERENCE TO TRANSITION METAL COMPOUNDS E. P. Wohlfarth Department of Mathematics, Imperial College, London SW7 2BZ, England

Solid state magnetism is one of the most complicated branches of science and technology. It is at the border lines of mathematics, physics, chemistry, metallurgy and materials science, electrical engineering, geology, zoology, medicine, etc. To help with the understanding necessary for further scientific and technological advances this subject must be studied from the broadest possible base using all the theoretical and practical techniques now available. It appears that transition metal compounds form an unusually rich class of materials for such studies and the present talk demonstrates this clearly, apparently for the first time. The subject has traditionally and unjustifiably been divided into two almost exclusive branches. Fundamental magnetism is concerned with such characteristic properties as the saturation magnetization M and Curie temperature Tc of ferromagnets and the magnetic susceptibility x(T) of paramagnets. These have been studied theoretically and experimentally to provide a scientific degree of characterization for both fundamental and applied purposes. Secondary magnetic properties are intimately connected with these primary properties and include such characteristics as the energy loss and coercive force related to the magnetic hysteresis cycle. These latter characteristics are also sensitive to materials properties and thus depend on the metallurgical treatment of the specimens. Both of these types of property have been widely investigated for the sorts of transition metal compounds with which this meeting is concerned. It is proposed to list a series of such compounds in illustration and in the prescribed order and to summarize in each case which particular property is interesting. TiH^ The magnetic susceptibility of TiHjj (1), (2) begins to develop sharper maxima at T= 3O0K when plotted as a function of the temperature and as x rises frcm about 1.70 towards 2.00. The original characterization of TiH2 as antiferromagnetic was found to be wrong as a result of neutron scattering experiments. The accepted interpretation of the sharp peak is that at this temperature the crystal structure of TiH2 changes from face centred cubic to face centred tetragonal. For the former phase a band structure calculation (3) shows the presence of a Fermi level in a region of very flat energy vs. wave vector dispersions giving a very sharp peak in the curve of the electronic density of states vs. energy. This peak leads directly to that observed in the susceptibility vs. temperature curve, since the tetragonal distortion inevitably leads to a decrease of the density of states at the Fermi level. Electronic heat data are compatible with this result (2). Fe-^B This compound has a very complicated magnetic structure, the same as that of Fe^P (4) where there are 3 inequivalent Fe-sites. As a result no band structure calculation has yet been performed. However, the compound is important in the study of amorphous alloys since these technologically Important magnetic materials have a short range order corresponding to this structure. A rudimentary tight binding band calculation which regards the Fe atoms as lying on a fee lattice has been performed (5) and shown to explain the variation with B concentration of the Curie temperature Tc of amorphous Fe-B alloys. The alloys have a potential as low loss soft magnetic materials for transformer cores etc. since the overall crystal anisotropy is zero. MB,M?B Magnetic measurements (6), for M* transition elements singly or in two fold combination, show a close correspondence with the corresponding TM alloys. The Slater-Pauling curves (giving the magnetization aa a function of composition) are reproduced in form but simply displaced by about 1 atomic species in going from TM alloy •*• M2B •*• MB. This results from an effective charge transfer mechanism. A systematic variation of the ratio of the magnetic moments obtained from PL VII

measurements above and below Tc is also obtained when plotted against the Tc of the compounds itself (7). This variation points to the occurrence of weak "itinerant" or band magnetism. Fe?C What has been called "a relatively rare hexagonal form" of this compound (8) was prepared as a fine powder and was then found to have a relatively high coercive force of at most 8OO oersted; the compounds had Curie temperatures Tc about 65OK under optimum conditions. These fine particles have permanent magnet potential due to the non-cubic crystal structure leading to large magnetic anisotropies and to the small size giving single domain behaviour. Although Fe2C has not yet been used in practice this study illustrates that potential technology can be based on fundamental magnetic concepts. Fe4W This compound has been prepared for possible use as a magnetic recording medium (9). For this purpose high coercivity, high saturation magnetization and a high Curie temperature are essential, as is the case for permanent magnet materials. The single domain character was established by electron microscopy of the fine particles produced by nitriding iron particles. The particles were elongated leading to shape anisotropy (maintained during the process) but consisted of a "stereo- network" of small crystallites about 1508 in size. The coercive forces observed reached about 8C0 oersted and the saturation magnetization about 145 emu/g if the original iron particles reached about 170 emu/g. Recording characteristics were comparable to those of more standard media. Fe2P High magnetic field and .high pressure measurements on this compound present an unusually rich set of phenomena (10), (11), (12). At zero pressure the low temperature ferromagnetism vanishes at a first order transition to paramagnetism at about 216K. On applying a sufficient pressure Fe2P (strictly stoichiometric) transforms to an antiferrcsnagnetic state with a triple point at 5 kbar and 170K. Ferromagnetism disappears completely at 13 kbar, but the antiferromagnetic state can be transformed back metamagnetically to the ferromagnetic state in a high magnetic field: At P= 14k bar, T=77K this field equalled about 6.5kOe. All these results point to an unusually interesting energy band structure (13) which it is planned to investigate. In addition Fe2P is hexagonal and the resulting large anisotropy makes it a potential permanent magnet material. Mn(As;Sb) These compounds are almost as interesting as Fe2P (14). At the As rich end ferromagnetism disappears at a "Curie temperature" in a first order way, Tc= 313K for MnAs, the crystal structure changing from hexagonal to orthorhombic. At the Sb rich end and up to 90% As the disappearance of ferromagnetism is second order, Tc for MnSb being 572K with no change of crystal structure. For the whoL? series of ferromagnetic compounds up to 90% As the pressure dependence of the Curie temperature has been interpreted (14) on the basis of the itinerant electron model giving quantitatively reasonable results. The first order transition for MnAs can no longer be interpreted in terms of a localized model (15) since a band structure calculation (16) gives a consistent picture. For MnAs metamagnetic transitions (17) will thus also have to be interpreted in band structure terms (13). MnBi This compound has been prepared in powdered form to produce magnetically hard single domain particles with a large anisotropy arising from the hexagonal crystal structure (for a review, see (18)). Coercive forces up to 12kOe have been reported. The material has never been produced on a large scale for this purpose. The same now goes for MnBi films which at one time seemed most promising for magnetooptic recording using the method of "Curie point writing" (19). There is a surprising lack of scientific data of the type just described for Hn(As,Sb). Co(S,Se)? These compounds are dichalcogenides with the pyrite structure and exhibit metamagnetic transitions from paramagnetism to ferromagnetism in high magnetic fields up to about 450 kOe for 70% S content (20). The data have been explained in terms of energy band effects (13) (20) making these materials unusually clear examples of itinerant metamagnetism. Cr(Te,X) (X-Se,Sb) These hexagonal materials are ferromagnetic and show interesting properties under high pressures (21). For X«Se dTg/dP'v-T^1 as given by itinerant electron theory; Tc vanishes at a critical pressure about 28 k bar. For PL VII

X - £b dT^dP ^-Tc which is compatible with magnetic heterogeneities on an atomic scale (22). High pressure data are thus useful in characterizing metallurgical features of this type. T1B«2 This Laves phase compound is paramagnetic with a x(T) maximum at about 10K. Again this is not evidence for antiferromagnetiaml If it is due to the fine structure of the electronic density of states curve (23) the observed but not fully resolved anomalies of the high field magnetization (24) may be correlated both with this phenomenon and with the surprising result that replacing Be by Cu produces an itinerant ferromagnet. Precise band calculations (25) also point to a great sensitivity of magnetism to the band structure. This list of transition metal compounds could be increased many times (MnP, FeGe, Fe3Si ). What appears to be of greatest interest is the interplay between the magnetic properties of these materials and their electronic and metallurgical structure.

(1) W Trzebiatowski and B. Stalinski, Bull. Acad. Pol. Sci. 1_ 131(1953) (2) F Ducastelle, R. Caudron and P. Costa, J. de Phys. 3_1_ 57 (1970) (3) M Gupta, Sol. St. Comm. 29_ 47(1979) (4) E J. Lisher, C. Wilkinson. T. Ericsson, L. Haggstrom, L. Lundgren and R. WSppling, J. Phys. C7_ 1344 (1974) (5) S N. Khanna and E. P. Wohlfarth, Physica 123B 69(1983) (6) M C. Cadeville, These Strasbourg (1965) (7) E P. Wohlfarth, Magnetismus (VEB Deutscher Verlag fur Grundstoffindustrie Leipzig) 21(1967) (8) W D. Johnston, R. R. Heikes and J. Petrolo, J. Phys. Chem. 6£ 1720 (1960) (9) S Susuki, H. Sakumoto, ¥. Omote and J. Minegishi, I.E.E.E. Trans. Mag. 2p_ 48(1984) and refs. therein (10) L. Lundgren, G. Tarmohamed, 0. Beckman, B. Carlsson and S. Rundquist, Phys. Scripta 17_ 39(1978) (11) H. Fujiwara, H. Kadomatsu, K. Tohma, H. Fujii and T. Okamoto, J. Mag. Mag. Mats. 21^ 262 (198O) (12) H, Kadomatsu, K. Tohma, H. Fujii, T. Okamoto and H. Fujiwara, Phys. Lett. 84A 442(1981) (13) E. P. Wohlfarth, High Field Magnetism (North Holland) 69(1983) (14) L, R. Edwards and L. C. Bartel, Phys. Rev. B5_ 1064(1972) (15) E P. Wohlfarth, J. Appl. Phys. 50 7542(1979) (16) R. Podloucky, J. Mag. Mag. Mats. 43_ 204(1984); J. Phys. F14 L145(1984) (discusses also other As compounds) (17) H. Xdo, T. Harada, K. Sugiyama, T. SakaJcibara and M. Date, High Field Magnetism (North Holland) 175(1983) (18) E. P. Wohlfarth, Adv. Phys. 8_ 87(1959) (19) Y. Togami, I.E.E.E. Trans. Mag. 1S_ 1233(1982) (20) K. Adachi and M. Matsui, High Field Magnetism (North Holland) 51(1983) (21) N. P. Grazhdankina and Y.. S. Bersenev, Sov. Phys. J.E.T.P. 44 775(1976) (22) D. Wagner and E. P. Wohlfarth, j. Phys. Fll 2417 (1381) (23) E. P. wohlfarth, Comm.Sol. St. Phys. 10 39(1981) (24) F. Acker, R. Huguenin, M. Pelizzone and J. L. Smith, J. Mag. Mag. Mats., to be published (25) T. Jarlborg, P. Monod and M. Peter, Sol. St. Conm. 47 889(1983)

B. P. Wohlfarth, Department of Mathematics, Imperial College, London SW7 2BZ PL VIII

CRYSTAL CHEMISTRY OF TRANSITION MEIAL BORIDES Yu. B. Kuzma Dindtscheskij Fakultet, Ul. Lomonosava 6, Luow 5, USSR PL DC

STRUCTURE, DEFECTS AMD PROPERTIES OP SOME REFRACTORY BORIDES

T. Lundstrom Institute of Chemistry, University of Uppsala, Box 531, S-751 21 Uppsala, Sweden

The refractory borides display several unique properties, which for a long period of time have been the subject of fundamental studies (1). Many of these properties are of great interest for technical applications. The borides are characterized by high melting points, great hardness and, most of them, good electrical and thermal conductivity. Further characteristics are a fair corrosion resistance, good wear resistance and a thermal 3hock resistance much better than that of oxide ceramics. Some of them are good thermionic emitters. There are a few well established applications, based on the above-mentioned properties, but the number of proposed applications is very extensive indeed.

The preparation of refractory borides often requires very high temperatures and as a consequence it is difficult to prepare homogeneous, single phase "borides of high purity that are desirable for basic research. On the other hand high—tenperature powder-metallurgical techniques are attractive for industrial manufacture. Borides prepared by methods, which do not permit close control over the composition, purity and crystal size of the product may show considerable variations in properties. It i3 very interesting to know what type of defect occurs in a specific case (impurity atoms, vacancies, twinning etc.) and how the defects influence the properties of the boride (2). In the present contribution t. review is presented of boride structures, whose defects have been investigated in detail and/or whose properties have been studied fairly well.

Borides of the composition MfB^

Borides of the composition M_B, crystallize in the Ru_B, type structure determined by X-ray diffraction techniques (3,*0. The structure of the closely related carbides Cr-C,, Mn_B^ and Fe_C, have recently been studied in detail using electron microscopy and diffraction (5-7). The carbides contain numerous defects (twins, anti-phase boundaries), which completely mask the real structure. The structures of the M_B_ borides (M = Ru, Rh, Re) have later been investigated using the same experimental techniques (8-10). The borides display only a concentration of defects, which establishes the hexagonal Ru_B~ as the true structure of M-B, (10). A very rare type of planar defect was identified in the borides, namely an inversion boundary between two parts of a crystal (10). PL IX

Monoborides

The main structure types of the monoborides are those of CrB, FeB and MoB. The structures contain zigzag chains of boron atoms and the difference between the structures originates in the different arrangements of these boron chains. It is convenient to describe the relationship between the structure types by vectorial transpositions of structural building blocks (11). Polymorphs of the same mono- boride phase often adopt tvo of these structure types (12-1U). At temperatures close to that at which the phase transformation takes place such a phase displays intermediate atomic arrangements or ordered intergrowth. Information on these defects hare been obtained from the broadening of the diffraction lines. Defects of these types have been demonstrated for WB, FeB, MnB and CrB.

Diborides

Several of the transition metal diborides, crystallizing in the A1B_ type structure, have extended homogeneity ranges, e.g. NbBg, TaB_ and MoBp. As mechanisms behind the composition changes has been proposed vacancies at the metal or boron positions as well as interstitial boron atoms (2). Definite experimental evidence is, however, lacking. Studies of microhardness within the homogeneity ranges have been reported for NbBp and TaB2 (15). Defects in other structures related to that of AlBp (Mo, B,, WB» , RUpB,, Rh^B,-) were recently discussed by the present author (2) but very little is known about the properties of these phases.

Hexaborides of the rare earth metals

The withs of the homogeneity ranges of the rare earth hexaborides and the mechanisms by which the composition changes take place have not been determined accurately for most hexaborides. It seems that most homogeneity ranges are relatively- small (16). Extensive homogeneity regimes have, however, been reported for SmBg (17,18) and for ThBg (19). The latter was studied in detail using neutron diffraction and it was demonstrated that the composition changes are connected with vacancies at the metal position (0-22 % vacancies). Reports on the ranges of homogeneity of LaBg have established that no variations in the cell dimensions occur (16,18). Recent diffraction studies indicate that the homogeneity range is very small and that vacancies at the boron position are probable (20). PL IX

References

1. V.T. MatkoTich (Ed.), Boron and Refractory Borides. Springer-Verlag, Berlin, 1977. 2. T. Lundstrom, in R.K. Viswanadnam, D.J. Hovcliffe and J. Gurland (Eds.), Science of Hard Materials, Plenum, Hev ïork, 1983, p. 219» 3. B. Aroasson, Acta Chem. Scand. J^ (1959), 109. 1*. B. Aronsaon, E. Stenberg and J". Aselius, Acta Chea. Scand. Jjt (1960), 733. 5. E. Bauer-Grosse, J.P. Morniroli, G. LeCaër and C. Frantz, Acta Metall. 2J9 (1981)» 1983. 6. J.P. Morniroli and M. Gantois, J. Appi. Cryst. _1_6 (1983), 1. 7. J.P. Morniroli, E. Bauer-Grosse and M. Gantois, Phil. Mag.. k8_ (1983), 311. 8. M. Khachfi, E. Bauer-Grosse, J.P. Morniroli, T. Lundstrom and M. Gantois, Rev. Chim. Min. 21 (1984), 370. 9. M. Khachfi, E. Bauer-Grosse, J.P. Morniroli, T. Lundström and M. Gantois, Proc. of the VIII International Symposium on Boron, Borides, Carbides, Nitrides and Related Compounds, held in Tbilisi, USSR, October 8-12, 1984. In the press. 10. J. Morniroli, M. Khachfi and M. Gantois, Contribution to the VIII. Int. Conf. on Solid Compounds of Transition Elements, Vienna, April 9-13, 1985- 11. H. Boiler and S. Parthé, Acta Cryst. _i£ ( 1963 ), 1095. 12. H. Boiler, W. Rieger and H. Nowotny, Monatsh. Chem. 95. (196 ), 1497. 13. T. KanaizuJsa, J. Solid State Chem. h± (1982), 195. 14. I. Smid and P. Rogl, Proc. from II. International Conference on Hard Materials, held in Rhodos, Greece, September 23-28, 1984. In the press. 15. T. Lundstrom, B. Lönnberg and I. Westman, J. Less-Common Met. 9§_ (1984), 229. 16. S.E. Spear, in A.M. Alper (Ed.), Phase Diagrams: Materials Science and Technology, Vol. IV. Academic Press, New York, 1976, p. 91. 17. K. Niihara, Bull. Chem. Soc. Japan kh_ (1971), 963. 18. Yu. B. Paderno and T. Lundstrom, Acta Chem. Scand. A37 (1983), 609. 19. J. Etourneau, J.-P. Mercurio, R. Haslain and P. Hagenmuller, J. Solid State Chem. 2 (1970), 332. 20. M.M. Korsukova, T. Lundstrom, V.N. Gurin and L.-E. Tergenius, Z. Krist. In the press. PL IX

Monoborides

The main structure types of the monoborides are those of CrB, FeB and MoB. The structures contain zi/jzag chains of boron atoms and the difference between the structures originates in. the different arrangements of these boron chains. It is convenient to describe the relationship between the structure types by vectorial transpositions of structural building blocks (11). Polymorphs of the same mono- boride phase often adopt two of these structure types (12—1U). At temperatures close to that at which the phase transformation takes place such a phase displays intermediate atomic arrangements or ordered intergrowth. Information on these defects have been obtained from the broadening of the diffraction lines. Defects of these types have been demonstrated for WB, FeB, MnB and CrB.

Diborides

Several of the transition metal diborides, crystallizing in the AlBg type structure, have extended homogeneity ranges, e.g. KbBp, TaBp and MoBp. As mechanisms behind the composition changes has been proposed vacancies at the metal or boron positions as well as interstitial boron atoms (2). Definite experimental evidence is, however, lacking. Studies of microhardness within the homogeneity ranges have been reported for NbBp and TaB« (15)> Defects in other structures related to that of AlBg (Mo, B,, WB2 0, RUpB,, Rh^B^) were recently discussed by the present author (2) but very little is known about the properties of these phases.

Hexaborides of the rare earth metals

The withs of the homogeneity ranges of the rare earth hexaborides and the mechanisms by which the composition changes take place have not been determined accurately for most hexaborides. It seems that most homogeneity ranges are relatively small (16). Extensive homogeneity regimes have, however, been reported for SmBg (17,18) and for ThBg (19). The latter was studied in detail using neutron diffraction and it was demonstrated that the composition changes are connected with vacancies at the metal position (0-22 % vacancies). Reports on the ranges of homogeneity of L&Bg have established that no variations in the cell dimensions occur (16,18). Recent diffraction studies indicate that the homogeneity range is very and that vacancies at the boron position are probable (20). PL IX

References

1. V.T. Matkovich (Ed.), Boron and Refractory Borides. Springer-Verlag, Berlin, 1977. 2. To Lundstrdm, in R.K. Viswanadham, D.J. Rowcliffe and J. Gurland (Eds.), Science of Hard Materials, Plenum, Nev York, 1983, p. 219. 3. B. Aronsson, Acta CTuwn- Scand. _13_ (1959)* 109. U. B. Aronsson,. E. Stenberg and J. Aselius, Act a Chen. Scand. _iU. (i960), 733. 5. E. Bauer-Grosse, J.P. Morniroli, G. LeCaer and C. Frantz, Acta Metall. 2£ (1981), 1983. 6. J.P. Morniroli and M. Gantois, J. Anspl. Cryst. \6_ (1983), 1. 7. J.P. Morniroli, E. Bauer-Grosse and M. Gantois, Phil. Mag. U8 (1933), 311. 8. M. Khachfi, E. Bauer-Grosse, J.P. Morniroli, T. Lundstrom and M. Gantois, Rev. Chim. Min. 2j_ (198U), 370. 9. M. Khachfi, E. Bauer-Grosse, J.P. Morniroli, T. Lundstrom and M. Gantois, Proc. of the VIII International Symposium on Boron, Borides, Carbides, nitrides and Related Compounds, held in Tbilisi, USSR, October 8-12, 1981*. In the press. 10. J. Morniroli, M. Khachfi and M. Gantois, Contribution to the VIII. Int. Conf. on Solid Compounds of Transition Elements, Vienna, April 9-13, 1985. 11. H. Boiler and E. Parthi, Acta Cryst. J_6 (1963), 1095. 12. H. Boiler, W. Rieger and H. Nowotny, Monatsh. Chem. £5_ (196 ), 1U97. 13. T. Kanaizufca, J. Solid State Chem. Ui_ (1982), 195. Ik. I. Smid and P. Rogl, Proc. from II. International Conference on Hard Materials, held in Rhodos, Greece, September 23-28, 198k. In the press. 15. T. Lundstrom, B. LSnnberg and I. Westman, J. Less-Common Met. 9|6 (198U), 229. 16. K.E. Spear, in A.M. Alper (Ed.), Phase Diagrams: Materials Science and Technology, Vol. IV. Academic Press, Hew York, 1976, p. 91. 17. K. Uiihara, Bull. Chem. Soc. Japan hh_ (1971), 963. 18. Yu. B. Paderno and T. Lundstrom, Acta Chem. Scand. AJf (1983), 609. 19. J. Etoumeeu, J.-P. Mercurio, R. Naslain and P. Hagenmuller, J. Solid State Chem. 2 (1970), 332. 20. M.M. Korsukova, T. Lundstrom, V.H. Gurin and L.-E. Tergenius, Z. Krist. In the press. PL X

NEW EXAMPLES OF CLUSTERS, EXTENDED METAL-METAL BONDING AND INTERSTITIAL DERIVATIVES John D. Corbett Department of Chemistry and Ames Laboratory-DOE, Iowa State University, Ames, Iowa 50011, USA

Synthesis of many new compounds with unusual stoichiometries and with' structures that exhibit strong metal-metal bonding has been achieved in the last 10 to 12 years in binary halides of the early transition metals. Such characteristics had previously been limited to metal octahedra in the well- n+ known (Nb,Ta)6X12 and (Mo.VOgXg"* halide clusters (6-12 and 6-8 types, respectively; X = Cl, Br, I). This new chemistry is found for highly reduced binary halides of the metals Sc, Y, La, Zr, Kf and some lanthanides. The metal-metal arrays in these may be classified as conventional 6-12 or 6-8 type clusters that have, in response to reduced X:M ratios, condensed by sharing metal-metal edges to form infinite chains or sheets of clusters. The structures and properties thus lie along the interface between metals and normal salts (1,2,3). The use of welded niobium or tantalum containers has been vital to these discoveries, most of the products being achieved in relatively long-term reactions at 600-1000'C. Some of the compounds have been found to bind a variety of small nonmetals within the metal interstices while others ire uniquely stabilized by such interstitial atoms. Structure and bonding correlations as well as valence photoelectron spectra and extended HOckel calculations will also be considered.

A section of the infinite chains in Gd2Cl3 - the first example discovered - is shown in Figure 1, these being derived from hypothetical metal octahedra

Fig. 1 PL X

face-capped by hal ide. Generally all of these metal-metal bonded arrays are well-sheathed by halide and are bonded together by halIde bridging between metal vertices. The alternate construction of infinite chains from 6-12 type

clusters (where halides bridge edges) occurs in Sc5C18 (Figure 2); here there

Fig. 2

is also a parallel chain of condensed Sc^Clg octahedra. Several other arrangements of infinite chains are known. Interchain condensation in the limit generates the novel metallic ZrX (X * Cl, Br) with strongly bound, infinite double-metal layers (Figure 3). The transition from normal to more

Fig. 3

metal-like structures that accompanies reduction in such a series is reflected in photoeiectron valence spectra, coordination number changes and metal-metal bond orders achieved. Contrasts with the chaicogenides will be noted. Interstitial atoms have been found to stabilize a number of scandium and zirconium clusters, several of which occur in structure types already known for the electron-richer (and empty) niobium and tantalum clusters. An isoelec-

tronic series occurs with Zr6ClX2Be, Zr6C1llfC, Zr6ClisN, Sc(Sc6C112N) and

CsZr6Cl14B. The character of the interstitial atom bonding in these will be considered. Interstitial carbon atoms likewise stabilize Sc7Cl10C2, with carbon atoms in all metal octahedra in double metal chains (Figure 4). The PL X

Fig. 4

carbon-free Sc7Cl10 has a related but distinctly different structure. The double-metal-layered monohalides show a prolific interstitial chemistry

involving H, B, C I and 0, viz., Zr2Cl2B, M2C12C (M = Sc, Y, Zr) and Zr2Cl2N with nonmetals in nominal metal octahedra and ZrClOo.i, and MCIH with tetrahedral occupancy. Bonding of interstitial atoms in these is similar to that in clusters. The Hel photoelectron spectrum of Zr2C!2C together with the superimposed density-of-states from an EHMO calculation are shown in Figure 5.

Finally, there is the challenge presented by the discovery of Zr6I^K - Figure 6 - in which the alkali metal atom is bound within the metal cluster. Factors

Fig. 6

related to its remarkable stability will be considered.

(1) J. D. Corbett, Adv. Chem. Ser., _186, 329 (1980). (2) J. 0. Corbett, Ace. Chem. Res., ^14, 239 (1981). (3) J. 0. Corbett, Pure Appl. Chem. 56, 1527 (1984).

Professor J. D. Corbett Department of Chemistry Iowa State University Antes, Iowa 50011 0 1

THE a TRANSITION IN PURE AND DOPED FeS : PHYSICAL AND STRUCTURAL ASPECTS G. Collin, M.P. Gardette, F. Keller-Besrest Laboratoire de Chimie Minerale Structurale Associe au CNRS - LA 200, 4, Av, de 1'Observatoire, 75270 Paris, France.

The a transition in pure and doped FeS is identified as a transition from a non-degenerated polaron gas at low temperature to a degenerated polaron gas at high temperature (1). This transition is accompanied by an increase of the electrical conductivity due to the energy lowering of the polarons and to a decrease of the magnetic exchange in the antiferromagnetic order. From the structural point of view the phase transition is characterized by a change from the low temperature superstructure (a\3, 2c) (2) to the high temperature one (2a,c) (a,c parameters of the NiAs type substructure) at T . This transition can be: - either temperature induced for FeS or alloyed materials with impurity levels deeply trapped such as Fe. Mn S, Fe, Co S. I —"X X I — X X - or impurity induced for materials in which electronic levels are populated in the band gap of FeS. In this latter case there is a critical concentration beyond which the polaron gas becomes degenerated. The experimental results are in agreement K = 08 for x 2+ one with the Mott predictions with cr±t'cal - cr ^ Fe s two polaron per chromium) and = .04 for i_xQx < polarons per vacancy). I - In the low temperature (L.T.) phase, the effect of temperature or of alloying is to introduce disorder in the L.T. correlation giving rise to disordered regions. This leads to a simultaneous reduction of the L.T. contribution which enhances the preexponential o\_ factor in the conductivity law, and of the coherence length which reduces the dissociation energy Ej of the polaron. The polaron conductivity can then be interpreted in terms of a constant hopping energy En and of a temperature dependent dissociation energy E*, this latter vanishing at T . 0 1

II - At high temperature (T > T) the conductivity is that of a degenerated polaron gas in which the hopping activation energy E_ remains only. This H.T. phase can be stabilized at room temperature by alloying, either with CrS (.08 c x <. .015) or with CoS (.15 £. x) rendering more convenient its study than with FeS itself, especially in the case of CoS showing a perfectly long range order superstructure. Under these conditions it can be confirmed that the crystal structure of this phase is hexagonal (2a,c) and that the tri-twinned "MnP type" model (3) cannot be retained. Especially the (k k o) reflections of the (2a,c) phase are perfectly observable in both kinds of materials (CoS and CrS) . This (2a,c) phase appears to be very sensitive indeed: In FeS.the superstructure is easily destroyed either by small amounts of vacancies or upon heating. For Fe,» Cr S and

Fe, O S with x > x ..ical the superstructure becomes unstable and produces the apparition of incommensurate short range order characteristic of the degenerated polaron gas instability. On the contrary cobalt alloying - with the overlap of Co 3d levels with the sulphur p band -, stabilizes the phase over a wide range of temperature and composition. Ill - We will present experimental results concerning transport properties and magnetic susceptibilities versus temperature or composition. We will also show from single crystal diffraction experiments the lattice constant variations and the changes in coherence length of the different kinds of reflections : contrasted behaviour of the substructure reflections retaining large coherence lengths, and of the superstructure reflections - low or high temperature type - which on the contrary are broadened by alloying and upon heating. Another kind of information will be provided by the changes in intensity of these different types of reflections with temperature, especially around the a transition. Finally results of conventional crystal structure determinations and of X ray diffuse scattering experiments will be presented. 0 1

All these results, physical and structural, will be used to establish the disorder models and to propose a quantitative confirmation of the polaron conductivity curves.

(1) G. Collin, M.F. Gardette, G. Keller, R. Comes Accepted by J. of Phys. and Chem. of Solids under press (2) E.F. Bertaut, Bu-11. Soc. Fr. Miner. Crist. 22/ 293 (1956) (3) H.C. King, C.T. Prewitt, Acta Cryst. B38, 1877 (1982)

Mr. G. Collin Laboratoire de Chimie Minerale Structurale, 4, Av. de L'Observatoire 75270 Paris Cedex 06, France 0 2

HIGH PRESSURE X-RAY DIFFRACTION STUDY ON

p-FeS2, m-FeS2 AND MnS2 UPTO 340 KBAR T. Chattopadhyay*. H. Fjellvag and H.G. von Schnering Max-Planck-Institut fiir Festkorperforschung, Heisenbergstrasse 1, D-7000 Stuttgart 80, Federal Republic of Germany * DRF/DN, Centre d'Etudes Nucleaires, 85 X, 38041, Grenoble Cedex, France

A large number of transition metal dichalcogenides and dipnictides with the general formula TX2 or TXY crystallize with closely related pyrite, marcasite and arsenopyrite structures |1|. These compounds have varied electrical and magnetic properties |l,2|. The chemical bonding in these compounds has been discussed by various authors |l-5| and band structures of a few of them have been calculated |6-8|. In these structures the transition metal is surrounded by six non-metal atoms, forming a distorted octahedron and the non-metal atom is bonded to three transition metal atoms and to another non-metal atom forming a distorted tetrahsdral arrangement. The existence of X2 or XY dumb-bells is typical for these structures. The relationships between these structures have been discussed in detail by Brostigen and Kjekshus |9|. These authors have also suggested a hypothetical transformation route from the pyrite to the marcasite structure. In the present investigation we have studied the effect of hydrostatic pressure upto 340 kbar on the crystal structure of p-FeS2i m-FeS2 and p-MnS2. Fe2+ in p-FeS2 and m-FeS2 is in the low-spin state whereas Mn2+ in MnS2 is in the high spin state. High spin -*• low spin transition involves volume contraction |10|. Therefore, application of hydrostatic pressure in MnSj is likely to induce high spin -•• low spin transition. We have undertaken the present high pressure X-ray investigation to check this. We have used natural crystals of FeS2 (pyrite and marcasite) and MnS2 (heurite) of high purity for the investigation. These natural single crystals were ground to fine powders for the high pressure X-ray investigation. High pressures were generated in a gasketed diamond anvil cell |ll|. The pressure generated was measured by the well-known ruby fluorescence technique |12|. X-ray diffraction patterns were obtained at room temperature in the energy dispersive mode.

Fig. 1 shows the variation of the reduced volume V/Vo of p-FeS2 and m-FeS2 as a function of pressure obtained from the high pressure X-ray investigation. No structural phase transition has been observed for these two compounds. m-FeS2 is found to be slightly more compressible than p-FeS2 • Fig. 2 shows the reduced volume V/Vo of MnS2 as a function of pressure. At about 140 kbar a structural phase transition is observed. The high pressure phase has been interpreted to have a marcasite type structure. With this interpretation the volume discontinuity is found to be about 15 %. The high pressure phase of MnS2 has much lower compressibility compared to that of p-MnS2 and has comparable compressi- 0 2

1.0

100 200 300 400 100 200 300 400 500 P(kbar) P(kbar)

Fig. 1. Variation of the reduced volume V/Vo of Fig. 2. Variation of the reduced volume V/Vo of the pyrite p-FeSj and the marcasite m-FeSj as i MnSj as a function of pressure. The discontinuity function of pressure.m-FeSg is more compressible in volume {about 15 X) corresponds to the first-order than p-FeSj p-MnS2 •+ ra-MnSj structural phase transition

(a) p-MnSj P.21.3 kbor (c) ! m-FeS3 P«l16kbor 28.8* • 26

: i 1 i -020*101 L. • i (A

-Mo *-200 220-^ -II 3 • K0'/2 1 -110 111 120 inc 1 II I 111 210 211 | : 311 222 ('(ncl 1 1 - '\J w 30 40 50 1 20 20 30 A40 50

Fig. 3. Energy dispersive high-pressure X-ray

diffraction spectra of : (a) p-HnS2 (21.3 kbar), (b) in-MnSj (290 kbar) and (c) ra-FeS2 (116 kbar). A conparison of spectra (b) and (c) indicates that the high-pressure HnSj has the marcasita type structure

30 40 SO E (KlV) 0 2

bility to those of p-FeSj and m-FeS2. In Fig. 3 we compare the energy dispersive X- ray diffraction spectra of (a) p-MnSj (21.3 kbar), (b) m-MnS2 (290 kbar) and (c) m-FeS2 (116 kbar). The spectra (b) and (c) look very similar and this led us to index the diffraction lines with an orthorhombic marcasite unit cell. The axial ratios c/a = 0.76 and c/b = 0.62 correspond well to the class B marcasite |11. Mn2+ in p-MnSg is found to be in the high spin state and has a moment of 5ug corresponding to the five unpaired electrons. It is known that high spin •»• low spin transition involves large volume contraction |l0|. It is therefore likely that p-MnS2 to m-MnS2 structural phase transition is accompanied by a high spin •+• low spin transition of Mn2+. A further indirect evidence in favour of this assumption is the Mossbauer resonance studies of Bargeron et al. |13| on the effect of pressure on the spin state of iron(n) as a dilute substitutional impurity in manganese sulfide. These authors found that the substitute Fe2+ in MnS2 is in the high spin state and the application of pressure initiates the high spin -*• low spin transition at 40 kbar which is completed at about 120 kbar. Reversely, one would expect that the Mn2+ substituted in p-FeS2 or m-FeS2 would be subjected to a chemical pressure leading to a low spin state of the substituted Mn2+. We are now performing similar high pressure X-ray diffraction study on MnSe2 and and the results will be presented at the Conference.

(1) G. Brostigen and A. Kjekshus, Acta Chem. Scand. J24, 2993 (1970) (2) F. Hulliger and E. Mooser, J. Phys. Chem. Solids 26, 429 (1965) (3) J.B. Goodenough, J. Solid State Chem. 5, 144 (1972) (4) J.A. Tossel, D.J. Vaughan and K.J. Burdett, Phys. Chem. Minerals 7, 177 (1981) (5) J.K. Burdett and T.J. McLarnan, Inorg. Chem. n, 1119 (1982) (6) M.A. Khan, J. Phys. C : Solid State Phys. 9, 81 (1976) (7) D.W. Bullet, J. Phys. C : Solid State Phys. 15_, 6163 (1982) (8) S.D. Wijeyesekera and R.Hoffmann, Inorg. Chem. 2Z, 3287 (1983) (9) G. Brostigen and A. Kjekshus, Acta Chem. Scand. 24, 2983 (1970) (10) A. Kjekshus and D.G. Nicholson, Acta Chem. Scand. .25, 866 (1971) (ll)W.B. flolzapfel, High Pressure Chemistry (Edited by H. Kelm), Reidel, Amsterdam (1978) (12) J.C. Barett, S. Block and G.J. Piermarini, Rev. Scient. Instrum. 44,1 (1973) (13)C.B. Bargeron, M. Avinor and H.G. Drickamer, Inorg. Chem. 10,1338 (1971) 0 3

CRYSTALS GROWTH, CHARACTERIZATION AND SOME PHYSICAL PROPERTIES OF RuS Hatern E2ZA0UIA, Jonathan Foise, Ouri Gorochov Laboratoire de Physique des Solides, C.N.R.S., 1 place Aristlde Briand 92195 MEUDON PRINCIPAL Cedex (FRANCE)

RuS. crystallizes with the pyrite structure (1). It is an exceptional material with regard to chemical stability, hardness and photoelectrochemical stability. Until recently, attempts to grow single crystals from classical methods have been unsuccessful.One of che authors has proposed a new technique of crystal growth using tellurium as a flux (2)(3) and also chemical vapor transport with IC1, has recently been reported (4). Nevertheless, in our group we have not been abLe to reproduce this second technique. We report here the crystal growth from tellurium flux by two techniques. The first involves the slow cooling of a silica tube (holder) containing a large excess of tellurium (up to 6 grams) and small amounts of RuS. and sulfur. Some typical conditions are given in Table 1.

Grams of Grams of Grams of" X-Ray analysis Microprobe analysis

tellurium RuS2 sulfur

6 0.030 0 RuTe2 Pure RuTe. 3 0.015 0.005 Mixture of RuS. RuS. pure, RuTe. and RuTe. witn 4% S

3 0.045 0 Almost pure RuS. RuS2 with 6% Te

6 0.030 0.010 Pure RuS. RuS2 with 27. Te

All of the above cooled from 1000° to 840°C at 1.5°/hour

Several temperature conditions and rates of cooling have been examined. The second technique involves concentration of RuS- in the flux by slow and conti- nous evaporation of tellurium in a temperature gradient (3). The halogen growth has been accomplished with chlorine as the transporting agent but with addition

of oxygen or RuO. to the RuS2 powder. It is clear that RuO- acts as an intermediate phase, which is necessary to obtain RuS- crystals. Oxygen concentration in these crystals is less than can be detected by microprobe ( < .57.). RuS- is a diamagnetic material according to the magnetic susceptibility (4.2 - 300K). Only a small paramagnetic contribution of impurities is observed below 10°K. The electrical properties (resistivity and Hall measurements) indicates that the charge carriers range from 10 - 10 cm" and that Hall mobility varies from 100 - 400 cm V s . Band gap determination from optical absorption 0 3

and from photocurrent vs. wavelength indicate that for pure samples of RuS_, 1.8ev Is measured. In some cases, lower values, down to 1.4ev, are related to the presence of impurities (ex: Te). An initial photoelectrochemical study has shown that n-type RuS- is a stable photoelectrode in various aqueous solutions, even in the absence of reversible redox couples (5). Nevertheless the oxygen evolution is slow below 1.2V/SCE and this can be related to strong recombination of hole - electron pairs at the surface or close to it. Several surface treatments have been examined to eliminate this problems, p-type RuS- crystals, obtained in critical conditions, were also studied. Our ultimate goal is to synthesize stable and efficient photoanodes and photo- cathodes for photoelectrolysis of water. This material is also of interest as regards to the catalytic properties of ruthenium.

(1) Sutarno, O.Knop and K.J.G. Reid, Can. Journ. of Chem., 45, 1391 (1967) (2) H.Ezzaouia, These de 3eme cycle, Paris XI, 9-7-1982. (3) H. Ezzaouia, R. Meindl and J. Loriers, J.of Mat. Science (to be published) (4) R. Bichel, F.Levy and H.Berger, J. Phys. C. Solid State Phys. 1]_, L19 (1984) (5) H. Ezzaouia, R. Heindl, R. Parsons and H. Tributsch, J. Electroanal. Chem. 165, 155 (1984).

Ouri GOROCHOV, C.N.R.S., Laboratoire de Physique des Solides 1 place A. Briand, 92195 MEUDON PRINCIPAL CEDEX 0 4

IN-SITU X-RAY CHARACTERIZATION OF THE INTERCALATED ZIRCONIUM DISELENIDE Lx2 C.LEVY-CLEMENT and J.RIOUX C.N.R.S. Laboratoire de Physique des Soltdes, 1 place Aristide Briand, 92195 MEUDON (France) W.R.McKINNON and J.R. DAHN National Research Council, Solid State Chemistry, Division of Chemistry, KIA 0R6 OTTAWA (Canada)

The lamellar semiconductor ZrSe_ belongs to the same IV B transition metal chalcogenide family as TiS. which is a good cathode for secondary lithium intercalation batteries. It has been shown that lithium also intercalates into ZrSe, (1) but results published on the physical properties of Li ZrSe. are conflicting. Berthier et aL. (2) found, working on powder samples, that upon intercalation of lithium the IT ZrSe_ structure was conserved, that no variation of the a and c lattice parameters was observed in the 0 < x < n,4 concentration range and that a nonmetal transition occured for x = 0.4. However, transport measurements done by Onuki et al. (3) and Klipstein (4) on Li ZrSe_ single crystals indicated a metallic behavior for all values of x including x = 0.

By studying with an in-situ X-ray diffraction technique a Li/Li ZrSe_ electroche- mical cell we effectively observe that in the 0 < x< 1 concentration range the IT structure remains but also that as soon as lithium intercalates the lattice parameters vary with a decreasing while c increases with x. We also show here that the intercalation is reversible.

We constructed Li/electrolyte/ZrSe. cells as described previously (5). To avoid the co-intarcalation of propylene carbonate (PC) which occurs in the compound of this MX_ family, a nearly saturated solution of LiAsF, in PC was used (6).The cells wer'; charged and discharged at constant current while the voltage V was measured (7). The quantity x, the number of Li transferred to the ZrSe, cathode per mole of ZrSe_ was calculated from the current and the cathode mass. Using in-situ X-ray diffraction from cells with berylium X-ray windows we studied the evolution of the structure of the host compound while variyng the amount of intercalated lithium. Homogeneous samples can be prepared easily in the X-ray beam by fixing the cell's voltage and waiting for the intercalation to reach equilibrium. As a result, changes in the intensity of Bragg peaks reflect changes within Li ZrSe. and not simply changes in the amount of material in the beam (8). 0 4

Highly crystalline ZrSe. powder was prepared by reacting a mixture in the ratio of 1 mole of Zr (99.9%) and 1.01 mole of Se (99.999%) in sealed quartz ampoule at 900°C for 1 week. The ampoules were quenched in air. The lattice

+ o o parameters a » 3.7732 - .0005A and c = 6.1315 + .005A are in good agreement with published values (9). , th The voltage versus x for the 4 discharge and charge for a Li/Li ZrSe, cell is shown on figure 1. The current correspond to a change flx = 1 in 48 hours. Below 1.3V the cell voltage drops rapidly with x presumably because all the octahedral sites are filled (1). Since the filling of octahedral sites should occurs at x = 1, the values of x shown in figure 1 have been normalized to x = 1 at 1.3V. The value of x at 1.3 volt calculated from charge transfer was x = 0.76 O.n not 1 because part of the intercalation x in UxZrSe2 electrode was not in electrical contact with the substrate and hence did not Fig.l - V(x) for Li/LixZrSe2 cells. intercalate. The offset of charge and

discharge is most likely due to the poor kinetics in the electrolyte because it is saturated.

Portions of X-ray diffraction profiles taken during the first discharge at a 50 hour rate of an in-situ Li/Li ZrSe x 2 X-ray diffraction cell are shown in figure 2. The values of x, not corrected for the inactive part of the cathode, are from 'ne back, 0.0086, 0.048, 0.088, 0.1274, 0.1667, 0.2062, 0.J458, 0.2835. 41 4.1 4» Scald-ring Anglo (Ucg) The cathode utilization was almost total because at the end of the eighth scan the Fig.2 - In situ X-ray diffraction cell voltage was 1.675V which according profiles,The Miller indices of to figure 1 corresponds to x = 0.29. peaks are indicated. As x increases Bragg peaks are shifted. For example; 001 peaks are shifted toward smaller diffraction angles indicating an increase of the c parameter and hhO moves in the other direction corresponding to a decrease of a. The unaffected (003) and (102) peaks are referred*. The variations of a and c obtained from 0 4

0.7 U.7H equilibrium measurements at fixed voltage, as well as from increasing and decreasing Che voltage, on the same cell, are shown fl.fi :).7B on figure 3. There was no evidence that there was hysteresis. Using the values of G.3 x determined from the average voltage values on figure 1 the variations of a 6.J and c versus x are given in figure 4. 1.1 l.fi I.II ;c (V) All the data points were obtained by least squares fits to the positions of fig.3 - Variation of the lattice at least 8 Bragg peaks for each data parameters a( ) and c (.) versus point. Data from figure 4 show that the voltage of the cell. for 0 < x< 1 our a and c values are in agreement with values reported in (2) G.7 , •-•^• but for 0

H.5 3.76

e.n •J.74 Fig.4 - Variation of the lattice parameters a ( ) and c(.) with x in (i.l I1.73 0.4 O.fi 0.0 Li ZrSe.. o.o 1.0 x 2 x in Li ZfSe2

References. (1) M.S.Wittinghja, Progess in Solid State Chem., ^2_, 41 (1978) (2) C.Berthier et al., Solid State Ionics, 2, 231 (1981) C.Bertheir et al., Nai:o "Davy" A.S.I., Cambridge 1983 (3) Y.ONUKI et al., Synthetic Metals, J>» 245 (1983) (4) ?.C. Klipstein, Ph.D.Thesis, Cambridge (1982) (5) D.C. Dahn et al., Solid State Comm., 44, 29 (1982) (6) W.R. McKinnon et al., J. Electrochem. Soc.to be published (7) J.R. Dahn et al., J. Electrochem. Soc, 13±, 1823 (1984) (8) J.R. Dahn et al., Can. J. Phys., 60, 307 (1982) (9) F. Hulliger in Structural Chemistry of layer-type phases,(vol.5 of Physics and Chemistry of Materials with layered structures edited by F.Levy) D.Reidel (1976).

C.LEVY-CLEMENT, C.N.R.S., Laboratoire de Physique des Solides 1 place Aristide Briand, 92195 MEUDON Cedex (France) 0 5

CURRENT TRANSPORT IN NIOBIUM TRISELENIDE BY CHARGE DENSITY WAVES K• Seeger, A. Philipp and W. Mayr Ludwig Boltzmann Institut fur Festkorperphysik, Kopernikusgasse 15, A-1060 Wien, and Institut fur Festkorperphysik der Universitat Wien, Strudlhofgasse "4, A-1090 Wien

The low-field electrical conductivity as well as thermopower and similar phenomena indicate metallic transport along the whisker shaped single crystals of niobium triselenide, NbSe-. At low temperatures, two of the three types of Nb chains in the unit cell become insulating at 143 K and 59 K, respectively, due to the formation of pinned charge density waves (CDW). The application of strong electric fields larger ti ?*n a threshold field causes depinning of the charge density waves and a restoration of the high conductivity. The order of magnitude of the threshold fields are about 1 V/an for the high-temperature CDW and about 0.1 V/cm for the low-temperature CDW. Two models for the depinning mechanism are presently under consideration: the classical overdamped oscillator model (1) and the tunneling model (2). A crucial test between the -wo models is the observation -of photon assisted tunneling at the application of a frequency which is larger than the pinning frequency of the CDW's (1). This phenomenon should already occur at field amplitudes which are below the threshold field for dc conduction if the tunneling model is valid, but not in the case of the classical model.

We have performed such an experiment at microwave frequencies and indeed observe photon - assisted tunneling at field amplitudes which are an order of magnitude below the dc threshold field (3). The technique applies phase - resolved harmonic mixing (PREHM). It suppresses the ohmic contribution to the conductivity which is due to the chains remaining metallic even at the lowest temperatures. Only the nonohmic contribution due to CDW depinning is recorded. PREHM involves mixing of the fundamental frequency v. and its second harmonic v. = 2 v at the nonlinear current voltage characteristic and the detection of the zero frequency current (v? - 2 v..), i.e. the resulting dc current. The variation of a phase $ introduced at will between the v. wave and the v^ wave varies the dc signal according to a cos(2) law, and by this variation the CDW signal is distinguished from a rectification at contacts or sample inhomogeneities. By an observation of the phase dependence relative to a reference sample of germanium at room temperature, the phase shift due to the dielectric displacement current is seen to be significant only at temperatures in the range of 20 to 30 K where the relative dielectric constant is up to ~ 10 9. 0 5

The dipole moment per unit volume in a CDW increases with applied dc electric field up to the threshold field, where depinning begins. In the depinned CDW's there is no polarization and the dielectric displacement decreases with further increasing in the field strength. It has been shown that in this range of a negative differential dielectric constant, electrical instabilities occur which have a fundamental frequency proportional to the CDW current and which are rich in harmonics. The mechanism of these instabilities involves the formation of a space charge equal to the effective charge of the depinned CDW's, and a field independent mobility. The instabilities have been observed 1979 by Fleming and Grimes in NbSe_ and, more recently, by others in TaS_, halogenized tetrachalogenides and blue bronzes.

The unusually large dielectric constant observed in the ohmic range is quantitatively calculated from the nonlinear conductivity by the application of the Kramers Kronig relations well known in optics, although the pinning frequency v is in the 107 Hz range only. The overall picture of the CDW condensation is that of a many - particle mechanism with a Peierls gap of about 10 eV, four to five orders of magnitude below the thermal energy k T of 10 - 10 eV at the condensation temperature. In particular, MbSe, seems to be a model substance for the investigation of CDW phenomena.

(1) For a review see e.g. G. Griiner,Comments Solid State Phys. _10, 183 (1983) (2) J. 3ardeen, Phys. Rev. Lett. J+2., 1^98 (1979); ibid. 4_5, 1978 (1980) (3) K. Seeger, A. Philipp and W. Mayr, Solid State Comm. _5£, 223 (198M-)

Prof. K. Seeger, Ludwig Boltzmann Institut fur Festkorperphysik Kopemikusgasse 15, A-1060 tfien 0 6

THE N7C3 fPWFe.Cr)) METAL-CARBIDES: A HIGH RESOLUTION ELECTRON MICROSCOPY STUDY OF A TWO-DIMENSIONAL BINARY ORDERING SYSTEM. J. Mahy, D. Van Dyck, J. Van Landuyt and S. Amalinckx University of Antwerp (R.U.C.A.), Groensnborgerlaan 171, 2000 Antwerp, Belgium

It was reported that the electron diffraction patterns of the M?C metal- carbides CM»[Fe.Cr)) contain planes of diffuse intensity [ 1]. These compounds, of which the detailed structure is described elsewhere [1,2] can be regarded as consisting of chains of face-sharing metal octahedra. Carbon-centered triangular prisms of metal atoms can be based on the triangular faces of the strings, with the prism axis sloping either upward or downward with respect to the string axis. On a single string however, all prisrcs have the same slope. This unambiguously discriminates between the two string types, except when the structure is projected along the string axis Cc-axis). It is therefore tempting to regard the M_C- metal-carbides as a two-dimensional binary ordering system, in which the ordering is caused by structural units rather than individual atoms [ 2] . Disorder in the stacking of the chains would then result in the appearance of planes of diffuse intensity in reciprocal space. In the [ 0001] zone however, the diffuse intensity would be absent, because the chains are identical when projected along the c-axis. It was shown in ref. 2 that the "cluster theory", which proved to be successful in the description of the "transition state" in binary systems [ 3] could also formally be applied to the PLC., metal carbides and enabled to explain the occurrence of the diffuse intensity contours in terms of a microdomain model. Observations by means of high resolution electron microscopy CHREM) confirm the proposed models, yet ambiguities remain concerning the number of types of possible defects occurring in the structure.

The specimens of metal-carbide were obtained by arc-malting of the constituent elements in nominal composition Fe.CrC- and subsequent ion-beam thinning. Upon b j tilting away from the hexagonal [0001] section, lines of diffuse intensity, arranged in a "Davidstar"-pattern in planes parallel to the basal hexagonal section were observed (fig. 1a). This pattern is similar to the ones reported by [ 1] in Cr_C_. In some areas of the studied sample, superlattice reflections could be distinguished, which are superposed on the lines of diffuse intensity Cfigs. 1b]. Dark field imaging using the diffuse intensity as well as the satellites in zones with their axis not parallel to c* revealed that the three symmetry related sets of parallel lines are associated with three sets of 0 6

parallel defects, which could be dstsrminsd as twin boundaries and antiphase boundaries with [1010] and [1120] orientation respectively. In some crystal parts, very weak lines of diffuse intensity were observed in the [0001]* section of reciprocal space Cfig. 1c3. These lines have tha [1010]* type orientationi they pass through the (hkio) reflections of the underlying hexagonal lattice. Their origin is not due to multiple diffraction affects. H.R.E.M. imaging in the [0001]* zone shows the hexagonal environment associated with the projection of the chains along their axis Cfig. 2a). No distinction can be made between the two chain types, resulting in the "average" R117B3- structure. In -fig. 2b it is seen that the defect planes can be arranged either periodical or not. In the areas contributing to the diffuse lines in the [0001]* section of reciprocal space, faint parallel defects are observed (fig. 2c).

From the analytical expression of the experimentally observed loci of the diffuse intensity, an ordering-relation was obtained [ 2j. The microdomain model derived in this way consists of ordered orthorhombic and hexagonal domains on an underlying hexagonal lattice, which are separated by defect planes satis- fying the ordering relation. A number of models was computer simulated and the calculated patterns of diffuse intensity were compared to the observed ones. Conversely, the H.R.E.M. images are also consistent with the microdomain model. However, since the projection of both types of strings is the same, no defects due to the string-disordering should be visible. The presence of the lines of diffuse intensity observed in the [0001]* zone, can thus possibly be explained either by assuming that the projection of both string-types is not identical, which is not'likely to be the case, or, more likely, that the distance between chains of the same type is different from the distance between chains of a different type.

References [ 1] J.P. Morniroli, E. Bauer-Grosse, M. Gantois : Phil. Mag. A 48, 3, 311(1983). [2] J. Many, D. Van Dyck, 3. Amelinckx : Phil. Mag. A _5£. 3, 411 (1984]. [3] R. De Ridder, G. Van Tendeloo, D. Van Dyck, S. Amelinckx : Phys. Stat. Sol. (a) 38, 663 (1976). 0 6

Fig, la : EDP tilted away over 10° from the [0001]* zone. The Qavidstar pattern of diffuse lines is marked.

Fig, "c : tiOP of the foTn]* zone with superlattire reflections indicating a trip- lir.g of me periodicity along [10T0]*, eventually superposed on lines of diffuse intensity.

* . ' snowing faint lines of t^^t^^,'''^-1^^''^'^'''''^ diffuse intensity along - • •-' —-:' —"^l+LH ..-!5!'BC-> ?¥* • .. [ 10~G]*

OOOO 0 cir. 2a: [ JGC"1] * zone axi irr^apg C.A.I.), with the evsraps ^u^B^ structure 5-cwn. ' -

Fig. 2b : [3111]* Z.r.I. showing periodical/non-periodical arrangement of [1010]* twin ooundaries.

Fig. 2c : [D0D1J* Z.A.I, in which defects associated with fig. 2a are observed.

J. Mahy, Lab. for High Voltaga Electron Microscopy university of Antwerp, RUCA, 3roenenborgerlaan 171, 3-2020 Antwerpen, Selgium. 0 7

TRANSITION METAL CARBIDE/-BORIDE PHASE BOUNDARIES, KEY FOR WEAR RESISTANT MATERIAL DEVELOPMENT H. Holleck Kernforschungszentrum Karlsruhe, Institut fur Material- und Fest- kOrperforschung, Postfach. 3640r D-75OO Karlsruhe 1

Transition metal carbides, nitrides and borides are widely used in wear resistant materials. In composites together with binder metals they give the high hardness to the material, whereas the metal phase is responsible for the toughness. This binder metal phase is however often the limiting factor for applications at higher temperatures or in corrosive environment.

Multiphase materials without binder metals but with a high amount of phase boundaries can show sometimes an extremely high wear resistance, high toughness and good application performance for turning and milling of steel.

Sintering behavior, properties and wear of such materials can be related to the constitution of the system, the microstructures and to the structures of the compounds.

As model system the section TiC-TiB2 was selected to study these relations between constitution, microstructures and properties.

Fig. 1 shows the constitution of MC-MB2 (M = Ti, Zr, Hf, V, Nb, Ta) systems (cf. /1/). Specimens in the system TiC-TiB- near the eutectic composition (^ 40 Mol.% TiB2) can be completely densi- fied in a normal sintering procedure at temperatures of 1600°C.

MiTi,Zr,Hf,VlNb,Ta I TS(MB2)

TSIMC)

Ts (Eut.)

MC » 40

MolV. MBj

Fig. 1; Section MC-MB2 in the ternary system -M-C-B (schematically, M = Ti, Zr, Hf, V, Nb, Ta) 0 7

Microstructures with regular grains,as shown in Fig. 2, with almost no grain growth are obtained. The hardness of the two phase composite is with about 2200 HV lower than that of the

pure compounds TiC (2800 HV) and TiB2 (2900 HV). The fracture toughness and transverse rupture strength of the composite however exceeds the values of the boundary phases. Wear tests in particular elucidate the predominant role of the phase boundaries. This has been proved in more complex multiphase

systems like TiC/TiB2/SiC, TiC/TiB2/Si3N4 or TiC/TiB2/Al203.

Fig. 2; Microstructure of a TiC/TiB_ composite

Only carbide-boride composites with the possiblity of semi- coherent phase boundaries (Fig. 3 shows the possiblitiy of

TiC(TiN) TiB2

in 0001 B A A

0.306oml lOJ03nm (0.300™)!

0-750 nm 0.323nm phast boundary A, B, C • close p«d»d Ti-ptan»t; BjC(N) • boraff.ewboibor nitragan B ond C atanu not Mi tt» paatr plan. Fig. 3: Structural relations and possibility of coherence at

the close packed planes of TiC, TiN and TiB2 0 7

coherence trough the close packed planes of TiC (TiN) and

TiB2 and with apparently low interfacial energies show the excellent wear resistance during turning or milling of steel. Other composites with sometimes higher hardness and toughness are not able to resist the local stress at the surface of the tool bit during operation.

/1/ H. Holleck; "Binary and Ternary Transition Metal Carbide and Nitride Systems",in german Gebr. Borntraeger Verlag Berlin, Stuttgart 1984

^. eppllkon: Fermenter, Tltro-Analyser LSLSecfroid: Gefriertrocknung Bavimad: Elnbettungs- und Färbeautomaten Lüdi: Metalikappen für die Mikrobiologie Binder. Universal- und Brut-Schranke Macherey- Buchi: Laboratoriumstechnik Nagel+Co: Filter- und Indikatorpapiere, Chromato- GFl: Laborapparate graphieerzeugnisse für DC, SC, GS und HPLC, Gltsorc Uquidhandllng.HPLC Wasseranalyse, Mikroliterspritzen Herrnle: Zentrifugen Revco: Kühl- undTiefkälteschränke H&C CSB, BS85l Meagerâte für den (für Labor und Blutbank) Umweltschutz Rosenthal: Labor- und Industrieporzellan Hüben Thermostate Sartorius: Laborwaagen llado: Rührwerke, Analysenmühlen Schott/Maine Laborgläser und -apparais Kinematics: Oispergier-Geräte Schott/Geräte:MeegerätefürpH,Oz.LF,Viskosimetrie, Köttsrrnann: Materialprüfschranke, Labormâbel Titrationssysteme Kühnen Schütlelgerät« Vemeret Laborbedarf aus Gummi, Silikon, Viton

Prospekte Beratung Angebote „*"?*54, POSTFACHC^^n« Referenzen Toto 135917 0 8

HYDROLYSIS OF CARBIDES OF TRANSITION METALS B. Haoek. P. Karen, V. Bro2ek Department of Inorganic Chemistry Prague Institute of Chemical Technology, 16b 28 Prague

The hydrolysis of carbides has been studied extensively particularly since the introduction of gas chromatography. Still, little attention has been paid to the fact that carbides of transition metals usually yield complex hydrocarbon mixtures whereas carbides of nontransition metals give rise to a single hydrocarbon, the crystal structures being often of the same type (SrCg

YC2, etc.). There are two particular questions that call for explanation: first, why the C, and C, groups in the carbide structure are transformed into mixtures of hydrocarbons and of predominantly even hydrocarbons, respectively, rather than to CH. or C2 hydrocarbons, etc, and second, why odd hydrocarbons are also present in the mixtures of even hydrocarbons, particularly for dicarbides of rare earth or actinoid elements. The first problem can be treated 1> y2) in terms of the concept of the reaction of the primarily formed CH* or C~H* radicals, the second, in terms of thermal"5' ; or radical ' '-"splitting of

the Co groups or of the defect of C, groups lying in the C5 sites of the carbide structure0' '. We suggest that the results of the hydrolysis studies have to be evaluated critically, particularly with respect to the problems of the phase purity of samples, systematic GLC errors, complete- ness of hydrolysis of the possibly multiphase sample, retention of hydrocarbons by the reaction system, reactions talcing place in the gas phase , etc. First, we concentrated on carbides containing isolated C, 8) 9 10) groups in their structures: UC,_X , Mn^ *, Mn5C2, Mn23Cg , ^2C00.05-0.1' Sc2COO.O5-O.l " **2m (M * A1' Ga' *»• Sn) ^' We found that 1) the hydrolysis can be so conducted that the H/C ratio in the gas phase correspond to the stoichiometry of the

hydrolysis reaction MCX(O ) + H20 —•• hydrogen + hydrocarbons + a+ M (OH)a> H/C * (a - 2y)/x; and 2) the gas mixture contains hydrogen and a natural sequence of saturated hydrocarbons in decreasing 0 8

concentrations, in 30me cases accompanied by a similar series of olefins in trace concentrations. So long as the primary formation of CH£*~Z'* was considered during the aemiquantitative treatmet of the problem, the following facts could not be accounted for properly: a) the appreciable

differences in the hydrogen content, e.g., between UCQ „- (H/C =

4.21, 14% H2), Mn?C3 (H/C * 4.67, 40* H2), and Mn5C2 (H/C = 5,

51% H2), where the ratios of the CH and H* radicals considered would be close to each other; e.g., 1:2.21, 1:2.57, and 1:3,respectively, for CH2*:H*; and b) the fact that whereas carbides of nontransition metals are hydrolyzed as salt3 of weak acid, carbides of transition metals probably are not, because the CHj^H" ratio of 1:2 could then be considered also for Be2C. This problem can be tackled proceeding from the hypothesis that the hydrolysis of the carbon groups as such is not different for transition and nontransition elements, the cause of the radical reaction for the former being the formation of active hydrogen H*, which takes part in the surface reaction of the carbide hydrolysis; this concerns carbides which can liberate hydrogen from water owing to their stoichiometry or structure-defect and bonding properties and thus behave as low-noble metals from the electroche- mical point of view. The hypothesis of the initiation of the radical reactions via reaction H* + CH. —*• CH, + H~ and other ra- 8\ 4 3 2 dical reactions (see ' and references therein), made it possible to account for a) the ration of the assumed H* and CH. components and the hydrogen content of the gas; b) the fact that A1.C-, and 10) BegC give pure methane; c) the fact ' that the abundance of higher hydrocarbons lowers considerably for carbides with higher H/C ratio; and d) the fact •*' that fine-grain UC, „ and UC, „ alloys •L~x *~y « yield more hydrogen and less methane than their physical mixtures. The problem of the occurrence of odd hydrocarbons in the hydrolysis products of transition metal dicarbides had to be approached in a different way. This problem is associated with the phase impurity and defects in the samples. The occurrence of CH., particularly for dicarbides of heavy rare earth elements, was a consequence of the presence of the residual carboreduction intermediate, M2OC, and possibly of the similar M(C,N,O,D) phase formed by atmospheric contamination during the preparation from the metal and carbon. The C.,, C=, and other hydrocarbons are formed 0 8

M C from phases such as i5 19' which gives 30% C-j, 55% CH., 10% C2, and higher hydrocarbons. The stability of this phase also increases with decreasing atomic radius of the rare earth element. The metal contamination from the atmosphere during the high-temperature carbide synthesis leads sometimes to the formation of a phase impurity also giving C,H* on hydrolysis; so far, we failed to identify this phase. For dicarbides of actinoids, the occurrence of methane in their hydrolysis products is ascribed primarily to the defect of the presence of C^ groups . It can be concluded that the relation between the structure of transition metal carbides and the composition of their gaseous hydrolysis products realizes in two forms, viz. 1. in a stoichiometrically consistent relation - a quantitative relation between the stoichiometry of the hydrolysis reaction, or the carbide stoichiometry, and the H/C elemental ratio in the gas mixture; and 2. in a structurally consistent relation - e qualitative and quantitative relation between the gas phase composition and the type and number of carbon groups in the carbide structure unit. This relation is given by the fact that C^} C2, and C, groups in the carbide give a natural, an even, and a C, , respectively, sequence of hydrocarbons of decreasing concentrations within each of them.

(1) J.G.Palenik, J.C.Warf: Inorg. Chem. 1, 345 (1962) (2) H.J.Svec, J.Capellen, F.E.Saalfield:~J.Inorg. Nucl. Chem. 26, 721 (1964) (3) F7H.Pollard, C.Nickless, S.Evered: J. Chromatogr. 15, 211 (1964) (4) N.J.Clark, R.Mountford, I.J.McColm: J. Inorg. Nucl. Chem. 2±, 2729 (1972) (5) N.N.Greenwood, A.J.Osborn: J. Chem. Soc. 1961, 1775 (6) J.S.Anderson, N.J.Clark, I.J.McColm: J. Inorg. Nucl. Chem. 20, 105 (1968) (7) M.Popl, V.Brozek, B.Hajek: Collect. Czech. Chem. Commun. 35, 3529 (1970) (8) B.Hajek, P.Karen, V.Brozek: Collect. Czech. Chem. Commun. 49, 793 (1984) (9) VTBro2ek, B.Hajek, P.Karen, M.Matucha, L.2ilka: J. Radional. Chem. 80, 165 (1983) (10) P.Karen, B.Hajek: Proceedings of this Conference (11) P.Karen, R.Krai, V.3ro2ek, B.Hajek: Proc 39th Conf. Czech. Chem. Soc., Olomouc, August 30th - September 2nd, 1983, p.40 (12) P. Karen, R.Kril, B.Hajek, J.Kubdt: Proc 40th Conf. Czech. Chem. Soc.,Banska Stiavnica July 2nd - 6th, 1984, p. 46 (13) M.J.Bradley, L.M.Ferris: J.Inorg.Nucl.Chem. 2J, 1557 (1965) 0 9 ARE FIRST AND SECOND NEIGHBOUR PAIRWISE INTERACTIONS RELEVANT TO ANALYSE THE STABILITY OF ORDERED SDBSTOICHIOMETRIC CARBIDES AND NITRIDES?

J.P. LANDESMAN , TREGLIA2, P. TDRCHI3'4 and F. DUCASTELLE3 (1) S.E.S.I., C.E.N., B.P. 6, 92260 Fontenayaux-Roses, France' (2) Laborato IT e de Physique des Solides, Bat.510, 91405 Orsay, France (3) O.K.E.R.A., 92320 Cnatillon, France (4) Laboratoire de Dynamique du Riseau et Ultra-sons, Tour 22, 4 place Jussieu, 75005 Paris, France

We present here a study of the ordering processes in substoichiome- trie transition metal carbides and nitrides which crystallise with the MaCl structure ( the transition metal atoms occupying one FCC sublattice whereas the carbon or nitrogen atoms occupy the other, referred to as "interstitial"). In most of these compounds, vacancies appear on the interstitial FCC lattice only. Under some conditions ( annealing temperature and duration ) these vacancies may order(l). The long range ordered phases observed in this case ( on the interstitial sublattice ) are analogous to those predicted, for FCC substitutional alloys, by theoretical calculations using an Ising model limited to first and second neighbour pair interactions, V. and V_ ( see figure 1 and ref.2 ).

Fig.l-Stable structures on the FCC lattice with first and second neighbour interactions.The structures A_B, A-B and A_B. are described in ref.2.

Segregation

Note that the ordered structures observed in carbides (nitrides) are found in the region VJ> 0, ( 7j> °» vi>2V ? We show that one can calculate from the electronic structure of the disordered state such pairwise interactions by extending to the interstitial case the generalised perturbation method previously developed in the case of substitutional transition metal alloys (3). Actually, although the total energy of such compounds cannot be written in terms of pair interactions, it can be shown that the configurational part of this energy can be written in such a way, i.e. as a sum of effective pair interactions. The idea is that the energy of any ordered compound can be obtained by a perturbation expansion from the disordered medium ( the electronic structure of which is derived from the band structure calculations of Schwarz(4) and treated within the C.P.A.(5) ): OS DS. 0 9 OS DS where E (E ) is the band energy of the ordered (disordered) state characterized by the number of metalloid-metalloid pairs, q ( q ) at a given distance R. V is the corresponding pairwise interaction, which only depends on concentration and band filling ( not on configuration ) through the electronic structure of the disordered state ( the complete expression can be found in ref.6 ). These pairwise interactions are shown in figure 2 for hemi-carbides and nitrides and for first and second neighbours, as a function of the band filling N.

Fig.2- First anc1 second neighbour pair interactions (V-i'V?) for hemi-nitrides (a) and carbides (b) as a function of the band filling N.

It can be shown that the fillings characteristic of carbides (nitrides) are such as .31? N s.45 ( .42$ N $.58 ). Therefore, carbides and nitrides correspond to the following hierarchy of pair interactions: V > 0, (carbides) 2 v2» V (nitrides), o. vh, >V22 in fair agreement with experiments and figure 1.

This hierarchy of pair interactions can be understood from simple moment arguments based on the linearized Green function method (7).The first non vanishing contributions to V^ and V. are due to the fourth moment v, of the density of states, multiplied by a universal function S"2(N) which presents two zeros as a function of the band filling (O

Fig.3- Various contributions to the first and second neighbour pair interactions. The different paths on the interstitial sublattice are numbered on the upper side of the figure. The metalloid (metal) atoms are represented by open (full) circles.

REFERENCES:

(1) De Novion C.H. and Maurice V., Journal de Physique, Colloque 3_8_ C7 (1977) 211 De Novion C.H. and Landesman J.P., this conference (2) Allen S.M. and Cahn J.W., Acta Met. 2£ (1972) 423 Kanamori J. and Kakehashi Y., Journal de Physique, Colloque 38_ C7 (1977) 274 (3) Ducastelle F. and Gautier F., J. Phys . F £ (1976) 2039 Ducastelle F. and Treglia G., J. Phys . F 10 (1980) 2137 Bieber A., Gautier F., Treglia G. and Ducastelle F., Solid State Comm. 39 (1981) 149 Bieber A., Ducastelle F., Gaucier F , Trlglia G. and Turchi P., Solid State Coiam. 45_ (1983) 585 (4) Schwarz K., J. Phys. C 10 (1977) 195 (5) Faulkner J.S., Phys. Rev. B 13 (1976) ?391 Klima J., J. Phys C 12 (1979) 3691 (6) Landedtnan J.P., Turchi P., Ducastelle F. and Treglia G.,"Phase Transfor- mations in Solids", Mat. Res. Soc. Symp. Proc. 21^ ed. T. Tsakalakos, (North Holland, N.Y., Amsterdam, Oxford) 1984, p.363 Landestnan J.P., Treglia G., Turchi P. and Ducastelle F., Journal de Physique, to be published (7) Turchi P. and Ducastelle F., Proceedings of "The Recursion Method and its Applications" Conference (London, Sept. 13-14,1984), Springer- Verlag Series 0 10

Effects of disorder on the superconducting temperature of cubic MoN D.A. Papaconstantopoulos and W.E. Pickett Naval Research Laboratory Washington, DC 20375-5000, USA

Abstract

calculations of the electron-phonon interaction, n, in ordered, fully stoichiometric Bl-structure MoN predict a superconducting transition temperature, T , of 29VK. In this paper we discuss the effects of disorder of both n and T on the basis of calculations using a) the electron-lifetime model (ELM), (b) a tight-binding form of the coherent-potential approximation (CPA) to account for vacancies of the nitrogen sites and c) the rigid-band model to describe the

Nbi-yMoyNi.o system- The ELM calculations involve a convolution of the density of states (DOS) of the stoichiometric MoN using a broadening function which is a Lorentzian whose half-width is related to the infrared plasma energy and the residual resistivity arising from disorder. The results indicate that this form of disorder will lower n and Tc from the calculated value for stoichiometric MoN. However, up to strong levels of disorder it gives a T higher than that of the high-temperature superconductor

KbN. 0 10

The second type of calculation we performed

utilizes the CPA and has as a starting point a Slater-

Koster fit to the MoN energy bands. In our CPA model we

have assumed the the metal sublattice remains unaffected

but the nitrogen sublattice contains random substi-

tutions of nitrogens by vacancies. We performed these

MoN CPA calculations for 1.0 < x ± 0.5 and found very

small changes of the total DOS, N(Ep), at the respective

Fermi levels. However, the Mo t» DOS and particularly

the N p-like DOS decrease substantially with decreasing

x. Based on this observation and information known from

experiment for NbC and NbN regarding an apparent

stiffening of the phonon moments with decreasing x, we have constructed a theory for the calculation of the electron-phonon coupling constant X and T . Our calculations show that T drops from 2S.2"K at x = 1.0 to 4.3"K at x = 0.5. Our result for x = 0.5 is in good agreement with recent measurements by Wolf et aJT for cubic Mo_N. However, when nitrogen implantation was applied to increase the nitrogen content, a 50% increase of T was achieved still far below the theoretical prediction. This is probably due to damage of the lattice due to the ion implantation.

We have also Investigated the variation of n» X., and T in the system Nb, Mo N, Q using our band structure results for NbV and MoN, within the rigid-band 0 10

model. Our results show a decrease of T as we start c from the NbN end (Tc * 17"K) to a value of 6"K at y =

0.3. This is followed by a steady increase of Tc with increasing yf which reaches 29.4"R at y = 1.0.

A more elaborate study of the Nb, ,Mo,,N, n system i-y y i.u using the CPA is now is progress.

References W.E. Pickett, B.M. Klein, and D,A= Papacon- stantopoulos, Physica, 107, 667 (1981).

D.A. Papaconstantopoulos, W.E. Pickett, B.M. Klein, and L.L. Boyer, Nature, 308, 494 (1984).

S.A. Wolf, S.B. Qadri, K.E. Kihlstrom, R.M. Simon, W.W. Fuller, D. VanVechten, E.F. Skelton, and D.U. Gubser, Proc. Applied Conf., IEEE Trans, on Mag. (1984). 0 11

EFFECT OF d-BAND MIXING IN Mn-TRANSITION METALS SPIN-GLASSES

J. J. Hauser and J. V. Waszczak

AT&T Bell Laboratories, Murray Hill, New Jersey 07974, USA

We recently reported (1) the magnetic properties of amorphous (a) Mn-Zr alloys which exhibited two remarkable and unexpected results. First, while Ni, Co and Fe display no local magnetic moment (2) when alloyed with Zr in metallic glasses, Mn has a local moment in a-Mn-Zr. Second, the existence of the Mn local moment does not result in a spin-glass interaction. These results can be understood by the study of the magnetic properties of a-Cu-Zr-Mn alloys presented here which shows that Che absence of spin-glass interaction in a-Mn-Zr is not due to the amorphous state but to a mixing of the Zr and Mr. d bands. This model has led to the discovery of a new spin-glass alloy: W-Mn. All the films of the present study were sputtered from master alloy cathodes onto sapphire substrates held at 260K. The films were then scraped with a sapphire slide in order to avoid magnetic contamination. The susceptibility of the resulting flakes was measured in a high dc magnetic field using tne Faraday method, in a low dc field (3 Oe) in a susceptometer using a super- aa conducting quantum interference 24 a m ° device, and a low ac field (4 Oe) 22 a ° 0 at 10 kHz. Fig. 1 displays the 201- ca. o well known susceptibility cusps ^•13% Zr a o characteristic of a spin-glass #n a _ a interaction for a film of the 18 a a archetypical Cu-Mn spin-glass (3) 14 " and for a-Cu-Mn-Zr films. Since as i " iliown in Fig. 1 the ternary amor- / 0%z "6 Stm " phous Cu-Mn-Zr alloys display a r - spin-glass interaction, it is a» 33% Ts clear that the absence of spin- al I«>I_ • • • 111 •— ~»_ «O •, •

° n • •,*.„_ glass interaction in a-Mn-Zr cannot be linked to the amorphous state. Furthermore, it' is also clear from Z\- • Fig. 1 that the spin-glass cusp

. 0,...,.. 2O3O4090«O70«*• >i has almost vanished when 20 at.% Zr TOO is substituted to Cu in Cu 9MnQ ... Fig. 1 ac susceptibility as a function This is further discussed in Fig. 2 of substituted Zr. 0 11

where one notices that with increasing Zr, the spin state (S) decreases smoothly from SPUTTEREO CATHOOES that of CV9Mnoa to »S that reported for

Mno92roa (1) while the spin-glass freezing temperature T__ which SG corresponds to the maxima in the susceptibility curves of Pig. 1 remains constant 16 24 32 40 <8 56 64 72 80 88 96 for amorphous films up

Fig. 2 Dependence of T and S on substituted Zr to the disappearance of for both films and cathodes. the spin-glass transition which occurs for Zr concentrations larger than 20 at.%. The constancy of T reflects the fixed SG Mi. concentration and the decrease in S can be understood in terms of the 6.0 decrease in the Mn moment by the mixing of Mn and Zr d-bands. One should •_ 3.5 at. % Mn also notice in Fig. 2 the opposite S.C o H^J.IOO oe behavior of the sputtering cathodes: 4.5 OS in this case T__ increases while S remains essentially constant with increasing Zr because Zr is insoluble ^.S.5 in both Mn and Cu and consequently, •o •s in these phase separated cathodes increasing Zr implies an increasing " 2.5 Mn to Cu ratio which leads (3) to an 2.0 increase in T SG' This result stress- IS es the necessity of obtaining single

10- phase alloys by film deposition. This point is further demonstrated in OS Fig. 2 by the values of S and T _ for SG 40 TC« the 2.5 at.% Zr crystalline film Fig. 3 ac susceptibility as a function which fall in between those for of applied dc magnetic field. crystalline CuQ 9MnQ 1 and those for 0 11

the amorphous alloy films. The absence of a spin-glass state which has also been observed in Ti-Mn films is again attributed to a mixing of Mn d states with the partially filled Ti d band. This mixing effect which can obviously not take place in Cu-Mn alloys because of the full Cu d band may not exist in certain other Mn-transition metal alloys such as W-Mn for a different reason. Although W has also a partially filled d band the density of d states is a minimum at the Fermi level and the W d states may therefore not perturb the Mn d band. This is demonstrated by the fact that the value of S in W-Mn alloys is very similar to that reported for Cu-Mn alloys (3). Furthermore, it is clear from Fig. 3 that a spin-glass transition occurs in dilute W-Mn alloys as shown by the susceptibility peak at T y. 15K and the rounding-off of the peak by the application of a dc magnetic field H. This new spin-glass transition was studied as a function of the Mn content.

(1) J. J. Hauser and J. V. Waszczak, Phys. Rev. B30_, 2898 (1984). (2) Z. Altounian and J. 0. Strom-Olsen, Phys. Rev. B27_, 4149 (1983). (3) V. Canella and J. A. Mydosh, Phys. Rev. B6, 4220 (1972).

J. J. Hauser, ATST 3ell Laboratories, Murray Hill, New Jersey 07974, USA 0 12

CrTe is an example of such phamomenon. Makovetskii and Shaklevich (17) proposed

a (CrTe) » 2.45uR for infinite field, but from their curve the extrapolated value for HsO is 2.40 u „- This values is due to the contribution of the two electron 3d3/. with u » 1.5 and a * 1.2n_. The two electrons with u«0.5, a »0,4u. are quenched. From the work (5) we have shown that in this compound there are two cation cation bonds. We can see, these two bonds are responsible for the quenching. Now we want to discuss the case of Rb-jCrCL. From this work the highest possible

value for the chromium ion is 3.2yR. In 1975 Anthony et al (7) published a value

o = 3.5uB at H x 35 k Oe. Their extrapolated value at H=0 is 3.3Oy_. Now for neutron diffraction studies large single crystals are required. With a new method Garton and Walker (21) got very good results. They underline that going from powder to single crystal the purity is improved. This point is important because in neutron diffraction study Fair et al (8) found in 1977 a =3-9MB on powder but3.3uB on single crystal. This value has recently been confirmed by polarised neutron diffraction. Mumminghoff et al (9) found in 1982 a = S^n.^ Similar results have also been found on CsCrCU in 1980 by Zandbergen and Ijdo with a = 3.16 uB by neutron diffraction study (7).

As from the classical spin only value we must obtain 4.00uB for Cr.., we think that the value of 3.2y~ as a maximum found for chromiun is a very good test to check this model. At this time it seems quite well verified.

2.4 2.8 3.2 uB | | | Ref.

Rb2CrCl4 |0 ||o (7,8,9) I. do)

_>e4 II' (11,12) CdCr2Se4 || (13,14)

HgCr2Se4 | (14)

CrTe || | (15,16,17)

RbCrl3 |, (18) CrN |J. (19,20)

Figure 1 : Chronium magnetic moments in some compounds. Experimental values from magnetization | and from neutron diffraction studies |0 . The number on the scale are the theoritical values. 0 12

(1) Dirac P.A.M., Proc. Roy. Soc. A 117, 610-624, (1928). (2) Oudet X. J. Phys. J3, L205-8, (1980). (3) Oudet X. Proceeding of the annual french meeting on the "rare earths and actimides with anormal valency". 7 and 8 november 1984 C.R.T.B.T., Grenoble. (4) Dirac P.A.M., Proc. Ray. Soc. A 118, 351-61, (1928). (5) Oudet X. Ann. Chtmie JJ, 483-507, 1983. In French, english version avaible from the author. ~ (6) Oudet X. J. Magn. Mat. (1984). (7) Anthony K. et al. J. Chem. Soc, Dalton Trans., 1306-1311, '1975). (8) Fair M.J., et al, Physica 86-88 B, 657-9, (1977). (9) Munninghoff G. et al. J. Physique, 43-C7-243-7, (1982). (10) Zandbergen H.W., Ijdo D.J.W., J. Solid State chem., 34, 65-70, (1980). (11) Plumier R these Paris 17-12-1968 p. 134. (12) Nogues M., our laboratory private communication, o = 2.84ug on single crystal. (13) Menyuk, K., et al., J. Appl. Phys. 37, 1387-8, (1966). (14) Baltzer P.K., et al., Phys. Rev. L5J.. 367-377, (1966). (15) Guillaud C, Berbezat S., C.R. Acad. Sci., Paris 222, 386-8, (1946). (16) Lotgering F.K., Gorber E.W., J. Phys. Chem. SofidTjJ, 238-49, 1957. (17) Makovetskii G.I., Shakhlevich G.M. kristall und Tecknik 14, 97-105, (1979). (18) Zandbergen H.W., Ijdo D,J=W,. S, Solid State Chem.. 38. 199-210. (1981). (19) Nasr-Eddine M., Bertaut E.F., Sd. Stat. Comm. 24, 4"57-492, (1977). (20) Nasr-Eddine M., Roubin M., Paris J., Sol. Stat. Comm. 32, 953-4, (1979). (21) Garron G., Walker P.J., J. Crys. Growth 33, 61-4, (197617

Xavier OUDET, Laboratoire de Magne'tisme, CNRS, 1, Place A. Briand 92195 Meudon, Cedex, France O 13

MAGNETIC ORDERING IN INTE8METALLICS OF THE ThCr2Si2 TYPE J. Leciejewicz Institute of Nuclear Chemistry and Technology, 03-195 Warszawa A. Szytula Institute of Physics, Jagelloaian University, 30-059 Krakow, Reytnonta A, Poland

In this paper, we shall concentrate on results obtained by a neutron diffraction study of a few RKT_X- intermetallic compounds. The ternary intermetallics RET-X- in which RE is a rare earth metal, T a 3d, 4d or 5d transition clement and X is Si or Ge, have been extensively studied during the past years. They crystallize in a simple structure of ThCr-Si- type which is tetragonal with space i;roup IA/immn (I)- The X-ray and neutron diffraction data indicate that: RE atoms arc in the position 2a): 0,0,0 T atoms take- 4d): 0,j,{, {,0,\, and X atoms prefer 4e): 0,0,z, 0,0,z Thv. structure can be described as a stacking of atomic planes perpoTidieular to the c-axis with the sequence of Rc—X-T-X-Rli. The value of the 7. parameter is abouL 0.375. These compounds allow to study many interesting physical properties such as superconductivity, valence fluctuations, the Kondo effect and the magnetic interactions (2). Magnetometrie and neutron diffraction measurements indicate that in most compounds (except those with Mn) the T component has no magnetic moment. On the other hand the rare earth moments usually order antifcrro- m.-ij;nctical ly or Cerroraas;netical ly at low temperatures. The temperature dependence of the magnetization and reciprocal magnetic susceptibility of REMn2X_ compounds (3) indicate two critical temperatures of magnetic ordering! - at low temperatures (UIT) the magnetic moments localized on Mn and i{£ atoms order, - .it hij-h temperatures (LNT and RT) the magnetic moments on Mn atoms order. Neutron diffraction measurements indicate that within the REMn.X- family, a col linear antiferromagnetic structure is observed for REMn-Si-

(RE- Cc,Pr,Nd), ErMn2Si2, ErMn.Ge-, and TbMn-Gej- Their magnetic structure 0 13

consists of a sequence of ferromagnetic layers of Mn atoms, stacked along the c direction. The magnetic moment of the Mn atom ( 2.5n_ at 4.2 K) is parallel to the c-axia. REMn.Ge, (RE- La-Sm) and LaMtuSi- are ferromagnetic (4). The magnitude of the magnetic moment localized on the Mn atom results from of an overlap of the electronic shells of the Ma and Si or Ge atoms, leading to spin transfer from the p shell of Si or Ge to the 3d shell of Mn. In the case of other 'RET-X- compounds similar collinear magnetic structures within the RE sublattices are observed:

F - ferromagnetic ordering is encountered in ErMn-X^ (X* Si,Ge)s

AF I - magnetic moments lying in the (001) planes are coupled ferromagnetically. Adjacent planes are coupled antiferromagnetically so that their sequence in the direction of the c-axis is +-+-. This type of magnetic ordering is obseri?ed for EEC^Si- and RECo-Ge- (HE= Pr-Tm),

PrCu^, RERh2Si2 (RE- Nd-Er) and TbIr2Si2. AF II- the magnetic structure consists also of ferromagnetic layers perpendicular to the c-axis coupled antiferromagnetically with

sequence +H—. This is observed in NdFe2X2, PrFe2Ge2 and ErFe2Si2, AF III-antiferromagnetic ordering within (001) planes with a +-+- stacking sequence along a fl 11 } direction. The magnetic cell aY2*, aPu] c.

This type is observed in TbNi2Si2 and CeRh-S^ (5), AF IV- the magnetic structure consists of ferromagnetic (101) planes of RE coupled antiferromagnetically. This type is observed in (Tb,Dy,Ho)

Cu2Si2 and (Tb,Ho)Cu2Ge2 (6). Also non-collinear magnetic structures are observed: - in the case of PrCo^Ge,, a cosinusoidally modulated longitudinal spin wave (LSW) propagating along the c-axis (the propagation vector is 0.73 c ),

- in the case RE&juSi- (RE- Tb,Dy,Ho,Er) and REOs2Si2 (RE- Tb,Ho,Er) a consinusoidally modulated longitudinal spin wave propagating along the b-axis (the propagation vector are 0.2-0.233 b* for RERu-Si- and 0.3 b*

for REOs2Si2). Investigations of the magnetic ordering in RET-^ compounds supply interesting information about the magnetic anisotropy and the exchange interactions in ternary compounds. In all compounds for RE» Pr-Ho and T» Mn-Ni, the magnetic moments are aligned along the tetragonal unique C-JUCLS. In the case RE* Er and Tm 0 13

they are normal to it. A different situation is observed among the RECu-Si- compounds. The magnetic moment of the Fr atoms is parallel to the c-axis, while that of Tb, Dy and Ho are normal to it. Orientation of the magnetic moments in the unit cell is connected with the sign of the B. coefficients. The CEF hamiltonian for a rare earth ion with point symmetry 4/nmn is ha • »°°? * "M • « * 'K * &,

where the operators 0 and the coefficients B are as defined by Hutchings (7). According to Greedan and Rao (8) the positive value of the B_ coefficient indicates that the magnetic moment lies in the basal plane or makes an angle 0 with the c-axis. The estimated values of the B_ coefficients for REFe-Si. compounds by means of the Mossbauer effect are in good agreement with the determined orientation of the magnetic moments (9).

The majority of RET,X2 compounds can be described as stacking of ferromagnetic (001) planes along the c-axis. RE ions which compose the (001) planes are coupled ferromagnetically at distance of about 0.4 nm. The nearest distance between two rare earth ions belonging to neighbouring planes is more than 0.5 nm, so that the magnetic interactions are apparently weaker. The distribution of rare earth ions in the crystal structure of ThCr^Si- indicates the anisotropic character of magnetic interactions between RE ions. The oscillatory character of the observed in RERu-Si- and REOs-Si- (RE» Tb-Er) magnetic ordering schemes suggests that also magnetic interaction via conduction electrons, postulated by the theory of RKKY should be called for to explain the stability of the above mentioned structure. The above conclusions suggest that within the RET.Xj family, the exchange interactions may take one of the following ways: - superexchange RE-X-RE, - exchange via the conduction electrons. The actual electronic structure will decide, which of them will be dominant in a particular compound. Superexchange interactions lead to collinear magnetic structures. If exchange interactions of RKK7 type are dominant, modulated structures appear to be Ptable. 0 13

(1) W.Rieger and E.Parthé, Mh.Chem. _100, 444 (1969) (2) F.Sceglicb, J.Aarts, C.B.Bredl, W.Lieke, D.Meschede, W.Franz and S.Schäfer, Phys.Rev.Lett. 43, 1892 (1979) (3) A.Szytula and I.Szott, Solid Stats Conmtun. ^0, 199 (1981) (4) K.ScV.L.Narasüsban, 7.O.S.Rao, W.E.Wallace and I.Fop, AIP Coof.Proc. 29, 594 (1975) (5) S.Quezel, J.Rossat-Mignod, B.Chevalier, P.Lejay and J.Etourneau, Solid State Commun. ^9_, 635 (1984) (6) P.Schobinger-Papamantellos, A.Niggli, P.A.Kotsaoides and J.K-Yakinthos (private communication) (7) N.T.Hutchings, Solid State Phys. \6_, 227 (1964) (8) J.E.Greedan and 7.S.U.Rao, J.Solid State Chem. 6_, 387, £, 368 (1973) (9) D.R.Noakes, A.M.Umarji, and G.K.Shenoy, J.Magn.Magn.Mat.39_, 309 (1983)

A. Szytula Institute of Physics, Jagellonian university, 30-059 Krakow, Reyaxmta 4, Poland 0 14

MAGNETIC AND ELECTRONIC PROPERTIES OF THE Eu, La S SOLID SOLUTION FROM 151Eu 1—X X MOSSBAPER SPECTROSCOPT J.P. Sanchez, J.M. Friedt Centre de Recherches Nucleaires, 67037 Strasbourg Cedex, France

K. Westerholt, H. Bach Ruhr-Universitat, Bocb-an, Experimentalphysik IV, D 4630 Bochum, W- Germany

The solid solution between the ferromagnetic semi-conductor EuS and the non-magnetic metallic LaS exhibits a magnetic phase diagram which is typical for frustrated systems : with decreasing concentration of magnetic atoms (1 - x). one observes ferromagnetic (FM), re-entrant spin glass (for decreasing T) and spin glass (SG) transitions (Fig. 1). An insulator to metal transition sets in at low x, giving rise to a rapid increase of the ferromagnetic T (x < 0.10) due to a change of magnetic interactions (from superexchange to PJQT interactions). The occurrence of a spin-glass phase corresponds to the oscillating character of the RKKY interactions.

Eu Mossbauer experiments are reported as a function of temperature (1.4 - 77 K) and of applied magnetic field (0 - 80 kOe) . The C. -,-ie temperatures. T , established from Mossbauer spectroscopy agree with those evaluated from ac - susceptibility whereas the Mossbauer spin glass freezing temperatures, T,, occur at higher temperatures ; this indicates a weak dependence of T. on the —8 —2 characteristic measuring time (~ 10 s for Mossbauer vs ~10 s for ac - x) Broad Mossbauer spectra are observed over a wide temperature range below T , T down to ~0.3xT , T . The temperature dependence of the spectralshape is assigned to relaxation effects rather than to a distribution of static hyperfine field (notice that at saturation the data are well represented by an unique set of hyperfine parameters). The relaxation is uniaxial and spherical below T and T, respectively (Fig. 2). The temperature dependence of the relaxation rate follows Arrhenius laws with activation energies, E , comparable 8 ^ to T , T, and a frequency prefactor of ~ 6 x 10 Hz. Relaxation persists in moderate applied fields of a few kOe.

The whole of the results suggests magnetic transitions implying the freezing of magnetization of interacting clusters over a wide temperature range below T , T..

The concentration dependence of the 151Eu isomer shift, & reveals a decrease of 16^1 (i.e. an increasing electron density at the nucleus) with 0 14 the La content ; this is attributed to the lattice pressure effect alone (compression of the Eu 6s shell) (Fig . 3). The change of electronic structure is of negligible effect. This is consistent with th« essentially d-character of the conduction band.

The concentration dependence of the Eu saturation hyperfine field, H, , , is determined by the competing effects of dilution by the non-magnetic La, change of electronic stricture, pressure effect and change of magnetic structure (however Che magnetic structure can be maintained "ferromagnetic" along the series by application of a large external field). The difference between the zero-field and the in-field hyperfine fields reflects directly the influence of changing magnetic structure. The lower value of |S,f | in zero-field (Fig. 4) clearly indicates the frustration effects which induce increasing misalignement of the spins for increasing x, even in the uiluted ferromagnetic regime. 0 14

MAGNETIC AND ELECTRONIC PROPERTIEJ OF THE Eu, La S SOLID SOLUTION FROM 151Eu 1—X X MOSSBAUER SP2CTROSCOPY J.P. Sanchez, J.M. Friedt Centre de Recherches NuclSaires, 67037 Strasbourg Cedex, France

R. Westerholt, H. Bach Ruhr-Universitat, Bochum, Experimentalphysik IV, D 4630 Bochum, W- Germany

The solid solution between the ferromagnetic semi-conductor EuS and the non-magnetic metallic LaS exhibits a magnetic phase diagram which is typical for frustrated systems : with decreasing concentration of magnetic atoms (1 - x) one observes ferromagnetic (FM), re-entrant spin glass (for decreasing T) and spin glass (SG) transitions (Fig. 1). An insulator to metal transition sets in at low x, giving rise to a rapid increase of the ferromagnetic T (x < 0.10) due to a change of magnetic inteiactions (from superexchange to RKKY interactions). The occurrence of a spin-glass phase corresponds to the oscillating character of the RKKY interactions.

Eu Mossbauer experiments are reported as a function of temperature (1.4 - 77 K) and of applied magnetic field (0 - 80 kOe). The Curie temperatures, T , established from Mossbauer spectroscopy agree with those evaluated from ac - susceptibility whereas the Mossbauer spin glass freezing temperatures, T,, occur at higher temperatures ; this indicates a weak dependence of T on the —8 —2 characteristic measuring time (~ 10 s for Mossbauer vs ~10 s for ac - x) Broad Mossbauer spectra are observed over a wide temperature range below T , T down to ~0.3xT , T . The temperature dependence of the spectralshape is assigned to relaxation effects rather than to a distribution of static hyperfine field (notice that at saturation the data are well represented by an unique set of hyperfine parameters). The relaxation is uniaxial and spherical below T and T, respectively (Fig. 2). The temperature dependence of the relaxation rate follows Arrhenius laws with activation energies, E , comparable 8 ^ to T , T, and a frequency prefactor of ~ 6 x 10 Hz. Relaxation persists in moderate applied fields of a few kOe.

The whole of the results suggests magnetic transitions implying the freezing of magnetization of interacting clusters over a wide temperature range below T , T..

151 The concentration dependence of the Eu isomer shift, 6lg, reveals

a decrease of |6ISI (i.e. an increasing electron density at the nucleus) with 0 14 the La content ; this is attributed to the lattice pressure effect alone (compression o'f the Eu 6s shell) (Fig. 3). The change of electronic structure Is of negligible effect. This is consistent with the essentially d-character of the conduction band.

The concentration dependence of the Eu saturation hyperfine field, H, . , is determined by the competing effects of dilution by the non-magnetic La, change of electronic structure, pressure effect and change of magnetic strue cure (however the magnetic structure can be maintained "ferromagnetic" along the series by application of a large external field). The difference between the zero-field and the in-field hyperfine fields reflects directly the influence of changing magnetic structure. The lower value of |H, | in zero-field (Fig. 4) clearly indicates the frustration effects which induce increasing misalignement of the spins for increasing x, even in the diluted ferromagnetic regime. TEMPERATURE (K) p.BSOBPTION ( *> c» o o W O i—i i—i—j- 1 ""H 1* 1 1—I—r—^7*"

• p T S 4 M ID \ 2

c 0. 4 m o o g" \ - 0) (A V ID o / o. at / p> 04 . 7 o b> I o Ml

ISOMER SHIFT (mm/*) SATURATION HYPERFINE FIELD (KOe)

0Q w O u w uo o o o o o 1 °7 Ml fl ID 0 O 3 i §3 O'I—V » - H- ID ,' I 01 0 to sr 0 rt 1 \ / l ID I ID P> •1 rt o a' V> O B* 0 r, ^,._, I-1- J > m MI a O 0) -•*—i r ID o' / » » a> ID a. ? » r ' -• 01 O. o / N • • i •--1 n s <" 3 C n Ml O CO I — — I ii rt c ID p. O i * ID H0 - M - Vf ID nID rt 0 15

ON THE PHYSICAL PROPERTIES OF SEVERAL MjMojXg COMPOUNDS (M = group IA, group IIIA metal; X = Se, Te)

J. M. Tarascon Bell Communications Research Inc., Murray HOI, New Jersey 07974 F. J. DiSalvo and J. V. Waszczak AT&T Bell Laboratories, Murray Hill, New Jersey 07974

Recently, a new family of ID compounds has been reported by Potel ex a/..(i> Their structure consists of infinite chains of (Mo3X3)i separated by the ternary element. The presence of these chains leads to a strongly anisotropic character of. the physical properties. The transport properties of the group IHA metal ternary molybdenum chalcogenides have been studied by Armici et al.m and Mori a o/.(3) when M = Tl and In respectively. Both compounds have been found metallic, whereas band structure calculations on these linear chains'4' suggest that some of these materials could undergo a "Peirls distortion" leading to a semiconducting ground state. However these calculations did not distinguish between the physical properties of group IA or group ULA ternary molybdenum chalcogenides, since these cations were assumed to make no contribution to the wavefunctions near the fermi level. Here we show how the physical properties of the M2Mo6X6 compounds change in going from group EUA to group IA metals. Transport and magnetic properties collected both on single crystals and on powder samples for several members of the M2Mo6X6 family with M being an alkali metal (Li, Na, • • • Cs) and with M = Tl for comparison will be reported.

Needle-like single crystals of Cs:MotTet, Rb2Mo6Te6 or Rb2Mo6Se, (figure 1) were grown by heating to 1100°C the stoicb ometric powder (synthesized by a low temperature ion exchange reaction3) in an evaculated silica or molybdenum tube with a temperature gradien: of about 100°C. The electrical properties of the alkali metal compounds measured along the needle axis exhibit a broad change from metallic to semiconducting behavior as the temperature is lowered (Figure 2). No anomalies iu the resistivity or its derivatives are seen that would indicate a phase transition temperature. On the other hand the compounds with group IHA metals (M = Ga, In, Tl) are metallic down to IK except for Tl2Mo6Se6 which becomes superconducting at 5K. The susceptibility of these materials is weakly diamagnctic and does not reveal anomalies correlated to the onset of the broad metal-nonmetal

"transition" except perhaps for Cs2MotTef. Such behavior may be due to differences in the interchain coupling (e.g. differences in die interchain overlap of the wave functions at the fermi lev?!) between group IA and group JH/» M^Mo,^ compounds. In weakly coupled 0 15

compounds (M *• group IA) a broad "transition" due to a fluctuating Peirls gap that is uncorrelatcd from one chain to the next could lead to this behavior. A second possibility is that impurities or defects sufficiently broaden the three dimensiona] ordering transition that it can not be detected.

(1) M. Potel, R. Chevrel, M. Sergent, J. C. Armici, M. Decrous, and O. Fischsr, J. Solid State Chem. 35, 286 (1980). (2) J. C. Aimici, M. Decroux, O. Fischer, M. Potel, R. Chevrel and M. Sergent, Solid State Commun. 33, 607 (1980). (3) T. Mori, V. Yokogawa, A. Kobayashi, Y. Sasaki and H. Kobayashi, Solid State Commun. 49, 249,1984. (4) T. Hughbanks and R. Hoffman, J. Am. Chem. Soc. 105,1150 (19S3); Inorg. Chem. 21, 3578 (1982). (5) J. M. Tarascon, G. W. Hull and F. J. DiSalvo, Mat Res. Bull, 19, 915 (1984).

CCNTIMCTEK

Figure 1. Monocrysttb of M,MotTet a) M • Ct b) M - Rb. 0 15

too

too IOC TIKI

Figure 2. Resistance vs. temperature of M2Mo6Se6 samples and log resistance vs inverse temperature for Rb2Mo,Se,.

. F. J. D1 Salvo AT&T Bell Laboratories 600 Mountain Avenue Murray Hill, New Jersey 0. i74 0 16

TRANSMISSION ELECTRON MICROSCOPIC INVESTIGATION OF V5Si3 PRECIPITATES IN

V3Si (A15 STRUCTURE). A. Ben Lamine, F. Reyaaud Laboratoire d'Optique Electronique du C.N.R.S., BP 4347, 31055 TOULOUSE cedes, FRANCE and J.P. Senateur I.N.P.G., GSnie Physique, ER 155, BP 46, 38042 ST MARTIN D'HERES, FRANCE

A new lattice defect has been found (1) by transmission electron microscopy in the V.Si superconducting compound (A15 structure). It looks like a pair of dislocation lines parallel to the <100> directions and it has been analyzed as an -T <100> faulted'dipole (2) by matching conventional electron micrographs with computer simulations. However, residual contrast has been observed (3) in the case of perfect theoretical extinction of the faulted dipole (g.b - 0, g-b.u * 0 and A » 0.97 at room temperature). We first wrongly suggested (4) that this defect could be a dispiration (5) dipole. We were then led to postulate the hypothesis that the faulted dipole accounted for the plastic deformation of the V,Si matrix caused by a needle-shaped precipitate. The chemical composition of the precipitate (V-Si_) has been assessed by electron energy loss spectroscopy (EELS) and its crystallographic structure (quadratic V,Si_) has been determined by convergent beam electron microdiffraction (6). The purpose of this article is to present the orientation relationships between the V.Si, precipitate and the V,Si matrix and the results of a high resolution electros microscopic investigation of the interface between the needle-like precipitate and the matrix. The orientation relationships bet- v ween the quadratic 5Si. and the cubic V_Si have been determined by convergent beam electron microdiffraction in a Philips EM 400 (120 kV) microscope equipped with a field emission gun. They are written as follows

[loo] 7 [100] v5si ' V.Si [oio] J J 7 [021] v3si V5Si3 ' 7 [0T2] [ooi] v5si3 '

They are different from the orientation relationships that have been observed in the V5Si3/V3Si direct\onally solidified eutectic (7) and in V5Si3 films grown by reactive pulverisation of silicium onto a V.,Si single crystal at 1500'C. 0 16

H v5si3 " M v3si OI 01 [ °]v5si3 » [ °]Vi

H V5Si3 " H 73S1

The interface between the 'cSi, precipitate and the V3Si matrix has then be studied by high resolution electron microscopy and the results have been explained (8) by the theory of the symmetry applied to interphase interfaces (9). First, the high resolution technique of perfect V.Si has been assessed : by comparing experimental micrographs with the corresponding computer simulations obtained with a multislice program, it has been shown (10) that the tilted beam illumination and the selection of the four simultaneous reflections 000, 100, 010 and 110 enabled us to determine rigorously the position of the vanadium atomic rows lying parallel to the incident electron beam in a {lOOJ thin foil of perfect V.Si. Next, a faulted crystal has been observed and figure 1 shows the cross section of a needle-like V,Si_ precipitate observed edga-on in the JlOOJ V.Si matrix. It can be seen that the large interface consists of ledges parallel to <100> y gj_« The nature of these ledges is now under investigation using the DSC and CSL theories of Bollmann (11).

(1) A.Ben Lsmine, F. Reynaud, C. Mai and J.P. Senateur, Phil. Mag. 38A, 359 (1978). (2) A. Ben Latnine, J.F. Senateur and F. Reynaud, J. 1'icrosc. Spectrosc. Elec- tron. .5, 745 (1980). (3) A. Ben~Laxniue and F. Reynaud, 10th Int. Congr. Electron Microsc. Hamburg (1932), vol. 2, 149. (4) F. Reynaud and A. Ben Lamine, Acta Met. ^9, 1485 (1981). (5) W.F. Harris, Phil. Mag. 22, 949 (1970). (6) A. Ben Lamine, F. Reynaud, C. Colliex, M. Acbeche, J. Sevely and Y. Kihn, ICX0M83, J.Phys. C2, 45, 709 (1984). (7) G. Bear, H.N. Quyen and K.H. Berthel, quoted by R. Madar in Current Topics in Materials Science, 8, 230 (1982). (8) A. Ben Lamine, R. Pottier, J.?. Senateur and F. Reynaud, submitted to Scripta Met.. (9) G. Kalonji, Ph. D. Thesis, M.I.T. Cambridge (D.S.A.) (1982). (10) A. Ben Lamine, M.J. Lahana, F. Reynaud and P. Stadelmann, J. Mater. Sci. Lett. 3_, 431 (1984). (11) W. BollmannT Crystal Lattices, Interfaces, Matrices, An Extension to Crystallography, Ed. W. Bollmann (1982). 0 16

Figure J : High resolution electron micrograph showing the cross-section of the V.Si. needle-like precipitate in a 100 V,Si matrix.

Francois REYNAUD Laboratoire d'Optique Electronique BP 4347 31055 TOULOUSE cedex, FRANCE 0 17

PREJ1CTI0I OF INK TniM.TKKn COMPOBZTZOV OF COMPOUNDS WITH CUTRID niOOIAL PKXSMff II R - •! - SI STSXINS f. HovastreTdt and X. Parthe Laboratcire da Cristallographie aux Rayon* X, Universite da Geneve 24, quai Ernest Anseraet, CH - 1211 Geneve- 4, Switzerland

Tha ternary systaas R-Mi-Si (R « rara-«arth alaoant) fora a most iatarasting fiald of invastigation for crystal chaaists. Thay ara offered &z unusual richness of diffarant structure types; the Ce-Ni-Si system, which h»r, baaa tha most thoroughly studied up to now, contains over 20 distinct phases, mostly fouad by Nys'kiv (1) aad Bodak & GladyshevsJcii (2).

Coordination of Ni and Si atoms A A A only trigonal prisms 0 O tT«ionai prisms and ,3 other polyhedra

.S "5 g 1||

'S TS 15 II

20 %Si— 30 ro Fig.l: Part of the Ce-Ni-Si phase diagraa showing coapouads containing stretched trigonal prisas. 5,2,3 indicates CaJligSi,. The structure series A, S, C and 0 have the following general foraulaa ana representatives:

A: R ir M B: C: R T K a*+3n+2 n»-n+2 n*+n z+r z r t+l t-l t+l a-1 H z-2 Gd.MiSi, t-1 ysi a»2 y Hi Si 3™ r»4 J Z a»2 c. »i si C. Hi Si t*l» •"J CaNiSi, 5 2 3 r*5 7 2 5 R5™5 b>l 2

n«3 CaSi t*2 iSi V 3 b"2 *3 *2 Ja n-4 t-3 R2™2

n—. RTM t— RTM 0 17

Io aost cases it was however not clear to see any correlation between the coaposition and the structure from the data available. For structures built up of centred rare-earth prisas, a rule describing the distribution of Mi and Si atoas on che trigonal-prisa-csstre si+iet (3) allowed to gain soae new insight.

,; f_. Gd3NiSi2

' filled io Hf3Pz type oP24, Pnma

hP22, P63/m

a+b *2 2o-2 4b-2 2

o«1,b«1

CeNiSi2 oC16, Cmcnt *Z-7"— .©M©/-«.©.

Fig.2; Representative structures of each of the 4 series A, B, C and D

Applying this waist-contact-restriction rule, which excludes R-Ni and Hi-Hi waist contacts, for each known structure type, a coapositios for highest transition-metal content and maxiaua order can be foraulated. With these corrected foraulae, it is possible to correlate soae of these structures by asans of structure series, four of which are shewn in Fig. 1. One structure of each series is represented in Fig. 2. Series A is built up of triangular coluans, a prises wide with octahedral voids where six coluans aeeo. Series B certains trigonal prisas of which in one building block x. prisa centres have no R waist contact and x prisa centres do. Series C has AlB,-type slabs whfch are $ 0 17

prisu thick and series 0 consists of 4 AlB2-type segments iatergrown with J la soae cases, the ideal coaposition, obtained by the waist-ccntact-restriction rule, was considerable different fxoa that previously reported. The following table contains a list of published Ce-Mi-Si phases together with their ideal composition.

Series formula Earlier reported Predicted Ezperifflcntax coaposition coaposition verification —_ Cei4(Ni8Si3» (1) C.14KisSin R * Ce (5)

Rn*+3n+2Tn*-n+2Mn*+n C.6Ni2Si3 —— (Series A) Ce.fNi-Si ) (7)

Ce15(Mi4Si13) (8) ct15wi7si10 R •= Pr (9)

R z+,r T 2H r Ce_Ni Si_ (10) Ce7Ni2Sig R '• Ce (10) (Series B)

Ct3«1a.48ii.fi» m R »> y (11) (Series C) Ce(Hi 5Si5 > (12) CsHiSi

Q in ij Ka+bx2aM4b CesNigSi7 (2) (Series D) CeNiSi2 (13) CeNiSi2 R = Ce (13) R > Ce3Ni2S.g .(14) C«3Ni2Si8 Ce (14)

References (1) M-G. Mys'kiv, Thesis, Ivano Franko Univ., L'vov, USSR (1973) (2) O.I. Bodak & E.I. Gladyshevskii, Inorg. Mater. 3, 1754 (1959) (3) £. Parthe, B. Chabot & E. Hovestrtydt, Acta Cryst. ££2, 596 (1983) (4 7a.P. VarzoJ.yuk, L.Q. Akselrud, yu.H.Grin', V.S. Fundaaenskii ft E.I. Gladyshevskii, Sov. Phys. Cryst. g±, 332 (1979) (5) E. Hovestreydt, J. Less-Cooaon Met. iP_2, L27 (1984) (S) O.I. Bodak, E.I.Sladyshevskii e O.I. Kharcheako, Sov. Phys. Cryst. 19, 45 .'1974) (7) O.I. Bodak, E.I. Sladyshevskii & M.G. Mys'kiv, Sov. Phys. Cryst. 17, 433 (1972) ~ (8) M.G. Mys'kiv, O.I. Bodak & E.I. Glauyshevskii, Sov. Phys. Cryst. 18, 450 (1974) ~ (9) E. Hovestreydt & S. Parthe, Acta Cryst. C, submitted. (10) M.G. Mys'kiv, sec Structure Reports JiA, 44 (1975) (11) X. Klepp 6 S. Parthe, Acta Cryst. BJJfi, 2026 (1982) (12) O.I. Bodak, M.G. Mys'kiv, A.T. Tyvanchuk, 0.1. Kharcfaenko 6 E.I. Gladyshevskii, Inorg. Mater. £, 777 (1973) (13) O.I. Bodak • £.1. Gladyshevskii, SOT. Phys. Cryst. H, 359 (1970) (14) J.A. Stapisn, K. Lukasiewicz, I.I. Gladyshevskii * O.I. Bodak, Bull. Acad. Pol. Sci., Ser. Sci. Chia. 20, 1029 (1972)

Mr S. Hovestreydt Laboratoire de Cristallographie aux Rayons »* #tn»i T-r*mm+. inurut. CH - 1211 0 18

GROWTH SIMULATION OF BORIDES COMPOUNDS. S. Hamar-Thibault, R. Hamar Institut National Polytechnique de Grenoble - Laboratoire de Thermodynamique et Physico-Chimie M£tallurgiques (L.A. 29) - E.N.S.E.E.G. Domaine Universitaire, B.P. 75 38402 Saint Martin d'Heres, France.

From the description of the structure of an inter-metallic compound, and from various energetical considerations, we developed numerical calculations which permit : - to deduce from attachment energy and superfical energy of a (hki.) face, the growth and the equilibrium morphologies.

- to simulate the formation and the growth of a nucleus on the different (hki) faces.

The first approach was previously presented in the case of the CrB compound (1), and led to the construction of the growth and equilibrium morphologies. These calculations were extended to the simulation of formation and growth of nuclei on some faces.

1 - ENERGETICAL CALCULATIONS.- Our model is based on the variation of free energy of a system after the deposition of the atom (p) on a nucleus of (p-1) atoms. = T *p <*p - *v The expression of aH and &S are obtained after some approximations from the bond energy EL(p,j) between the atom p and the other j atoms of the nucleus. If n is the number of bonds of the atom p with the crystal, and Np the number of bonds of the atom j in the crystal : % N P AHp = % EL(P.J) and ^p * 2T Z

At the temperature T = Tu + &T "p " AS- - X EL(p.j) - I (1 - jl) X EL(p.j) P j = l £ TM j-1 The bond energy EL(i,j) is calculated from a pair interaction potential of Leonard-Jones type : 0 18

where m * n/2 * 6.

For a binary compound, three pairs of parameters (e?. and £?.) must be known. These parameters are obtained from thermodynamical and crystallographical considerations.

The energy which is necessary to the deposition of a group of atoms (from the atom of range Pj to the atom of range p~) is evaluated with the under- cooling AT rev : P, np p2 Nn o ? X EL (p,j)/ Z L EL (p,j)

2 - RESULTS.- To predict the nature of the faces of a crystal, Hartman and Perdok (2) introduced the notion of periodic bond chains (PBC - assembly of strong bonds which are periodically repeated according to certain crystal directions). This concept allows to distinguish three crystalline faces F, S and K according to the number of PBC contained in them. The "F" faces are parallel to two or more PBC ; they are growing from a bidimensional nucleus. These faces form the morphology of the crystal because their growth kinetics are more slowly than the other faces. The "S" faces are parallel to 1-PBC and are growing from a unidirectional nucleus. The "K" faces are growing atom after atom.

The calculations were used to examine the growth of nucleus on some faces of CrS (020), (110), (002), (130), (220), (131), (041), (021). We report the results for three typical faces.

The {020} face grows from two square nucleus : first a nucleus of 4 atoms UT4.» 693 K), then a nucleus of 9 atoms (AT3 = 490 K). The nucleus growth continues with the deposition of CrB, group (AT. » 230 K) followed with CrS group (AT »0K), alternatively in the [100] and [001] directions. This face is a "F" face with a two dimensionnal nucleus and a growth layer by layer. 0 18 001

•

D. •. • D« ^""brBs* "ft. / / */ ioo / DgJ < ^ — •» \ •« D. \ n D» Da Da

The {111} faces grows from a nucleus in one direction : for example the [1CI] direction, which is the direction of the [101] PBC. This face is a "S" face.

ATCrB= 426K * B D Cr

The {131) face is a face o* "K" type and the deposition occurs on all the face without preferential directions.

3 - REFERENCES.- (1) S. Hamar-Thibault, R. Hamar, VII International Conference on Solid Com- pounds of transition elements, Grenoble 1982. (2) P. Hartman and W.G. Perdok, Acta Cryst. 8, 49 (1955). (3) P. Hartman, Crystal Growth, Ed. P. Hartraan, North Holland, 1973.

S. Hamar-Thibault, Institut National Polytechnique de Grenoble - Laboratoire de Thermodynaniique et Physico-Chimie M€tallurgiques (L.A. 29) - E.N.S.E.E.G. Oomaine Universitaire, B.P. 75 38402 Saint Martin d'Heres, France. Q 19

CRYSTAL CHEMSTKY, STABILITY AND ORDER IN TERNARY METALLIC PNICTTDES

R. MADAR, E. DHAHRI, P. CHADDOUET, J.P. SENATEDR, R. FRUCHART

ER. 155. CNRS. ENSIEG. Dotoalne Universitaire BP 46, 38402 Saint Martin d'Heres, France.

B. LAMBERT Laboratolre de Cristallographie du CNRS 25, avenue de3 Martyrs, 166 X, 38042 Grenoble Cedex, France.

Unlike the classical crystallographic approach based on a rlgourous analysis of the symmetry elements, the crystallochemical analysis starts from the principle that the groups of atoms reflect the strength of chemical bonding. From a geometrical point of view, this analysis is more concerned by the local environments, which contrary to the classical crystallographic analysis, offer the possibility to neglect in a first approximation the local distortions of the lattice. Starting from the geometrical principles for the construction of numerous phases, the crystallochemical analysis leads to a schematic structural description which enables a systematic analysis of the bonding forces and related physical properties. We have used the previous crystallochemical analysis of the M2P phases (1) based on the metalloidic environment of the metal for the description and the prediction of the physical properties of alloys with the general formula

(Mx M'x_x)2 X where M is a transition metal, a rare earth or a void, M' a transition metal of the 3d and 4d groups and X represents phosphorus or arsenic. The corresponding families of alloys are listed in table 1. This crystallochemical description permits in a first step to give an uniform representation of different structures, starting from an array of triangular channels of metalloid. 0 19

Then this approach is used for the prediction of the substitutional

order, the explanation of the superstructures, and as a consequence it must be

considered as a powerful technique for the research of new compounds with

choosen properties.

These different applications of the crystallochemical analysis will be illustrated by recent examples from our present studies on ternary rare earth transition metal pnictides.

References :

(I). R. FRUCHART Ann. Chim. Fr., _7, 563, (1982).

Table 1

Alloys (Mx M'i-X)2 X

1-x R formula structural type Example X

0.5 1 M M'X Hexagonal Zr Ru Si Zr Ru P HT

Orthorhombic Ti Ni Si Zr Ru P BT

Orthorhombic Ti Fe Si Hf Ru As BT '

Tetragonal Pd Co P Rh Ni As

0.33 2 M2 M'4 X3 Hexagonal Hf2 Co4 P3 Zr2 C°4 P3 Orthorhombic D2 M'4 x3 Rh4 P3 Rh* ?3

0.23 3.33 «6 M'20 X13 Hexagonal Zrg Ni20 ?i3 Ln6 Ni20 P13 1 O6 M 20 X13 Hexagonal Rh2 0 S113 Rh20 Si!3 ,' f:

1 0.14 5 M2 M 12 X7 Hexagonal Zr2 Fei2 P7 Ln2 Nii2 P7 1 a2 12 X7 Hexagonal Cr^ 2 **7 Rhi2 As7

R. MADAR. ER 155. CNRS. ENSIEG. Doaaine Oniversitaire, BP 46, 38402 Saint-Martin-d'Heres, France. 0 20

INVESTIGATIONS INTO CRYSTALLINE AND AMORPHOUS COPPER - ARSENIC - CHALCOGENIDES R. Blachnik Anorganische Chemie Universitat Osnabrtlck, Barbarastrafle 7, D-4500 Osnabruck G. Kurz Anorganische Chemie Universitat-GH-Siegen, D-5900 Siegen

The sections Cu_S + As_S3 and Cu2Se + As2Se3 (1) of the ternary systems Cu - As - S (Se) have been investigated by means of thermoanalytical and x-ray methods. Both sections are quasibinary and contain a series of ternary phases. Single crystals of these and other ternary Cu - As - S (Se) phases were grown from salt melts. The structure of the phases

Cu3As3S4, Cu4As2S5, CulaASgSe21 and

Cdt-Cu.,. As_Se24 are given. The structure determinations

revealed that all phases on the Cu2X - As2X3 sections are derived from the sphalerite structure of the high-temperatur modifications of Cu_X. Phases with higher chalcogen contents

(Cu3AsSe4, HT - Cu3AsS4) are also based on this type, whereas the CuAsX phases have close packing of the atoms with mixed hexagonal or cubic stacking sequences. The first class of compounds differ in the number and position of anion vacancies, the latter is characterized by metal-metal bonds. Super structures are obtained in the first class by different ordering of the vacancies, in the second class by different stacking of the layers.

Differential scanning calorimetric and FJR investigations of the glasses in the ternary systems revealed two regions of different structure in the Cu - As Se system. The border running approximately along the As55Se4_ - Cu.gSeg.. section. 0 20

The classes with higher Se-contents are based on the As-Se, structure, those with lower Se-contents contain probably metal- metal bonds, similarly to the structure of the solid compounds. The structural base of the Cu - As - S glasses is the As-S,- network. The different recrystalization behaviour of these glasses is discussed and correlated to some atomic parameters.

(DR. Blachnik, G. Kurz, J. Solid State Chem. 1984 0 21

STRUCTURAL AND MAGNETIC PHASE DIAGRAM FOR THE MnAs — CrAs SYSTEM H. Fjellv&g and A. Kje&shus Department of Chemistry, University of Oslo, Blindern, 0315 Oslo 3, Norway

MnAs takes a special position among the 3d transition metal monoarsenides due to its intriguing transformation properties. The prominent features are a second order NiAs ,P £ MnP,P [P = paramagnetic] type transition at T_ = 390 K and a first order, hysteresis accompanied MnP.P to NiAs ,F [F = ferromagnetic] type

transition at 306 [Tn ,]—317 K [T_ .]. o,a L , i numerous studies have been devoted to the ternary solid solution phases Mn. .T.As and MnAs.. X with T = Ti — Ni, Mo and X = P, STD , Ge , Sn. In order to provide an example of the complex structural and magnetic behaviours often encountered in these phases, ve have chosen to present the phase diagram [Fig. l] for

1200

1000

N3,i, -I

N3,d

MnP, Hc

0.00 Q20 QAO 0.60 Q80 1.00

Fig. 1. Phase diagram for Mn, .Cr.A3. 0 21

the MnAs — CrAs system (1-3). The following transition temperatures are included in Fig. 1:

i) TD: HiAs.P * MnP.P

ii) To . and T_ . : MnP.P •*• HiAs ,F and MnP.P • NiAs.F, respectively

iii} T : NiANi, s ,F •*• MnP,H [Hfl * a-axis helimagnetie ; Pnma setting c>a>b]

iv) I». . and T_T . . : MnP,P *• MnP,H and MnP,P •»• MnP.H , respectively

v) T ana TJJ : MnP,H - MnP,H and MnP.H^, ->- MnP,H , TN2,N2 i and TH_ . d c a respectively [H = c-axis helimagnetic]

vi) TN3: MnP5P * MnP,Hc

vii] Tw, . and Tw., . : MnP,P •<• MnP,H and MnP.P ->• MnP ,H^ , respectively

These phase transitions separate five different phases; NiAs,P, HiAs,F, MnP.P, MnP.Ii and MnP,H , of which three exhibit nagnetic long range order. A triple point exists at t % 0.33 and

T % 200 K, where the first order lines T_. and T^2 meet the second order T__ line. There occurs a tricritical point close to CrAs, t % 0.90 and T % 290 K, where the MnP,P to MnP,H phase transition converts from first to second order. A few comments on the transitions i) - vii): i) The NiAs ,P •*• MnP,P type transition involves continuous, displacive shift of the atoms from their NiAs type positions. Parallel with the crystallographic MnP type deformation there occur a reduction of the paramagnetic moment which -inter alia is manifested by anomalous magnetic susceptibility curves.

ii) The first order MnP ,P -»• HiAs.F type transition is accompanied by hysteresis [AT > 10 K]. The increase in the unit cell volume of the low temperature ferromagnetic phase (u_ % 3.6 w_, if a spins in the basal plane! is mostly due to magnetostriction. 0 21

iii) The pressure induced MnP ,H type state [characterized by neutron diffraction] transforms on heating at atmospheric pressure irreversibly to the HiAs ,F type state. ivl The combined structural and magnetic MnP ,P * MnP ,H type transition is accompanied by some 2% discontinuous reduction in unit cell volume,but only minor changes in the positional parameters occur. The hysteresis upon heating and cooling is only slight [say AT < U K]. The helimagnetic parameters which specify the H state [propagation vector, phase difference between the independent spirals and magnetic moment] vary with t.

v) The transition between the helimagnetic H and H states a a is connected with a large hysteresis. This transition is accompanied by large changes in the unit cell dimensions , whereas the positional parameters are only slightly affected. There appears to be no significant difference in magnetic moment between the H and H states . a a vi) The second order MnP,P j: MnP,H type transition goes through a broad maximum at t ifc 0.8, T % 290 K. Apart from the propagation vector the helimagnetic [as well as the crystallographic] parameters for the H state vary very little over the range 0.39 < * < 0.90 [1.00]. vii) The MnP.P to MnP,H type transition is of first order for ^0.90 < t <_ 1.00. The unit cell volume of the H state exceeds that of the P state. MnAs substituted CrAs behaves as a "negative" pressure system and the first order behaviour is be- lieved to originate mostly from magnetostriction.

(1) N. Kazama and H. Watanabe, J. Phys. Soc. Jpn. 3_0, 1319 (1971) (2) K. Selte, A. Kjeishus, P.G. Peterzens and A.F. Andresen, Acta Chem. Scand. A _32, 653 (1978) (3) H. Fjellvag and A. Kjetshus, Acta Chem. Scand. A £&_, 1 (193U) 0 22

STRUCTURE AND BONDING OF TRANSITION METAL CARBIDES AND HYDRIDES J. Hauck Institut fur Festkorperforschung, KFA Julich, D-5170 Julich, FRG

Binary transition metal carbides and hydrides with face-centered cubic or body- centered cubic metal matrices are compared. Both systems contain non-stoichiometric solid solutions with a disordered distribution of C and H atoms on octahedral or tetrchedral interstices at high temperatures. Transition metal carbides order below ~ 1000 °C, transition metal hydrides below "0 °C because of the weaker bonding. For the same reason the lattice of the ordered hydrides is less distorted. The high melting point of carbides can be correlated to the strong bonding. The melting point of hydrides can roughly be determined by nsec laser pulse experiments to avoid the decomposition occuring in classical methods.

The specific ordering of carbon and hydrogen atoms were analyzed for the type of bonding. Ordered phases of transition metal carbides MC at x = 0.88 (V), 0.83 (V,Nb), 0.63 (Ti) and 0.5 (Ti,Zr,RE) can be described by a structure model considering repulsive interactions of carbon vacancies (or carbon atoms, resp.). Each carbon vacancy exhibits the same number of first, second and third nearest neighbors at compositions x = l-n/24 (n = 0,1 ... 12). The comparison of the different ordered structures of this model shows that short and long range ordering at x = 0.83 is favoured by a maximum of Coulomb energy, while covalent bond energy with shortened M-C bonds favours x r 0.5 and x = 1 compositions. Some unusual physical properties like the maxima of melting temperature, lattice constants, critical resolved shear stress and activation energy of diffusion at x — 0.8 can be correlated to the maximum Coulomb energy.

Many transition metal hydrides like PdH 5> /S-NbH and £ -VH order by the same principle with a further limitation. Each metal atom exhibits equal numbers of 1 ** 3 H atoms at compositions MH , x = 4, -f, -rt 1 for te«.rahedral coordination and 111 x = -?, •=•» "2 • • • for octahedral coordination of H atoms. Calculations of Coulomb energies considering repulsive forces between H atoms with the charge ~ 0.3 exhibit maxima at the distortion of the metal lattice as it is observed by x-ray diffraction. £-VH exhibits the maximum Coulomb energy, while this is reduced by •* 1 % in hydrides with (&-NbH structure. The difference of PdH _ . structure compared to TiCn _ or GdC_ c structure can be attributed to different bonding mechanisms of three H or C atoms in octahedral coordination. Th< 0 22

meridial configuration of three H atoms in PdHn _ is favoured by repulsive Coulomb interactions which is increased by 1.9 8. The facial configuration of C atoms in TiC_ 5 and GdC_ - is stabilized by covalent bonding with reduced M-C bond distances, e.g. reduced by 0.03 8 in ordered ZrC -

Hydrogen atoms in transition metals left of the sixth group are considered with a partial negative charge, while transition metals right of the fifth group exhibit a higher electronegativity thus reducing the electron density at the H atoms. Non-stoichiometric carbides of the third and fourth group can be loaded with hydrogen, while hydrogen uptake of carbides with transition metals left of the fourth group is more difficult depending on stoichiometry and structure of the carbide. 0 23

INTERBAND DIELECTRIC RESPONSE IN TRANSITION METAL HYDRIDES M. Fliyou L.S.O.C.S. (LA 232 CNRS) - Institut de Physique 5, rue de 1'Universite 67084 Strasbourg Cedex (France) R. Riedinger Universite de Haute-Alsace, Laboratoire de Physique et de Spectroscopie Electronique - 4, rue des Freres Lumiere 68093 Mulhouse Cedex (France) M.A. Khan L.M.S.E.S. (LA 306 CNRS) - Universite Louis Pasteur 4, rue Blaise Pascal 67070 Strasbourg Cedex (France)

The transition-metal-hydrides have been very actively persued during the last few years because of the Hydrogen storage and superconductivity (1). Their physical properties such as specific heat, susceptibility and superconductivity critical temperature, have been experimentally measured and theoretically calculated (1, 2). This particular interest in metal-hydrides has been also the origin of many band structure calculations (3). In spite all this, the study of optical properties are still rare and few. As compared to metals, their hydrides will have their own characteristic absorption peaks.

In the present note, we wish to illustrate the fact that the metal-hydrides have their characteristic optical properties by comparing the imaginary part of the dielectric constant e-(u) (oi is the photon energy) of PdH and NiH with those of Fd and Ni respectively.

The interband contribution to the imaginary part of the dielectric constant (in atomic units) is (4) :

1 i2 dS 2irai2 nn' S |Vu ,—j where S is the constant energy surface defined by :

S " { * ; En'k " Enk "

F ,r- is the dipole matrix element between the occupied |nk> and unoccupied | a'lt> states :

6 is the usual step function, E_ the Fermi level and H is the Slater-Roster •: Hamiltoniau. 0 23

Using the band-energy schemes of Papaconstantopouloa et al (3) and of Switendick (3) for FdH and NiH respectively', we have calculated the interband contribution to e.Cu) (4, 5). Contrary to Pd and Ni their hydrides are opti- cally transparent up to a certain limit of photon energy. The threshold energy for the band to band transitions is .408 eV for NiH and .952 eV for PdH. This particular behaviour can be explained from the band schemes where the Fermi surface of these hydrides intersects only non-degenerate bands, which is not true in the case of Ni and Pd.

There are many structures in e»(w) versus to for NiH as well as for PdH. The most intense peak in the case of NiH is situated at 6.46 eV. The corresponding structure in pure Ni is observed by Johnson and Christy (6) at 4.8 eV. Laurent et al (7) obtained a doublet at 5.1 and 5.6 eV in the optical conductivity from their band scheme of Ni. Hence, the main absorption peak in Ni is shifted towards higher energy in the presence of Hydrogen.

In the case of PdH we obtain a small peak at 3.4 eV and a very intense peak at 5.58 eV (4). Besides these two main peaks, we also obtain a number of other small structures. Johnson and Christy (6) have reported structures at about 2.2 eV and 4.8 eV in the optical conductivity of Pd.

Unfortunately, there have been no optical measurements on NiH and PdH as yet, but the interband transparency and the shift of the main absorption peaks towards the higher energy as compared to the corresponding transition metals are very interesting theoretical observations to be persued experimentally.

(1) "Hydrogen in Metals", Topics in Applied Physics, Vol. 28 and 29 (Springer Verlag, Berlin) 1978. (2) R.B. Me Lellen and C.G. Harkins, Mat. Sci. Eng. _1£, 5 (1975). J.E. Schirber and C.J.N. Northrup, Phys. Rev. B J£, 3818 (1974). (3) A.C. Switendick, Ber. Bunseng. Phys. Chem- J6, 535 (1972). J.S. Faulkner, Phys. Rev. B _13_» 2391 (1976). D.A. Papaconstantopoulos, M.B. Klein, J.S. Faulkner and L.L. Boyer, Phys. Rev. B ^J», 2784 (1978). M. Gupta and J.P. Burger, in "Recent Developments in Condensed Matter Physics", 71. 4, eds. J.T. Devreese, L.F. Lenmenj, V.E. Van Doren and J. Van Royen (Plenum, NEv York, 1981) p. 309. 0 23

(4) R. Riedinger and M.A. Khan, Phil. Mag. B 84, 547 (1981). M.A. Khan and R. Riedinger, J. Phys. (Paris) 43, 321 (1982). (5) M. Fliyou, R. Riedinger and M.A. Khan, Phys. Letters A _104» 379 (1964). (6) P.B. Johnson and R.W. Christy, Phys. Rev. B ^, 5056 (1974). (7) D.G. Laurent, J. Callaway and C.S. Wang, Phys. Rev. B 20, 1134 (1979).

M.A. KHAN LMSES (LA 306 CNRS) Universite Louis Pasteur 4, rue Blaise Pascal 67070 STRASBOURG CEDEX France P 1 A 1 STRUCTURAL AND MAGNETIC PROPERTIES OF COBALT GALLIUM SULFIDES:

a-CoGa2S4 AND y-CoGa2S4. E.AgostinellijL.GastaldirfM.G.Simeone and S.Vtticoli ITSE-CNR, Area della Ricerca di Roma,CP 10 - 00016 Monterotondo Stazione (ROMA) - ITALY

CoGa2S4 has been recently synthetized as powder and two crystallographic phases have been found(1): a low temperature tetragonal a- phase and a high temperature orthorhombic j3-phase. In the* framework of a wide investigation on the chemical and physical properties of compounds with ABjX. stoichiometry, we report in this paper the crystal growth and the single crystal X-ray investigation on CoGa_S- system.A correlation between crystallographic and magnetic properties is presented.

The a-phase of CoGa_S4 was grown using the vapour-phase chemical- transport, while any attempt to obtain crystals of the j3~phase failed. Moreover a new phase of CoGa2S4 ( y-phase) was grown using KBr as flux.The magnetic properties were investigated in the temperature range 4.2 * T< 300K. a-CoGa_S4

Single crystals of a-CoGa2S4 were prepared by vapour-phase chemical transport using I- as trans-

port agent (Td=780*C;Ts=860*C).Af- ter a week of thermal transport, /£T\ black metallic crystals in oblong parallelepiped form were obtained.

# Co The a-CoGa2S4 (Fig.1) crystallizes in the space group IT (a*5.2538(8) 0 Ga A;c=10.393(2)A).While the Ga(2)-S distance (2.311A) is in agreement with the literature data, the Co-S and the Ga(1)-S distances are unex Fig.1 pectedly similar (2.281A). The tetragonal distortion of the unit cell along 'the c axis (c/a-1.98) is anomalous with respect to the

analogous CdGa_S4. These structural results suggest the presence P 1 A 1 of little degree of disorder between Co and Ga(1) ztoms in the cationic sublattices. The reciprocal magnetic susceptibility of

a-CoGa2s4 is reported i-n Fig.2 . A Curie-Weiss law is followed in t-he range 100«T$300K with 9 =-78K and C=2.1 .Below T=100K the magnetic susceptibility de- viates noticeably from the Curie- Weiss law. This behavior is typi

cal ofv the occurence, of short ran ge magnetic order (2), which can be associated, in our case, to the little degree of inversion between Co and Ga(1) ions among the cationic sites. The maximum in the susceptibility at T = 7+ ito 109 0.5K was interpreted as a N§el temperature.

The y-form of CoGa2S4 was obtained starting from a mixture of pow dered a-CoGa2S4 and KBr as flux. The mixture was kept for two days at 900 C and then the temperature was lowered to 700°C at a rate of 1°C/h. Finally the mixture was quenched to the air. The

7-CoGa2S4 (Fig.3) crystallizes in the space group F4~3m, and in the cubic cell (a=5.221(2)A); the metal ions and the vacancies are randomly distributed aroag all the available cationic sites. This metal distribution deter- mines an increase in the number of Cc-S-Co paths by which the magnetic interactions occur Fig.3 The magnetic susceptibility(Fig 4) follows a Curie-Weiss law in the range 200 4 T.$3OfiI\; and the high negative 9 value'. 9 «-42OK) reflects a strengthening of the Co+1/4Q+1/2 Ga antiferromagnetic interactions ad a conseguence of the increased relative Importance of the PI

Co-S-Co superexchange pathway with respect to the Co-S-Ga-S- Co one. Below T-200K the inverse susceptibility shows a downward curvature to the origin . This behavior can be related to the presence of small clusters of exchange-ccupled Co ions with a resultant magnetic moment and, consequently, their inverse susceptibility will be loo no zero at T=0K .

REFERENCES (1) M.P.Pardo, Mat. Res. Hull., YT_, 1477 (1982) (2) D.Fiorani and S.Viticoli, J. Phys.Chem. Solids, 41_, 1041(1980)

Dr. S. Viticoli ITSE-CNR, Area della Ricerca di Roma, CP 10 00016 Monterotondo Stazione (ROMA) - ITALY P 1 A 2

PHASE RELATION AND AGING EFFECTS IN Fej^O^S SYSTEM E. BarthSlemy, C. Carcaly Laboratoire de Chimie Minerale Structurale Associ§ au CNRS - LA 200, 4, Av. de 1'Observatoire, 7527C Paris, Frauce.

Continuous solid solution (NiAs type) between FeS and Cos can be obtained by quenching from high temperature (2; 700° C) to room temperature. I - Crystallographic data The system exhibits two phases, which have been determined by X-ray powder diffraction methods : one, for 0 >$ x < 0.17 ; the other, for 0.17 < x .$. 1. The c parameter of the first phase is greater than that of the second one. A mixture of the two phases is usually obtained for x = 0.17. For x < 0.17, a superstructure (a /3\ 2 c) of the NiAs type is obtained, while the other phase (x > 0.17) exhibits a pure NiAs form (a, c) - {1 C-phase). For x > 0.60, compositions are not well-established because a mixture of the 1 C-phase and CogSg is generally precipitated. It is necessary to quench the 1 C-phase. If this is not the case, the composition is segregated into 1 C, 2 C-phases and CogSR. II - Phase transition

At Ta = 425 ± 2K, FeS shows a first-order phase transition accompanied by a structural change. By heating, the superstructure (a /T, 2 c) vanishes ; the transition 2 C-• 1 C is endothermic and is associated with a drastic volume contraction. Changes are also found in magnetic susceptibility and resistivity. Co s In Fe x system, such a transition is also observed. This transition is sensitive to Co contents, aging effects and thermal treatments : these effects are reported here. - Influence of Co contents : These effects have been investigated mainly by DTA and magnetic measurements. Results on unaged specimens are summarized as follows : In the forward direction (i.e. : heating rate for x > 0.17 ; cooling rate for x < 0.17).

Ta decreases with increasing x.

For x * 0.17, Ta lies near room temperature. P 1 A 2

The amplitude of the transition is strongly reduced with increase in Co content ^S2c—*>1C ^ ^ J K"1*1101""1 for x = 0, 1 AS1C _^2C=#1.7 J K-lmol" for x = 0.22). The transition is not detected for x > 0.30. The width of thermal hysteresis is within several degrees and seems to increase for x > 0.17. - Aging effects When specimens are aged at room temperature, a typical behaviour is observed for each phase. - The 2 C-phase

In the forward direction (first heating run), Ta gradually increases with the lapse of time and reaches its final value about 3 weeks later.

- Increasing in Ta depends on the composition (about 20 degrees for x = 0.15, while no aging effects are detected for x 4 0.05). No crystallographical differences are found in X-ray charts.

Entropy changes at Ta are detailed and indicates that the 2C-phase is stabilized by aging. Thermal treatments through the transition on aged and unaged specimens are described. We assume that dislocations phenomena control in part the aging process. - The lC-phase In the forward direction (first cooling run), we observe :

at first, the decreasing of Ta and the stabilization of the phase by aging,- without changes in lattice parameters. 3 months later, the broadening of X-ray lines, while c parameter is increasing, and the vanishing of the transition. for very long time aged specimens (2 years), no transition at low temperature. Evidence for a 1C—» 2C transformation is shown by magnetic measurements. Thermal treatments at low temperature, on the unaged 1C- phase or on the lC-phase in equilibrium state indicate that this phase is not very sensitive to the subsequent cooling-heating runs through the transition. This result differs considerably of that obtained with P 1 A 2

the 2C-phase and explains why this phase can be considered to be relatively soft. However, further experiments have to be carried out to understand clearly these phenomena.

Mailing address of author : Melle E. Barthelemy Laboratoire Chimie MinSrale Structurale 4, Av. de 1'Observatoire 75270 Paris FRANCE P 1 A 3

PHYSICAL PROPERTIES OF THE SOLID SOLOTIOUS TlCu- Fe , 0

The compound TlCu^Sep is a p-type metal owing to one valenceVband hole per formula unit (1). As in other chalcogenides (except oxides}, copper is nonovalent (2).vBy partial substitution of copper by metals with a higher valence, the valence band may become filled. For the composition TlCu. cnMeQ ...Se^, Me= Al, Ga or Fe, semiconducting properties vere found (3), indicating that these sub- stituents are all trivalent, supplying tvo more electrons per atom to the valence band as compared with copper.

We have prepared solid solutions TlCu5 Me Se0, Ck x <1 (Mep Fe or Mn). The compqsitional dependence of the cell parameters shows- a pronounced differ^ ence between the two solute atoms- (Fig. ]). indicating different valencies. We have studied the iron system in more detail in the range 0< x <0.50 where iron behaves as a trivalent metal. The valence band is then partially filled leaving p'= 1 - 2x holes per formula unit. Electrical transport measurements support this view. The magnitude of the electrical resistivity (p) of the solid solutions is that of metals (hole conduction), but only tor the lowest iron concentrations was- a normal p<* T dependence observed. '' The Hall effect shoved a narked field-dependence at low temperatures. In magnetic materials, the Hall resistivity (p - E /'j ] contains an anomalous con- * xy y x tribution: p 3 R B + R y M, which above T , assuming a unity demagnetization xy o s o c factor (thin sample), may be expressed in terms of the applied field as p = y xy (R + R x «)5 • Due to the'magnitude of the extra term, it proved difficult to experimentally determine the "ordinary- Hall coefficient'1, R . Magnetization data had to be used. We conclude that p'=» 1 ^ 2x is- most likely valid also for compositions where this procedure could not be applied, The temperature dependence of the "spontaneous Hall coefficient", R , suggests that, although spontaneous-nagr netization occurs, the system does- not attain complete order even at TF 0 K, Thermoelectric measurements showed anomalous (magnetic 2 contributions to the SeebecJc coefficient (a) at low temperatures,-while above =300 K typical metallic behaviour (a« T) with —^ -?-"^=- >0, indicating hole conduction. We found a broad-band relation between the Fermi energy and the valence-band hole concentration (Ep- p2'^, assuming p'» 1 - 2x (Fig. 2). P 1 A 3

1.0

§• as

0.5 1.0 -27 Fig. 2. The Fermi energy (from thermoelectric measurements) as a function of the volume hole concentration (p= 2p'/V with p'= 1 - 2x) drawn asc?5eir loga- rithms. The slope of the straight line is 0.67 to be compared with 2/3 expected for a broad band.

1. The concentration dependence of the cell parameters (tetragonal type) of the solid solutions TlCu Me Se : O, Me= Fe; 0. Me= 2 x e22: The distance Tl-Se (iron system) is also snovn as ccalculated for z_ = O.36 illustrating the change in effective Se radius when the valence band is being filled (0< x <0.50).

The magnetic susceptibility may, at higher temperatures, be described by x= T - 0 + XP' the latter term representing Pauli paramagnetism. The paramagnetic moment deduced from the value of the Curie constant was smaller than expected for high-spin trivalent iron in a tetrahedral environment. However, if the Pauli paramagnetism is taken into account, the itinerant holes talcing part in the magnetic interactions, a decrease of the Curie constant with respect to the spin-only value may be anticipated (k). The asymptotic temperatures (e) are positive except for the iron concentrations approaching *- 0.50, the limit where the valence-band holes (giving metallic conduction), disappear. Spontaneous magnetizations (Faraday method) were measured below characteristic temperatures of the order of 50 K. The magnetic moment per iron shows a maximum in the middle of this concentration range where the end members are respectively a Pauli paramagnetic metal (x« 0) and an antiferromagnetic senicon^ P 1 A 3

ductor (x* 0.50). The magnetization obtained at x» 0.25 may be well explained (5) by assuming a moment of 5v-a for iron {x per formula unit) and 1n_ for polarized valence-band holes (p'» 1 - 2x per f .u- i. This s-inple picture does not hold for the other compositions. At low iron concentrations, only long-rrange interactions- can coaje into play. Here, the valence-band holes must take part which yields- a coupling of the iron spins of predominantly ferromagnetic character (0 is- positive!. At i'n-; creasing iron conteuu, the occurrence of iron near neighbours gets-more likely which creates an increasing possibility of (short-range), superexchange of an antiferromagnetic character. The tvo kinds of interaction thus have a) different sign, bl different range and c) different strength, both being concentration dependent. This leads to competition (partly observed in the sign of 91 which, at temperatures below -20 K, yields- spin^ordering frustration reflected in hysteresis phenoaena. As to higher iron concentrations, 0.50<_x <2, the valence-rband holes- axe no more pres-ent, which excludes any long-range exchange of importance. This- nay lead to very weak coupling between iron atoms 3ituated rn adjacent layers, for these compounds have a large c/a ratio bringing the layers- far apart. Recent MSssbauer spectroscopy measurements on (non-^stoichiometricl TlFe-^p ^ave snovn (6) that it may be considered a 2-difliensional Xsing antiferromagnet (!„= 1*50 K).

(1) R. Berger and C.F. van Bruggen, J. Less-Common. Met. 99, '13 ( 198M (2) J.C.W. Folraer and F. Jellinek, Ibid. 76, 153 (1980) (3) G. Brun, B. Gardes, J.-C. Tedenac, A. Raymond and M. Maurin, Mater. Res. Bull. Jii, 7*»3 (1979) {h) M. Otto et al., University of Groningen, Private communication (198M (5) J.C.W.'Folmer, R.J. Haange and C.F. van Bruggen, Solid State Comn. 36, 7U1 (1980) (6) L. flaggstrom, H.R. Verna, S. Bjarman and R. Berger, Unpublished: Contribution to this conference

Correspondence to: Dr. C.F.van Bruggen, Chemische Laboratoria, Nijenfaorgh 16, NL- 9747 AG Gronfngen, The Netherlands P 1 A 4

STRUCTURAL CHARACTERIZATION OF BULK AND THIN FILM PHASLS OF

Ni1+xTe2 (C*x«l}

Suraj Bhan and Mrityunjaya Singh department of Metallurgical Engineering, Institute of Technology, Banaras Hindu University, Varanasl 221005, India

Chalcogenides of iron, cobalt and nickel (1), of which NiTe2 is one, represent a special class of materials having a rather wide homogeneity range (2,3). The extent of homogeneity of these phases depends on the extent of filling up of interstitial positions by metal atoms. Though many studies have been carried out on the massive phases of these transition metal compounds (alloys), very little is known regarding the structural and physical characteristics of the vapour deposited phases. In view of this, the emphasis in the present investigation was laid on the structural characterization of bulk and vapour deposited 0 phases in the Nlj+xT^ \< x N< 1) composition range.

(A) Characterization of Ni^Te^CO ^ x ^ 1) in bulk form

In the Ni-Te system the Nii+JCTe2 is known to have a wide homogeneity. In the present investigation homogeneity range of this phase, samples annealed at 743K and quenched in water,was

found to be NiTelo8-NiTe2(51.92-66.66 at>£Te). X-ray diffraction data of the alloys of this composition range has been given in Table 1.

Table 1. X-ray diffraction data of Ni1+J£re2 (O^x^l) alloys quenched from 743K

Alloy Composition Lattice Parameters(nra) Axial Ratio (x) a c c/a 0.00 0.386 0.530 1.370 0.12 0.388 0.531 1.369 0.33 0.390 0.532 1.364 0.45 0.391 0.533 1.363 0.64 0.394 0.536 1.360 0.65 0.395 0.537 1.359

The unitcell expands with increasing concentration of nickel related to filling up of interstitial by nickel atoms. P 1 A

Lattice constants of Cdl_ type chalccgenides exhibit a characteristic feature. One of the body diagonal length of the hexagonal lattice is kept nearly constant over the entire homogeneity range. Nakahira and Kayashi (4) correlated these crystallographic parameters and the heat of formation of transition metal dichalcogenides with octahedral coordination.

(B) Characterization of thin film phases of ^ In the Ni-Te system some interesting results were observed

for NiTe2 composition (5). Therefore this phase was investigated in detail and results are reported. Electron diffraction pattern of the thin film in as deposited condition resembles the known bulk phase with lattice parameters a = 0.39 nm and c = 0.53 nm. In order to investigate possible structural transformations, the as-deposited NiTe^ films were annealed in-situ in the electron microscope under the vacuum prevalent in the microscope column ( ~10 Torr). Annealing treatment at a temperature around 473K, resulted in the growth of single crystal regions. Structural characterization of these regions revealed the presence of a new cubic phase with a = 0.74 nm which is different from the initial phase. This phase has no simple relationship with the bulk Cdl« type phase. Further annealing treatments resulted in the growth of large single crystal regions containing planar faults. Structural characterization of these regions indicated the presence of a long period phase with a=b=0.39 nm and c»1.16 nm. This value of 'c* is modulated and corresponds to a 2 H or 6 R polytyplc modification depending on hexagonal or rhombohedral structure. When the same films were further annealed at the same temperature fox long time, some striking changes showed up resulting in a highly disordered structure. These disordered regions contain a high density of random stacking faults.

1 (1) T.F. Conally, 'Solid state physics literature guide , Vol.3,, IFI/Plenum, New York (1972) (2) M. Ettenberg, K.L. Koroarek and E. Miller, J. Solid Stat.Chem., 1 (1970) 583 (3) J. Flahaut, MTP International Review of Science, Vol. 1C (Solid State Chem), Ed. L.E.J. Roberts, p.149 P 1 A 4

(4) M. Nakahlra and K. Hayashi, Mat. Re*. Bull. 13 (1978) 1403 |5) M. Singh and S.Bhan, Thin Solid Films (in press)

Professor Suraj Bhan, Department of Metallurgical Engineering, Banaras Hindu University, Varanasl 221 005 (INDIA) P 1 A 5

THREE-NUCLEAR Ta-CLOSTERS IN THE COMMENSURATE CDW-STATES OF

2H-TaS2 AND 2H-TaSe2

T.Blitz*' **, A.Lerf++, and S.Saibene* +Physik-Department, Technische Universitat Mttnchen, 8046 Garching

++Walther-MeiBner-Institut fur Tieftemperaturforschung der Bayeri- schen Akademie der Wissenschaften, 8046 Garching, FRG

The layered transition metal compounds 2H-TaS_ and 2H-TaSe2 undergo structural phase transitions at low temperatures which are commonly associated with charge density wave formation (CDW)[1]. The conduction electrons are considered the main source of instability. 2H-TaSe2 forms a commensurate 3x3 superstructure below

85K [2]. 2H-TaS2 exhibits a phase transition at 78K. Up to now the detailed sequence of CDW phases in 2H-TaS2 is unknown. It is unclear whether 2H-TaS2 also forms a commensurate 3x3 superstructure at low temperatures or not [3/4]. We report on a re-investigation of 2H-TaS, and 2H-TaSe, by means 181 of time dependent perturbed angular correlation (TDPAC) on Ta. We measured the symmetry and strength of the electric field gradient tensor at the inequivalent Ta sites at 10K including their 181 relative population. The 482 keV state of Ta has a nuclear spin 1=5/2. This level splits into +_5/2, +3/2, and ^1/2 sublevels under the influence of an extranuclear electric field gradient (EFG). By TDPAC precession frequencies corresponding to the energy differences E±3/2-E±1/2= Yuv E±5/2-Ei3/2=foi2, and E^-E^-Jfo^ with

(ijjsu^j can be observed. The position of u>1 depends on both the value of V (=largest component of EFG tensor in the principal coordinate system) and the asymmetry parameter n= |(V„ -V )/V | The asymmetry parameter can be most easily determined from the ratio WJ/UJ- or OJ3/CJ2. Previous studies [5] were performed with Nal(Tl) scintillation crystals as Y-detectors. The time resolution (^2.5 nsec FWHM) was insufficient to observe tu3 in the 2-3 GRad/s range. Furthermore, the frequency resolution was too low to allow for an unambiguous determination of the number of inequivalent sites and their rela- f 1 A 5

29M« P TAS2BK a-€.l•.84 i«K-16B«MV BAF2 KR (34328) 0 U 1 1 E 1 B R 18756 1

S 'A •• P '/I ' E 125M C T A R U M €2» - \ A « iAA^Ayv^*v 1/uuWV iAAA 2vW WvV ( NRAD/SEC )

Fig.l Fourier transformed TDPAC spectra showing the nuclear guadru- pole precession frequencies in 2H-TaS2 at 10K.

tive population. The use of fast BaF, scintillators, which produce ultrafast light pulses in the ultraviolet region, provides excellent timing (time resolution ^500 psec FWHM for the 133 keV - 482 181 keV cascade in Ta) combined with good detection efficiency and resonable energy resolution. This allowed for the first time the observation of all precession frequencies. A Fourier transformed TDPAC spectrum for 2H-TaS_ at 10K is shown in fig.1. The observed peaks can be unambigupusly assigned to three sites A,B, and C with roughly equal population. According to these data the CDW in 2H-

TaS2 consists of a commensurate 3x3 superstructure. From the 1:1:1 (or 3:3:3) population we conclude that the CDW has orthorhombic symmetry [7] and not hexagonal symmetry (this would lead to 6:2:1).

A preliminary analysis shows that the asymmetry parameters are nA= 0.38(3), ri =0.24(4), and ru=0.11 (6) . We note that V* is lower B C ZZ than V (no CDW), obtained by linear extrapolation from the high temperature region. Furthermore, the largest component of the EFG tensor is found to be perpendicular to the layers as in the absence of the CDW. From these facts it follows that the perturbing charge density consists of electron density in the x-y plane (=layer), sharply localized at the center of a Ta-triangle. A picture of this three-nuclear Ta-cluster is shown in fig.2. The near equivalence of the sites B and C (nearest and next nearest neighbours to the A- sites) and their relatively small ru „ requires a sharp localiza- P 1 A 5 tion of electron density. Basically the same observations were made for 2H-TaSe2 with the follo- B B wing details: nA=0.64(2), nB c*0...0.15, and the B- and C-sites are slightly more inequivalent. This means that the total electron density localized at the center of the 9-atom "snow- flake" is larger in

2H-TaSe_ than in 2H-TaS2. Fig.2 "Snow-flake" 9-atom Ta-cluster in the commensurate CDW states In conclusion we postu- of 2H-TaS2 and 2H-TaSe_. late three-nuclear Ta- clusters with a 3x3 superlattice lattice for the commensurate CDW state in both 2H-TaS<- and 2H-TaSe_. Such clusters were previously discussed for /3x/3 and 2x2 superstructures in hexagonal layers [6J. A harmonic CDW picture appears to be completely inadequate to describe the sharp localization. Thus it appears questionable whether the conduction electrons are the main source of instability.

[1] J.A.Wilson, F.J.DiSalvo, and S.Mahajan, Adv.Phys.24(1975)117 [2] R.M.Fleming, D.E.Moncton, D.B.McWhan, and F.J.DiSalvo, Phys. Rev.Lett.4J5 (1980) 576 [3] J.P.Tidman, O.Singh, A.E.Curzon, and R.F.Frindt, Phil.Mag.30 (1975)1191 [4] G.A.Scholz, O.Singh, R.F.Frindt, and A.E.Curzon, Sol.State Comm. 4.4 (1982) 1455 [5] T.Butz, A.Hiibler, A.Lerf, and W.Biberacher, Mat.Res.Bull. 16 (1981)541 [6] C.Haas in:"Physics of Intercalation Compounds", L.Pietronero and E.Tosatti eds.,Springer, Berlin:1981, p. 158 [7] M.B.Walker and A.E.Jacobs, Phys.Rev.B25(1982)4856 P 1 A 6

ELECTRODEPOSITED TUNGSTEN SELENIDE FlUSt STRUCTURAL; OPTICAL AND ELECTRICAL CHARACTERIZATION.

S* Chandra and S*N Department of Physics* Banar&s Hindu university varanasi - 221 005* India

Tungsten diselenide (WSe2> is a potential candidate material for photoelectrochemical solar cells because of it's favourable band gap ( f~l*6 eV) and stability against photocorrosion (1/ 2)• However/ for fabrication of devices like solar cells* large areas are needed* The most convenient method to obtain large area is to use polycrystalline materials in thin film form* preparation of WSe_ in thin film form poses many problems (2/ 3) and no studies have heen so far reported for MSe, film* The present investigation reports for the first time a simple electrodeposition technique for obtaining WSe_ film and their characterization. Tungsten selenide films have been electrodeposited cathodica- lly on Titanium substrate from an aqueous electrolyte consisting of H2W04+Se0 +KH 0H« Initially, 5.682 gm of H2W0 was dissolved in 100 ml of 10»e NNH.OH and 40 mg of seo was dissolved in 100 ml of water* The electrolyte in the electro- lytic cell consisted of their respective solutions in the volume ratio It 10. Reasonably good films of WSe. were obtained at 60 C with electrolysis current density 6mA /cm and 15-20 minutes of electrolysis time. We have found that SeO2 concentration is the rate determining factor for VSe- film deposition. The first deposition is always Se which is then followed by W deposition. Once* vf-rich layer is deposited a fresh layer of Se is formed which starts co-depositing vt« This deposition cycle was repeating itself* At certain electrolysis time a layer of right stoichiometry WSe, was obtained. The above observations have been confirmed with EOAX analysis (fig* 1)• The early deposition of se in comparision to W can be attributed to a higher value of e lee troneg a tiv ity for Se (=2*4} thenW(=l*7)* P 1 A 6

««»•». -.••» *.-•« i (1 I I]

0>PMiti«n tiiM,t

Fig. It The at* wt. % of Se (dotted curve) and W (solid curve) with time of deposition of film as derived from EDAX analysis* The shaded portion gives approximate stoichiometry WSe • The mechanism of the film deposition can be described with the following reaction*

-4- TJW OW * ( Mil \ W C\ H 2H2O

W0|-+ (6-n)e tfn+ oxide/hydroxide

seo, + H2O * + OH"

HSe| +2e~ Se(S) 2 + 1/2 H2

Se(s) + 2H20 3H H+ + e(y.se) H(Se)y ,n+ oxide/hydroxide + nH(y.Se) Where n is the valence state of W/ Se(s) is the elemental selenium deposited on the cathode (Ti)j H(se) is the atonic hydrogen held, on Se< x and y are the number of Tungsten and selenium ions deposited on Titanium* Thus* it appears that the atomic hydrogen held on Se is responsible for induced co- deposition of W* Optical absorption studies were also carried out by a cary-14 spectrophotometer to determine the band gap* Further/ the space charge capacitance, Csc/ of the VSe^/2T~t I electro- P 1 A 6 lyte system has been measured at IK Hz to characterize the interface. C depends on the applied bias. V as (Mott- Schottky relation) t 1/0 Ic * C^G, eND)) (v - V£b - kT/e)

Where v£h is the flat-band potential. The plot o£ l/CgC Vs v gives the values of ND# V£b etc. Same of the important parameters obtained from the above studies on a 1.7 /urn thick film are summarized below*-

carrier type n- Band gap, E<_(eV) ^1*0/ Indirect 3 Donor concentration/ ND(cm- ) 1.6 x 3 15 Trap density, Nfc (cm" ) 1.2 x 10 Density of states in the conduction 2«5 x band* % (cm"3) Redox potential of the electrolyte. 0*295 / () 2 SCE

Flat-band potential V£b (VSCE) -0.19

Band bending, Vb(v) 0.48 Depletion width, W( /urn) 0.02

Conduction band edge, EC(V_CE) -0.27

valence band edge, Ev (VSCE) 0.70 Referencesj 1. J. Gobrecht, H« Tributsch and H. Gerischer» J. Electro- chem Soc 125, 2085 (1978). 2* S> Chandra and S»N. sahuj J* Phys* D> Appl. Phys*, _17_ 198 4 (in press) . 3. G. Djemal, N- Muller, U. Lachlsch and D. Cahenj Solar Energy Material 5,, 403 (1981). P 1 A 7

ELECTRIC C0NDUCTT7TTT OP AgCrTiSj^ ABD Z.Cybulski Tecfan. Coll.,75-620 Koszalin, Poland

The first investigations on the synthesis of type \gX B^ S, four-component chalcogenides »»ere conducted by Hahn and al. (i\ They determined the structure type and lattice parametars for a number of compounds in that group. Further work on the synthesis of four-component chalcogonides, particularly with the use of transition metals, was carried on by Cybulski (2,3) • A number of theso compounds are semiconductors of low electric conductivity activation energies (k). Xnv?«?tiffations and results. Both compounds, AgCrTiS. and AgCrZrS, , were synthesized, as previously, from pure elements by heating them in sealed ampules ('0""Tr) (2). Preparations were obtained in powder form. This powder was subsequently pressed into tablets under 3-10 n/m" prassure. The tablets obtained wore polished and washed in hot acetone. On the specimens thus prepared, temperature-dependent electric conductivity measurements were carried out over the ran ge from 160 to ^26 K. Electric conductivity was measured by us a modified four-point square probe instead of a raoro connnonly usad liiioar one. The temperature-dependent conductivity determined by measurements yields a curve which is characteristic for semiconductors (Fig.1J,

4 AgCrlrS* 10 19 30 40 30 60 70

, 1. Dependence of the electrical conductivity as a function of temperature. P 1 A 7 Three sections are distinctly discernible on that curve,i.e. those relating to high, medium and lov temperatures. The conductivity activation energies calculated Tor tbe particular temperature ranges assume the following values:

T (K) Compound 160-193 0.026 193-288 0.0G8 AeCrTiS^ 288-378 0.185 378-h26 0.468 165-185 0.039 185-2^9 0.066 AjjCrZrS. 2k9-3k$ 0.130 372-^15 0.390

(1) G.Stricl:, G.Eulcnbereer, 357 [h-6] ,33S (1968) . (2) Z.Cybulski, Zosz.Kaul^.VSI Koszalin 1/2/76 Ji jlTllJ, 12. f3) Z.Cybulsl:!, II Srod.EonT. Chen. TJj: Poznru'i 197S, p. 123. [k) Z.Cybulski, J.Strzelecka, III Srod.ironT.Chea. T..V Toznajl 1983, p. 26.

I>r Z.Cybulslti FL-75-620 Kosaalin 1 A 8

STRUCTURAL AND MAGNETIC STUDY OF THE Cr^-Ga^ SYSTEM. L. Gastaldi and S. Viticoli I.T.S.E. - C.N.R. Area della Ricerca di Roma C.P. 10 00016 Monterotondo Scalo (ITALY) J. Flahaut, M. Guittard, A. Tomas and M. Wintenberger Laboratoire de Chimie Minérale Structurale - Associé CNRS N" 200 Faculté des Sciences Pharmaceutiques et Biologiques de Paris- Luxembourg, 4, Avenue de l'Observatoire, 75270 Paris Cedex 06 (FRANCE)

The ternary compounds of AB-X. formula where A = Zn, Cd, Hg ; B = Al, Ga, In ; X ~ 0, S, Se, Te have three possible structures : cubic spinel structure, tetragonal defective ZnS structure and layered ZnIn_S. structure. In this communication we discuss the X-Ray and magnetic data of the layered compound having the formula CrGa- cgS.. Single crystals are obtained by heating Cr-S, and Ga_S, in stoichiometric amounts with iodine in a sealed quartz ampoule at T = 1000°C. X-Ray powder diagrams show a solid solution bet-

ween CrGa. ggS, and CrQ ,gGa2S.. The crystal structure of

CrGa1 6gS4 (a = 3.606 A, c = 11.972 Â, space group P3m1 or P3m1, Z = 1) is close to FeGa^S.IT (1);Cr ions are surrounded by six nearest neighbours forming a triangular lattice perpendicular to the ternary axis. The magnetic susceptibility of CrGa.. ggS. has been investigated in the temperature range 4.2-300K and indicates for 6OK < T < 295K a predominance of ferromagnetic interactions (J/k = 4.IK). For T < 6OK the curve shows antiferromagnetic behaviour At T = 10.3K a maximum of x susceptibility has been observed.

(1) L. Dogguy-Smiri, Nguyen Huy Dung and M.P. Pardo Mat. Res. Bull., 15, 861-866 (1980)

Mailing adress of author : Mr. J. Flahaut Laboratoire de Chimie Minérale Structurale 4 Av. de l'Observatoire 75270 PARIS Cedex 06 France P 1 A 9

PHYSICAL PROPERTY CHANGES OF (T,M)1+xNb3_x WITH T = Fe, Cr AND M=Nb, Ti H. Gruber Institut fiir Festkbrperphysi k, Technical University, Petersgasse 16 A 8010 Graz

In these quasi-one-dimensional compounds the trigonal prismatic

chains of NbSe3 alternate with zig-zag chains of octahedra with octahedral bonded transition metal atoms to the anions {Fi g. 1) {1,2,3).

Fe, Nb,_ Se-|0 shows a broad metal insulator transition due to the forming a incommensurate charge density wave (CDW) at 140 K with q = {0, 0.27, 0). The CDW vector is only slightly different from that of the high temperature CDW in NbSe,. Recently the first measurements of the physical properties have been reported (4, 5, 6). For Nb Se tne are on1 (T.M) ]+x 3_x ] o y y partially known because they dep-end sensitively on the transition metal ions in the octahedral chains. Therefore the influence of their preparation and the relating changes of their physical properties with a different Fe, Ti , Cr and Nb content in the octahedra chains have been investigated in this paper. The single crystals were prepared by heating the mixed pow- FIG. 1 Structure of FeNb3Se10 ders of the sintered alloys from theme tallic components and Se, which were sealed in high evacuated quartz tubes. 10' A small temperature gradient at 973 K was maintained for three weeks. The stoichiometry of the crystal specimens was analysed by an X-ray microprobe. The crystal structure was determined by X-ray powder diffraction. Fig. 1 shows

the monoklinic structure of FeNb,Se,0.

Cr Nb Se as we11 as Fe Nb i c 9 a in iJ.v Q v^n exhibit a resistivity variation of more than 7 orders of magnitude (7). The re- 10 f , 100 1000 TfK] sistivity variation with a lower Cr-con- FIG. 2 Resistivity ratio tent is remarkably smaller than for the versus temperature P 1 A 9

compounds containing Fe (Fig. 2). The phase range for the compounds

with Cr extends from CrNb3Se10 to.Cr^ gNb2 4Se1Q. Fig. 3 shows 1rhe magnetic susceptibility which in the compounds containing Fe slowly increases with increasing Fe-content.

100 200 300 400 500 T[K| (SO 550 FIG. 3 Temperature dependence FIG. 4 Temperature dependence of the magnetic susceptibility of the inverse magnetic susceptibility At low temperature a Curie contribution to appears and can befit-

ted for T<100 K with the Curie-Weiss relation Z=C /(T+0) + ^o [lj .

C =Curie constant, 0=Weiss temperature, *o=temperature independent term. The compounds Cr, Nb^ Se,Q show a weak paramagnetic behaviour with increasing x in the measured temperature ranges (Fig.5). The compounds containing Ti can be fitted at T<100 K according [l], (Fig. 5). Fig. 6 shows the inverse magnetic susceptibility of the compounds containing Ti . The values of the magnetic susceptibility

ISO . .250 T|k] FIG. 5 Temperature dependence FIG. 6 Temperature dependence of the magnetic susceptibility of the inverse magnetic susceptibility Cr Nb Se show a vepy weak measured at 12.1 kb'e for ]+x 3_x To decrease from 77 to 170 K. In Fig. 6 the variation of the absolute thermopower for different quasi one-dimensional compounds is shown. 9 Nb Se in the wnole For the compounds ^ 7+x 3_x ]o temperature range a P 1 A 9

positive thermopower has been -found. The positive values increase with decreasing temperature down to 20K, falling then down strongly till 5K. In this region a large peak in the internal friction at 30 K was found (8). Cr, ojNb? g^Se,^ however, at higher temperature has a small thermopower and a change at low temperature to higher negative values. A possible explanation is that this compound has a comparable number of electrons and holes and their carrier mobilities have FIG. 7 Temperature dependence similar values. of the absolute thermopower The Mossbauer measurements (5,6,8) on the compounds containing Fe indicate that the strong distortion of the charge density of Fe caused by surrounding Nb atoms is related to the large differences in Q for different environments and the general decrease of Q with the number of Fe neighbours (9}.From the isomer shift at 170 K for the different types of Fe sites a low spin state Fe2 + may be indicated.

References (1) S.J. Hillenius and R.V. Coleman, Phys. Rev. B 2_5, 2191 (1982) (2) A. Meerschaut, P. Gressier, L. Guemas and J. Rouxel, J. Mat. Res. Bull. 16, 1035 (1981) (3) H. Gruber, W. SitteTnd H. Sassik, J. Physique 45, 1231 (1984) (4) S.J. Hillenius, R.V. Coleman, R.M. Fleming and TTTJ. Cava, Phys. Rev. B 23, 1567 (1981 ) (5) R.J. Cava, F.T7 Disalvo, M. Eibschiitz, J. V.Waszezak, Physy . Rev. B 27, 7412 (1983) (6) H. Gruber, M.TTei ssner and W. Steiner, Proceedings The Physics and Chemistry of low dimensional Synthetic Metals, Abano Terme 1984, J. Molecular Cryst. (1984), in print (7) A. Meerschaut, A.Ben Salem, L. Guemas, J. Rouxel, P. Monceau and H. Salva, Proceedings Synthetic low dimensional Conductors and Superconductors, J. Physique Coiioq 4444, C3-168C316811 (1983) (8) J.W. Brill, P. Booichand and G.H. Lemon Solid State Comm., 51, 9 (1984) (9) FT Gruber, E. Bauer, M. Reissner and W. Steiner, International Conference on Charge Density Waves in Solids, Sept. 1984, Budapest, Hungary P 1 A lo

MAGNETIC PROPERTIES OF NON-STOICHIOMETRIC TUe,Se L.Baggstrom, H.&. 7erma and S. Bjarman Institute of Physics, Uppsala University, Box 530, S-751 21 Uppsala

R. Berger Laboratory of Inorganic Chemistry. Materials Science Centre of the University, Nijenborgh 16, NL- 9747 AG Groningen

The compound TlFe.Se., as well as a few other ternary thallium 3d transition- metal chalcogenides of the corresponding stoichiometry, were reported to crystallize with the ThCr-Si type structure (1). In this structure, the 3d metal atom coordinates the non-metal atom in a tetrahedral arrangement, the tetrahedra being interconnected to form an infinite network in two dimensions. The third kind of atom, in this case thallium, links the slabs of filled Cetra- hedra together. In contrast to most representatives of this structure type, these thallium sulfides and selenides have a large c/a ratio of the conventional body-centred cell, the thallium forming bonds only with the chalcogen. In that way, these representatives may be considered as layer compounds, a fact which is reflected in the crystal habit. The structure is depicted in Fig. 1. The separation between the 3d atoms in adjacent layers is rather large, and magnetic interactions between the d-electron spins might therefore be re- stricted to these layers unless a long-range type of interaction can occur by means of itinerant charge carriers. That seems to be the case for the solid solutions TlCu, Fe Se_, 0< x <0.5, of the same structure type where ferromagnetic type of interactions occur even for moderate amounts of iron (2). In this concentration range the compounds are metallic, since the p-type broad- band character of the parent compound TlCu_Se_ prevails (3). According to preliminary electrical transport measurements, the resistivity is rather high for the iron compound so we would expect magnetic interactions of a two- dimensional character. It proved difficult to synthesize well defined material of the stoichiometric composition. We could recover unreacted iron from the synthesis product. The same difficulties were encountered for the sulfur analogue (4}. X-ray' diffraction analysis showed that there were additional lines that could not be indexed on the tetragonal cell reported by Klepp and Boiler (1). However,

On leave from PhysicsDepartment, Punjabi University, Patiala, India P 1 A lo

all lines were accounted for by choosing an a-parameter /5 tines larger. The occurrence of this "supercell" is an indication that vacancies may occur. Vie therefore be live that our product is iron deficient corresponding to the formula TLFe. Se», where x^ 0.4 according to a preliminary chemical analysis. The situation might be similar to the corresponding sulfur system where an orthorhombic supercell occurs for TIFe. JS^- The ThCr_Si» type cell may there be Obtained on heating (5). Mossbauer spectra were recorded in the temperature range 100 - 460 K by means of a conventional constant-acceleration spectrometer with a s7Co(Bh) source kept at room temperature. The target was a mosaic of single crystals mounted on an aluminium foil. The flaky crystals had an average thickness of 50 ym and were oriented with the y-ray beam perpendicular to the flakes. We found a first-order magnetic transition at about 450 K with some hysteresis effects. Below the transition temperature, a principal four-line magnetic pattern was obtained together with a weak multiplet (Fig. 2). On basis

458 K

V.,

V* T 295 K

1 I ' 4.0 *.a -i. r -4. o -a, a a. 2.0 o—o v (enm/s)

Fig. 2. Mossbauer spectra from above and below the transition temperature. H Se Cu.Fe

Fig, t. The ThCr.Si. type structure. The a-vector of the supercell may be obtained from sunning a + 2b. P 1 A lo

of the orientation of the target with respect to the beam, we conclude that the iron spins in the ordered state are oriented along the c-direction. Since the material is not attracted by a permanent magnet, we assign an antiferromagnetic ordering of the spins below 450 K, such that the compound must be considered a 2-dixoensional Is ing antiferromagnet or very close to it. From a detailed analysis of the spectra above and below the transition temperature we deduced both the sign of 7 and the orientation of the electrical field gradient tensor with respect to the structure (6). The quadrupole splitting, eQV /2 , is -0.48(2) nnn/s at room temperature, and V and V 22 22 XA are situated in the {OOl} plane and V along <001>. We note an increase in isomer shift as compared with the "isostructural" TICuFeSe-, from 0.47(1) mm/s (2) to 0.55(1) mm/s in our compound. There is also a marked change in the quadrupole splitting with eQV /2* +0.96(1) aan/s for the copper-containing compound. TICuFeSe, is a semiconductor which favours the hypothesis of high-spin divalent iron. The change in parameters is an indication that the d-alectrons get more localized and that the electron cloud is more spherical in TlFe_ Se_ as compared with TlCuFeSe_.

(0 K. Klepp and H. Boiler, Monatsh. Chea. _109_ (1978) 1049 (2) R. Berger and C.F. van Bruggen, University of Groningen. To be published. (3) R. Berger and C.F. van Bruggen, J. Less-Common Met. 99 (1984) 113 (4) M. Zabel and K.-J. Range, Z. Naturforsch. 2£b (1979)~T (5) M. Zabel and K.-J- Range, Rev. Chim. miner. V7. (1980) 561 (6) G. Caer, J.M. Dubois, L. Haggstrom and T. Ericsson, Nucl. Instr. Meth. 157 (1978) 127 P 1 A 11

NEW TERNARY CHALCOGENIDES OF THE COINAGE METALS WITH THALLIUM(I) OR ALKALI METALS Kurt O. Klepp Institut ftir Anorganische Chemie der Technischen Hochschule Aachen, Professor Pirlet-StraBe 1, D-5100 AACHEN (F.R.G.)

Investigations of the ternary systems A/M/X (A=Na, K, Tl; M=Cu, Ag.Au; X= S,Se, Te) have led to the detection of a number of intermediate compounds with A/M-ratios ranging from 3:1 to about 1:10. Their crystal structures, many of them representing new structural types, have been determined by single crystal techniques. They are in most cases characterized by short M-M-contacts. According to the topology of the M-X-partial structures these compounds may be subdivided into:

A) Compounds containing isolated groups: Na,AgS2, Na,AuS, In both compounds the coinage metal is in linear coordination by the chalcogen atoms, forming [ S-M-S] complex anions. Similar anionic

groups are likely to exist in Tl,AgS2> Tl,AgSe2 and Tl_AuSe2 which are obtained as glass-like products by preparation from the melt. B) Compounds with layered M-X-partial structures a) Planar trigonal layers are formed in KAgTe b) Layers formed by MX.-tetrahedra sharing edges and apices are found

in Tl3AgTe2 c) Layers formed by edge sharing tetrahedra: NaAgTe, TICuX (X=S,Se)

TlCu2X2 (X=Se,Te) (single layers); TlCu4X3 (X=S,Se) (double layers) d) Complex layers: TlCu,S2 C) Channel type compounds with three-dimensional M-X-frameworks a) The structure of TIAgX (X=S,Se,Te) can be visualized as formed by chains of edge-sharing MX.-tetrahedra which are further connected via common X-atoms to form a framework with narrow channels occupied by Tl. b) Similar tetrahedral chains, but a somewhat different type of connection is found in TlAg, X, (X=S,Se,Te). Two channels with different geome- tries are formed. The one shows trigonal-prismatic sites occupied by Tl, the other is hosting a chain of face sharing Ag-octahedra which are partially centered by additional chalcogen atoms. The structure of PI All

Table 1 : CRYSTAL DATA Compound s.g. a (A) b (A) c (A) Z; Type Ref. a (°) T12A820-xTell R3c 11 . 436(4) 41 •98(1) 6 new P6 /m 10 . 600(4) 4 . 240(1) 2 new T1A«6-x?3+y 3

T1A Se P63/m 10 .951(2) 4 . 384(1) 2 new *6-x 3+y

TlAgS, Te,x P63/m 11 .423(3) 4.617(2) 2 new 6-X 3+y

Tl7Ag36Te22 Fm3m 18 .731(2) 4 new

TlCu4S3 P4/mmm 3,.894(1) 9.33(1) 1 KCu4S3 (1)

TlCu4Se3 P4/mmm 3.. 974(4) 9,. 84(2) 1 KCu4S3 (1) 14.,63(1) 3.863(1) 8.. 298(4) TlCu3S2 C2/m 4 : & 111 • 72(4lt'Yt}l TlCu3Te2 P4,/nnm 8. 427(4) 14.'192(6) 8 new

TlAg3S2 Pbcn 8. 154(5) 8. 792(6) 7. 029(4) 4 new (3)

T1A S P21/m 15. 811(5) 4.095(1) 21. 907(3) 4) H) §3-X 2 1 0^.87(3) T1A Se P63/m 15. 119(5) 4. 408(2) 6 new §3-X 2+y TICuS P4/nmm 3. 913(2) 8. 16(1) 2 PbFCl TICuSe P4/nmm 4. 082(1) 8. 164(2) 2 PbFCl NaAgTe P4/nmm 4. 515(3) 7. 458(4) 2 PbFCl

KAgTe P63/mmc 4. 764(1) 9. 488(1) 2 Ni2ln

TIAgS Pnma 7. 228(3) 4..466(1) 8. 331(2) 4 NiTiSi (4) TIAgSe Pnma 7. 476(1) 4. 637(1) 8. 390(1) 4 NiTiSi (4) TIAgTe Pnma 7. 759(1) 4. 868(1) 8. 773(2) 4 NiTiSi (4)

Na3AgS2 Ibam 6. 380(2) 12. 581(3) 6. 896(4) 4 Na3AgO2 (5)

Na3AuS2 R3c 7. 623(2) 16. 672(5) 6 new (6)

Tl3AgTe2 P21/c 11. 002(5) 7. 416(2) 9. 875(3) 4 new 116. 62(4) ) determination of the crystal structure not completed P 1 A 11 TlAg, Se, is closely related. Here the tetrahedral single chains are replaced by double chains of edge-sharing AgSe.-tetrahedra. c) Compounds based on a packing of 343 4-nets of chalcogen atoms: TIAg-S,, TlCu^Te- (Fig. 1). Both structures show quadratic antiprismatic channels which host the Tl-atoms. The M-X-frameworks are built up by distorted stellae quadrangulae sharing edges and faces. D) Complex tetrahedral frameworks

a) Tl_Ag,/Te22: The Te-atoms form centered Friauf-polyhedra sharing common edges. Ag-atoms occupy tetrahedral sites within and between the Te- polyhedra. A system of interpenetrating channels, hosting the Tl-atoms is formed

b) Tl2Ag2Q Te, ,: The crystal structure is formed by slabs similar to those of TlAg, Te_ alternating with slabs exclusively built up by AgTe.- tetrahedra.

A detailed description of the crystal structures and their relationships will be given on the meeting.

(1) K. O. Klepp, H. Boiler and H. Vollenkle, Monatsh. Chem. m_, 727 (1980) (2) K. O. Klepp and K. Yvon, Acta Cryst. B_3£, 2389 (1980) (3) K. O. Klepp, J. Less-Common Met. , in the print (4) K. O. Klepp, Monatsh. Chem. IU_, 1433(1980) (5) K. O. Klepp and W. Bronger, J. Less-Common Met. , in the print (6) K. O. Klepp and W. Bronger, J. Less-Common Met. , in the print

Fig. 1 : The crystal structure of TlCu,Te2 as seen in projection along [001] (the occupancy for the Cu(4)-position is 0. 75)

Dr. Kurt O. Klepp Institut fur Anorganische Chemie der Technischen Hochschule Aachen, Professor-Pirlet-Strafie 1, D-5100 Aachen (BRD) P 1 A 12

FAR-INFRARED AND X-RAY INVESTIGATIONS ON MIXED TRANSITION METAL DICHALCOGENIDES WITH SEMICONDUCTING/METALLIC BEHAVIOR

G. Kliche Max-Planck-Institut fiir FestkSrperforschung 7000 Stuttgart 80, Heisenbergstr.1, West-Germany

The electrical properties of the transition metal dichalcogenides can be systematically varied by anionic or cationic substitution, semiconductor-metal transitions are e.g. possible

in the alloys FeS2-CoS2~NiS2, CoPS-CoS2, or PtS2-PtSe2-PtTe2. The metallization of the compounds is connected with increasing free carrier concentration, which can be determined by far- infrared reflection measurements.

In the cationic substituted mixed crystals, e.g. FeS2-CoS2 (1), metallization is caused by adding mobile electrons in a 3d: e band leading to strong but heavily damped plasma modes of free carriers in the far-infrared reflection spectra even at low excess-electron concentrations. On the other hand, [p,d)-band overlap was proposed to be responsible for the increase of free carrier concentration and for discontinuities in the lattice constants of the mixed anionic series PtS2~

PtSe2-PtTe2 (2).

In this work we studied the CdI7-type mixed crystals HfS, Te . We found, similar to the case of the platinum dichalcogenides, a change in the behavior of the lattice constants and in the x dependence of the plasma resonance frequency of the free carriers present near the composition HfSg 4Te1 ,, which may be critical for (p,d)-band overlap in the compounds under investigation. Optical phonon frequencies are given for the two-mode system HfS- ,/Te . L ~X X (1) E. Anastassakis and C.H. Perry, J. Chem. Phys. 6±, 3604 (1976) (2) G. Kliche, J. Solid State Chem. 1984, in press P 1 A 13

FAR-INFRARED REFLECTION SPECTRA OF PYRITE AND MARCASITE TYPE MANGANESE, IRON, AND PLATINUM GROUP CHALCIDES H.D. Lutz, 6. Schneider, G. Wfischenbach, and G. Kliche Laboratorium fQr Anorganische Chemie der Universitat.Adolf-Reichwein-Str., D 5900 Siegen

The far-infrared reflection spectra of hot-pressed samples of the pyrites MX, with M=Mn,Fe,Ru,Os and X=S,Se,Te (1) and of single crystals of the marcasite FeS2 are presented in the range from 40 to 700 cm . The spectra of the pyrites show five reststrahlen bands and more or less free carrier contribution due to deviation from stoichiometry. The oscillator parameters o>., p., y., w , y and the phonon frequencies WJ-Q, n+ and a _ are presented. The uncoupled longitudinal phonon frequencies U>,Q were determined from -Im(1/e) of the plasmon-free reflection spectra calculated from the oscillator parameters neglecting the free carrier contribution. The observed intensities of the reststrahlen bands (and also the TO/LO splittings) do not reflect the "inherent" intensities of the individual lattice modes, given by the oscillator strengths, because an intensity transfer from low to high-wavenumbered modes takes place.

From single crystal studies all 7 IR-allowed phonon frequencies

The effective ionic charges (Szigeti charges) reveal an increasing covalency of the pyrites in the order Fe>RuK)s>Mn compounds (pnictides>chalcides) and FtSg marcasite >FeS2 pyrite. Both the phonon frequencies and the force constants reflect the increasing strength of the metal-chalcogen bonds on going from 3d to 4d and 5d metal-compounds, discussed in former work (2).

(1) H.D. Lutz, G. Schneider, and G. Kliche, J. Phys. Chem. Solids (in press) (2) H.D. Lutz, G. Schneider, and G. Kliche, Phys. Chem. Minerals 9, 109 (1983) P 1 A 14

POLYSULFIDOAURATE(l) AND THIOAURATE(I): SYNTHESIS AND STRUCTURE 2 4 OF AuS9", Au2Sg " AND Au12Sg " G. Marbach and J. Strahle Insritut fur Anorganische Chemie dei Universitat, Auf der Morgenstelle 18, D-7400 Tubingen

Very little is know about gold sulfides and thioaurates. Au.S crystallizes in the cuprite structure; the structures of Au-S, and AuS are still unknown. References to the thioaurates(I) [AuS]~ and [AuS-] ~ as well, as to the tri- sulfidoaurate(l) [AuS,]" are to be found in older publications. In a systematic study of gold-sulfur compounds we have also attempted the synthesis of thio- and polysulfidoaurates and to elucidate their structure. Besides other thio- 2- and polysulfidoaurates we were able to isolate the novel anions AuSn", Au_Sg and AU.-SQ ~ and to determine their structure.

1. [Ph.As][AuSn]. (1) is formed as main product in the form of bright yellow crystalline platelets on reaction of K[Au(SCN)-] with tetraphenylarsonium polysulfide in absolute ethanol. The anion AUSQ" consists of a puckered AuSg ring (fig. l), in which a chain of nine sulfur atoms is bound as chelate ligand to the gold(I) atom.

Fig. 1. Structure of the nonasulfido- aurate(l)ion [AuS_]". Selected bond lengths [pm] and angles [ ] : Au-Sl 227.7(1), Au-S9 226.5(3), S-S 203.6(4) to 206.1(5); S7 S-S-S 105.9(2) to 108.2(2), S2 S-Au-S 176.0(1).

The shape of the AuS0 ring is essentially influenced by the streched S-Au-S group. The compound crystallizes in the triclinic space group Pi.

2. [Ph1As32[Au2Sg]. By mixing a solution of K[Au(SCN)2] or HAuCl4-4H,0 in ethanol with an aqueous solution of ammonium polysulfide a solution of polysulfidoaurate(I) is formed. Addition of Ph.AsCl yields [Ph.AsJ-EAu-S-] in form of yellow platelets. In the cyclic anion Au-Sg ~ two gold(l) atoms are linked 2 together by two S4 -cnains forming a twisted ten membered ring (fig. 2). P 1 A 14

Fig. 2. Structure of the polysulfidoaurate(l)ion [Au_Sg] ". 54 Selected bond lengths [pm] and angles [°]: Au-Sl 228.5(2), Au-S3 228.4(2), S-S 201.6(3) to 207.3(3), Au-Au1 312.4(1), S1-AU-S3 166.71(8), S-t-S 106.8(1) to 109.3(1).

As frequently observed in other gold(I) compounds a short Au-Au contact of 312.4 pm is present in Au_S» ~ as well, causing a remarkable bending of the SAuS- group. [Ph.AsJ-tAu-S-J crystallizes in the monoclinic space group C2/c. The compound is isotypic with [Ph.PJ-CAu-Sg] (2).

3. rPh.AsjJAu.^Sp]. (3) The gold sulfides Au-S and Au.S, as well as tetra- chloroaurates(lll) can be dissolved in concentrated aqueous solutions of Na-S; whereby the Au(III) compounds are reduced. Upon addition of Ph.AsCl the thioaurate(I) [Ph.As]4[AUj-S_] precipitates quantitatively as light yellow needles.

Au6 Fig. 3. Structure of the thioaurate(I) r T4- ion [Au12SgJ . Selected distances [pm] and angles [°]: Au-S 223.7(5) to 234.6(5), Au2 Au-Au 317.9(1) to 335.1(1); S-Au-S 177.8(2) to 179.4(2), Au-S-Au 86.7(2) to 93.1(2).

The anion [AU^SQ] ~ forms a cubane-I&e structure whose corners and edge midpoints are occupied by sulfur and gold atoms (fig. 3). Gold(I) thus achieves its preferred linear coordination; the S atoms each bridge three Au atoms. The Au atoms are arranged cluster-like in the form of an cuboctahedron.

[Ph4As]4[Au12Sg] crystallizes isotypic with [Ph4P]4[Cu12Sg] (4) in the monoclinic space group P 1 A 14

Praliminary results show that new compounds can be obtained by stepwise reduction of the anion Au.-Sg - The substitution of S atoms in the cyclic polysulfidoauratesO) by Au atoms may yield other species with ten or more atoms in the ring.

(1) G. Marbach, J. Strahle, Angew. Chem. 9£ (1984) 229; Angew. Chem. Int. Ed. Engi. 23 (1984) 246. (2) A. Muller, M. Romer, A. Bogge, E. Krickemeyer, K. Schmitz, Inorg. Chim. Acta 85 (1984) L39. (3) G. Marbach, T Strahle, Angew. Chem. 96 (1984) 695; Angew. Chem. Int. Ed. Engl. 23 (1984) ... (4) P. Betz, B. Krebs, G. Henkel, Angew. Chem. 96 (1984) 293; Angew. Chem. Int. Ed. Engl. 23 (1984) 311.

Prof. Dr. J. Strahle, DipL-Chem. G. Marbach, Institut fur Anorganische Chemie der Universitat, Auf der Morgenstelle 18, 0-7400 Tubingen 1 (FRG) P 1 A 15

METASTABLE PHASES OBSERVED IN THE SYSTEMS Ga^-MS WITH M = Mn AND Fe M.P. PARDO AND J. FLAHAUT Laboratoire de Chimie Minerale Structurale Associe au CNRS - LA 200, 4, Av. de 1'Observatoire, 75270 Paris, France.

The Ga_S3-MnS system forms in the Ga_S3-rich region, several tetrahedral phases (1). Some of these phases are stable in convenient regions of temperature and compositions (n = Mn/Mn + Ga at.)

- a solid solution of wurtzite-type for 0.02 v< n ^ 0.16, ana for temperatures upper than 900° C. - a solid solution of blende-type for 0.03 ^c n ,< 0.15, and for temperatures lower than 900° C.

- a non-stoichiometric phase of tetragonal CdGa2S.-type (designa- ted by

- a stoichiometric MnGa_S4 compound. At low temperature, it has a monoclinic structure of the MgGa_S4-type. At t = 965° c it undergoes a transition to a B-MnGa^S. compound which is a superstructure of the wurtzite and probably has the same structural type than the ZnAl2S4 compound. The structure of 6-MnGa-S, was described by Bakker (2) : it is closely related to the monoclinic structure of a-Ga2S.j (3) , by ordered substitution of Mn to the

Ga atoms and to some vacancies. The MnGa2S4 compound melts with a peritectic decomposition at 995° C. (Fig. 1) Metastable phases By fast cooling of the melts (quenching from 1100° C) from n = 0.02 to n = 0.33, two phases are obtained : - the wurtzite type solid solution for 0.02 ^ n^ 0.20

- the 6-MnGa2S4 phase, superstructure of the wurtzite, which probably has a small homogeneity range and is pure for 0.28 ^ n ^ 0.33. These two phases are in a metastable state at ordinary temperature. P 1 A 15

By slow heating, these 2 phases show the following behaviours. 1) from the wurtzite-type solid solution. It remains unaltered until about 400 - 450° C. At this temperature appear various superstructures of the wurtzite, which were called Q for n = 0.05 ; 2 for n = 0.14 ; 3 for n = 0.20 The formation of these phases is coming with liberation of heat (exothermic peak by DTA at about 430° C).

By subsequent heating, the preceding i>n phases are transformed into the stable phases which are characteristic of the phase diagram : blende type solid solution for 0.05 N< n ^ 0.14 and mixtures of this phase and aMnGa.S, for n = 0.20. The important feature of this second transformation is a second liberation of heat (second exothermic peak at about 630 - 690° C). This behaviour establishes the metastable character of the n (n = 0 to 3) phases. 2) from the 6-MnGa-S. phase. It remains unaltered until about 650° C. At this temperature, the stable phases of the phase diagram are formed (mixture of tetragonal CdGa2S«-type g phase and of a-MnGa-S.) . This formation is characterized by a liberation of heat (exothermic peak at about 630° C). 3) from mixtures of wurtzite-type solid solution and a-MnGa_S. phase (compositions 0.20 <: n s< 0.28). The heating of these mixtures forms two other metastable superstructures of the blende, called $4 and $5, which have a similar behaviour as the other n phases (n = 0 to 3) . (Fig. 2) Hysteresis phenomena By slow cooling of the melts (at 6°/min or 2°/min), the high- temperature phases of the diagram are maintened to the low temperature region. Moreover at 6°/min, the wurtzite-type solid solution is maintened at such a low temperature (= 600° C) that the formation of $3 superstructure can be observed. In conclusion, depending on the thermal treatment, various non equilibrium phase diagrams are possible. P 1 A 15

Fig. 1

1000

•f.h Fig. 2

(1) M.P. Pardo, P.H. Fourcroy and J. Flahaut, Mat. Res. Bull., 10, 665 (1975) (2) Bakker, Thèse de Doctorat, université de Leiden, Pays-Bas (1982) (3) G. Collin, J. Flahaut, M. Guittard and A.M. Loireau-Lozac•h Mat. Res. Bull., jn_, 285 (1976) Mailing adress of author : Mr. J. Flahaut Laboratoire de Chimie Minérale Structurale 4, Av. de l'Observatoire 75270 Paris Cedex 06 Pranro P 1 A 16

HYDRATED LAYERED PHASES M (B.O) CrS_ (M=» Alkali, Alkaline Earth Metal) DERIVED FROM 2H- AND 3R-K (H.O) CrS. BY TOPOTACTIC ION EXCHANGE

R. Quint and H. Boiler Institut fflr Physikalische Chemie der Ohiversitat Wien, Wahringerstrafle 42, A 1090 Wlen

H.Blaha

Institut fur Anorganische Chemie der Oniversitat Wien, Wahringeratrafle 42, A 1090 Wien

Introduction KCrS is easily oxidized to K (HO) CrS hiving a 3R-stacking order (1) In this phase belonging to the family of intercalated transition metal disulfides, potassium can be substituted by other alkali and alkaline atoms by ion exchange in aqueous solutions. These phases are described to have also 3R-structures with interlayer separations depending on the hydration of the intercalated ions. No structural details, however, are given (1) . Recently we described another polytype of K (H.O) CrS with a 2H-structure (2). Ion exchange in 2H-K (H-O) CrS. leads to a new series of phases having a two layer stacking sequence. We also reexamined the phases with three layer stacking sequences in order to elucidate more precisely their crystal structures and phase stabilities.

Experimental KCrS. was prepared by tempering a mixture of K.CO. and Cr powder (27g:lg) at 1000°C under H.S for 7-8 hours. Oxidation of KCrS, by oxygen in aqueous suspension yields K (H.O) CrS (xs*0.4j as described by Schailhorn (1). 2H-K (B.O) CrS. was the reaction product of polysulfide melts with K,CrO (2) after treatment with water. For ion exchange crystals or powdered samples of the potassium thio- chromitas ware treated with aqueous solutions of the respective alkali or alkaline earth ions and then examir.ad by X-rays (Deiy;e-Scherrer, rotating-crystal and Weissenberg method) and thermcj^<»vimetry. P 1 A 16

Results In 2H- and 3R-K (H,0) CrS K+ was exchanged against Li+, Na+, Bb+, Cs+, 2+ 2+ 2+ * ^2+ Mg , Ca / Sr , and Ba . The resulting phases have layer stacking sequences (2s- or 3s-) according to the starting material. Ions with lower hydratlon tendency (K , Rb , Cs> and Ba ) form hydrates having an interlayer separation approximately one water molecule thick (hydrates 1). The ions Li , Na , Mg , Ca and Sr form hydrates with interlayer separations corresponding to approximately bimolecular intercalated water layers (hydrates 2). These hydrates can be dehydrated at room temperature to hydrates 1. TGA showed that hydrates 1 lose their water between 50°C and 120°C either reversibly (Li , Na , K ) or irreversibly (alkaline earth ions).

O Q 0 O o o O \ o 6° 6 O o o o O1O 'O *O 2H(Ib) 20(1) 3R(Ib) 3R(IQ)

Fig. 1 : Schematic representations of the stacking in M (H_O) CrS. phases (Sections through the (110) plane of the hexagonal close packed layers. In the 20 structure the bold typed section is shifted by a_/4 above the paper plane) P 1 A 16

The crystal structures of hydrates 1 and hydrates 2 differ not only in the Intarlayer separation, but also in the lateral displacement of adjacent CrS 3labs: In hydrates 1 sulfur atom3 belonging to neighbouring slabs lie one above the other (3R(lb) and 2B(lb); Fig.t). This arrangement is not changed on dehydration. In hydrates 2 the corresponding sulfur atoms are in a staggered position. The lateral shift vector (in units of the hep mesh a ) H is (1/3, 2/3, 0) or (1/2,0,0) leading to a 3R(la) and a new orthorhombic 20 structure respectively (Fig.l). Because of the symmetry lowering in the 20-phase domain crystals are formed. The structural data of 20-Sr (H_0) CrS_ are: x 2 y 2 a= 3.33 8 b= 5.77 £(= a V*3) c= 22.80 X 18 Space group: Cmca - D

Atom positions: 4 Cr in 4a: 0,0,0

3S in 8f: 0,y,z; y= 1/3, z= 0.059

Sr and H_O could not be localized.

In addition a weak 8-fold superstructure in the (ab)-plane is observed.

The occurrence of the 20 structure with the quite uncommon shift vector of 1/2 a is explained by the symmetry principle (conservation of the H center of symmetry): A shift vector of (1/3,2/3) would lead to an acentric 2H-structure (P6,) with the sulfur atoms in split positions.

(1) R. Schollhorn, R. Arndt, and A. Kubny, J. Solid State Chem. 29^, 259 (1979) (2) R. Quint, H. Boiler, and H. Blaha, Monatsh. Chem. 115, 975 (1984)

Acknowledgement Thi3 work waa supported by the Fonds aur FSrderung der wissenschaftlichen Forschung, project P 4605. P 1 A 17

A HYPERFINE SPECTROSCOPIC STUDY OF POLYMORPHIC PHASE TRANSITIONS

IN TaS2 AND TaSe2

S.Saibene+, A.Lerf"*"*", and T.Butz+'++ +Physik-Department, Technische UniversitSt Miinchen, 8046 Garching ++Walther-MeiBner-Institut fur Tieftemperaturforschung der Bayeri- schen Akademie der Wissenschaften, 8046 Garching, PRG

The layer compounds TaS_ and TaSe- have the 2H-structure with trigonal prismatic metal coordination at low temperatures and the IT-structure with octahedral metal coordination above about 1200K. The 1T modification can be retained by rapid quenching from high temperatures to room temperature. At intermediate temperatures 4H, and 6R structures are formed with mixed metal coordination, i.e. the coexistence of both coordinations in alternating layers. A shear transformation of one chalcogen layer with respect to the metal layer is required to change the coordination from trigonal prismatic to octahedral. The system TaSe- was thoroughly investigated by in situ X-ray diffraction [1], contrary to TaS2- We have investigated the polymorphic phase transitions in TaS, 181 and TaSe2 via the Ta nuclear quadrupole interaction, measured by time differential perturbed angular correlation (TDPAC), for the following reasons: (i) for the preparation of monophase compounds and for high temperature intercalation reactions a detailed knowledge of the sequence of phases in TaS2 is necessary; (ii) practically nothing is known about the dynamics of the shear transformation in TaS2 and TaSe2; (iii) the nature of the mixed coordination phases is rather unclear. Between 500-600K a rather sluggish and irreversible transition from 1T-TaS2 to 2H-TaS2 involving a shear transformation takes place (see fig.1). in this regime we observe severe linebroadening.

The same transition is observed in TaSe2 between 400-450K. The dynamics of this transition is very slow and it appears plausible that the system foolows a seguence of non-equilibrium states. The 2H modification transforms irreversibly to a mixed coordination phase at 1020K and 1060K for TaS2 and TaSe2, respectively (see fig.s 1 and 2). In the case of TaS2 we observe a gradual loss P 1 A 17 of the static TDPAC signal already 40OK below the 2H- mixed coordination phase transition, indicative of sublattice melting. There is no bulk Ta diffusion but rather a rapid intar- conversion between the two adjacent Ta-coordinations (=confined diffu- TEMPERATURE sion) which is responsible 181, for the observed dynamics. Fig.1 Observed IO Ta nuclear quadru- This follows from the fact pole precession frequencies versus temperature in TaS2- The that a metal-rich compound, arrows indicate the heating heated up to 1000K in ex- and cooling cycle. cess sulfur for several days, does not attain stoichiometry. The sulfur- deficiency is cured as soon as the mixed coordination phase prevails. The full static TDPAC signal is re- covered not before the material has transformed to the 1T-modification. In the 411 It* case of TaSe2 no loss of TEMPERATURE the static TDPAC signal was 181. observed. This means that Fig 2 Observed lolTa nuclear quadru- the confined diffusion is pole precession frequencies versus temperatures in TaSe_. either too rapid (intercon- The arrows indicate the heating version rates >10 /sec) and cooling cycle. thus leading to motional averaging, or too slow and confined to a narrow temperature range. In this case a coexistence of the 2H modification and the mixed coordination should be observable. Our data do not allow an unam- bigous discrimination between both possibilities. A remarkable feature is the fact that we observe a unique, well defined Ta-site in the mixed coordination phase, despite the coexistence of two P 1 A 17

different metal coordinations, suggesting motional averaging. This would require massive diffusion perpendicular to the layers. At 1140K and 1120K an irreversible transformation to the 1T-modification takes place for TaS- and TaSe~» respectively. Upon slow cooling both materials transform back to the 2H-modification via an intermediate mixed coordination phase with a hysteresis of 150K

(TaSe,) and 200K (TaS2)• In this low temperature mixed coordination phase we observe substantial linebroadening which could be due to incomplete motional averaging. Since both the heating and cooling cycles for both materials were performed in evacuated ampoules some chalcogen loss might be anticipated. In the case of TaS- we indeed observed after re-transformation into 2H-TaS_ the formation of an additional new spectral component in the TDPAC spectra which we identify with a small fraction of self-intercalated material because it disappeared upon heating in sulfur excess at 1050K. No indication for Se-loss was observed after prolonged heating in an evacuated sealed ampoule (about two weeks). In addition to these phase sequences we investigated the irre- versibility of the 2H-mixed coordination and 1T-mixed coordination transitions. Two points are worth mentioning: (i) the low and high temperature mixed coordination phases can only approximately be interconverted; (ii) upon temperature reversal the system did not follow the same path within the mixed coordination phases. This strongly suggests that the mixed coordination phases, possibly stabilized by slight deviations from stoichiometry (the existence range of the high temperature mixed coordination phase under a large sulfur vapour pressure was less than 30K in TaS-!), are metastable longlived configurations. Regarding the fact that the data collection time at each temperature point was of the order of one day this means that equilibration requires several days or weeks. The X-ray characterization of the mixed coordination phases appears problematic in the light of our results: (i) in the high temperature existence regimes there is considerable diffusion which is difficult to study by in situ X-ray diffraction and (ii) the quenched mixed coordination phases are very likely severely disordered. [1] R.Huisman and F.Jellinek, J.Less.Comm.Met.J2M963) 111 P 1 B 1

INELASTIC NEUTRON SCATTERING AND LATTICE DYNAMICS OP HfSe , SnSe, and TiS

M. SchaTli and W. BUhrer Labor fUr Neutronenstreuung, ETH ZCIrich, 5303 WUrenlingen, Switzerland P. Levy Institut de Physique Appliquie, EPP Lausanne, Switzerland

HfSe_ and SnSe? belong to the well known class of transition-metal dichalcogenides (MCh~) showing a layered structure. They crystallize in the Cdl_ structure (spacegroup D^,) with one MCh_ molecule in the hexagonal unit cell. The layer character arises from weak interlayer interactions resulting, at least macroscopically, in rather anisotropic properties. The chalcogen ions do not occupy centers of inversion, and the induced static dipole moments are important for the understanding of the static and dynamic properties of these compounds. Thus, these compounds present the challenging feature of combining a simple chemical formula with a non-simple dynamical structure. Phonon dispersion curves have been determined by coherent inelastic scattering of thermal neutrons. HfSe- and SnSe :

HfSe- and SnSe_ are semiconductors with an indirect band-gap of the order of 1 eV. They show very similar lattice parameters, while the mass of the metal ion and the ionicity of the bonding is distinctly larger in HfSe_ as compared to SnSe_. The accoustic modes show a nearly identical shape for both compounds, supporting the fact, that the accoustic modes are mainly determined by ion-masses and lattice parameters. The typical features of the accoustic modes of layered crystals, namely the anisotropy of the longitudinal modes and the concave curvature of the transverse mode propagating in the basal plane are less pronounced or absent, respectively. The optic modes of the two compounds differ remarkably from each other with respect to the energetic position as well as to their sequence. PI SI M K

1 1 1 1 1 1 1 1 1

0 0 30 - O— .0.. ft, 4, • •o. •k

- * a ..a *

'••m. E — • •- - — 20 o. 0 0. ••o 6

»••. 'o..

Ir M< > ' • 0) 9 q

a a-'

C 10 • LJJ 0 o _,*• - • • •••. .••*"•' •••• *' 'V* t

1 I i Wavevector q (A"1)

Fig, 1: Experimental phonon dispersion curves of HfSe2. The dotted line is a guide to the eye. Polarization vectors: transversal •, longitudinal o and mixed o.

Recently, Harbec et al, (1) have performed a calculation of the vibrational spectrum of SnSe2 basing on the extended-shell model concept of Benedek and Prey (2). This calculation reproduces very well the optical zone center data and those neutron data which were used in the adjustment of the model parameters. Discrepancies exist, however, for modes along directions or with symmetries, which were not included in the model adjustment procedure (3). That means, that even the parameters of a physically realistic model only can be satisfactorily determined if enough orthogonal information from r.any q-vectors in the Brillouin-zone is available. PI B 1

The origin of the observed metallic properties of TiS_, which is now generally accepted to be a degenerated semi-conductor with an indirect band-gap of 0.1-0.3 eV, is still in dispute. The electrical resitivity shows a Tp -dependence, where 1.6«p«2.3, depending on the stoichiometry of the sample. Several inter- pretations of this temperature dependence have been given in terms of first-order electron-phonon or electron-electron scattering. Kaveh et al. CO have proposed the existence of a soft longitudinal accoustic interlayer mode propagating in direction perpendicular to the basal plane. From symmetry considerations this mode couples to the conduction electrons only quadratically, leading to a T -depencence of the electrical resistivity. We have looked for this mode, which cannot be detected by optical methods, using inelastic neutron scattering. The experimental determination of phonon dispersion curves in TiS. is extremely difficult due to the lack of sufficiently large single crystals. The neutron data indicate the existence of a small structure at the zone boundary (O,O,TT/C) at an energy of about 3 meV, i.e. slightly above the value of 1.7-2.5 meV estimated by Kaveh et al.. In order to attribute unambiguously the observed structures to the proposed phonon mode, further measurements are in progress.

(1) J.Y. Harbec, B.M. Powell and S. Jandl, Phys.Rev.3 28, 7009 . (1983). (2) G. Benedek and A. Frey, Phys.Rev.B 21, 2482 (1930). (3) W. Buhrer and F. Levy, to be published. (4) M. Kaveh, M.F. Cherry and M. Weger, J.Phys.C: l£, L789, (1981).

M. SchSrli, Labor fUr Meutronenstreuung, ETH Zurich, CK-5303 Wurenlingen, Switzerland P 1 B 2

MAGNETIC AND THERMOELECTRIC PROPERTIES OF SINGLE CRYSTAL CoTe AND NiTe H. Schicketanz, P. Terzieff, K.L. Komarek Institute of Inorganic Chemistry, University of Vienna, A-1090 Wien, Austria

In the past much attention has been paid to the broad range of homogeneity of the NiAs/Cd(OH)--type phases CoTe and NiTe. The magnetic properties have been studied repeatedly, however, to some extent the results are still inconclusive. The deviations from the Curie-Weiss behaviour observed in polycrystalline CoTe have been assigned to the superposition of a Pauli-and a Curie-Weiss-term (1-3). In polycrystalline NiTe a magnetic anomaly at about 130 K has been reported by some authors (3-4), while others observed a smooth variation down to 100 K (5-6). In view of the inconsisten- cies we decided to repeat the measurements on single crystals of CoTe and NiTe. The magnetic susceptibility of CoTe, measured both

1.6 1 0 em i

X >, 16

Ja ^_ o.

100 150 200 250 300 Temperature, T (K)

Fig.1 Magnetic susceptibility of single crystal CoTe and NiTe measured normal ( Xj. > and parallel ( X,, ) to the c-axis. P 1 B 2 parallel and normal to the c-axis, was clearly temperature dependent, but considering the weak susceptibilities the interpretation in terms of a Curie-Weiss-type behaviour seems to be questionable. The experimental data could not be fitted by splitting the magnetic susceptibility into a temperature independent and a Curie-Weiss- term. The measurements on NiTe yielded a weak constant paramagnetism, a considerable degree of anisotropy, and a smooth variation down to 100 K without any indication of the anomalies which have been assigned to antiferromagnetic interactions (3,4). Data on the thermoelectric properties are scarcely available for CoTe (2), and contradicting in the case of NiTe (4,7). For CoTe our measurements revealed a negative Seebeck coefficient increasing in magnitude non-linearly with temperature (Fig.2). The Seebeck coefficient of single crystal NiTe was found to be also negative with a maximum located at about 160 K. This is in remarkable contradiction to the positive Seebeck coefficient and the anomaly at

100 1S0 200 2SQ Ttaptrjfurt, T (K)

Fig.2 Seebeck coefficient of single crystal CrTe (+,• run 1,2 ) and NiTe (x,o run 1,2 ) measured normal (Sj.) and parallel ( S|| ) to the c-i PI B 2

about 240 K which have been reported for NiTe (7). In addition to the measurements on single crystals, the variation of the magnetic susceptibility and the Seebeck coefficient with the composition was studied by using polycrystalline samples.

(1) H. Baraldsen, F. Grtfnvold, and T. Hurlen, Z.anorg.allg.Chem. 283, 144 (1956) (2) G. Saut, C.R. Acad.Sci.Paris 26J_, 3339 (1965) (3) E. Vandenbempt, L. Pauwels, and K. De Clippeleir, Bull.Soc. Chim.Beiges 80, 283 (1971) (4) G. Saut, C.R. Acad.Sci.Paris B263, 1174 (1966) (5) J. Barstad, F. Grcinvold, E. Rtfst, and E. Vestersjtf, Acta Chem. Scand. 20_, 2865 (1966) (6) E. Uchida and H. Kondoh, J.Phys.Soc.Japan V\_, 21 (1956) (7) A.G. Rustamov, J.G. Kerimov, L.M. Valiev, and S.Ch. Babaev, Neorg.Mat. 7, 1123 (1971)

H.Schicketanz, P.Terzieff, Institute of Inorganic Chemistry. University of Vienna, A-1090 Wien, Austria P 1 B 3

WSea HOMO- AND HBTEROJUNCTIONS R.r. Sgaeh, M.Ch. Lux-Steiner, M. Obergfell and B. Bucher Fakultaet fuer Physik, (Jaiversitaet Konstanz, Postfach 5560, 7750 Konstanz, Fed. Rep. of Qeriany

In our search for alternative solar cell materials, various transition metal chalcogenide single crystals have been grown, such as WSea, MoSea, ZrSe3 and ZrS3. Homojunctions and heterojunctions of these materials have been prepared and characterised . The weak bonding between the layers of the single crystals implies that surface reconstruction will be negligible during the formation of hetero.junctions, so that no interface states may be introduced and lattice match is not important. Whereas the values of the energy gaps of WSes , MoSea and ZrSe3 promise high photovoltaic conversion efficiencies for these materials, ZrS3 wiJl be suited as window material of photovoltaic cell because Ox its larger energy gap. Single crystals of WSea and MoSea can be made both p- and n- type depending on the transport agent used for the vapour phase transport reactions. Single crystals of the other transition metal chalcogenides mentioned above exist only as n- type material when they are grown by vapour phase techniques. Fig. i represents the electrical resistivity for the single crystals. The Hall mobility of these crystals is depicted in fig. 2.

10'3* RT

Fig. 1 Resistivity of the Fig. 2 Hall mobility of the single crystals single crystals P 1 B 3

The Measurements of the electrical properties were carried out perpendicular to the c- axis of the crystals. Gallium iodiun or tin indiun alloys were used for ohmic contacts to n- type materials and evaporated gold, or silver paste for p- type Materials. Diodes were prepared by chemical vapour transport and deposition of epitaxial layers on single crystal substrates in sealed quartz ampoules. The layers were grown by the same technique used for the corresponding single crystal growth. The individual growing conditions of these single crystals determine the procedure. The WSe2 homojunctions and the n- MoSe2/p- WSez heterojunctions were prepared by deposition of p- type layers on n- type single crystals. K- ZrSe3/p- WSea and n- ZrSa/p- WSe2 heterojunctions were fabricated reversely. Fig. 3 shows the IV- characteristics of these junctions at room temperature.

— pn WSe2

— MoSe2/WSe2

--ZrSe3/WSe2

10"

-10" -1 -0.5 0 0.5 1 VOLTAGE [V]

Fig.3 IV- characteristics of typical WSe2 homojunctions and MoSes/WSes and ZrSe3/WSe2 heterojunctions.

For th«? WSe2 homo- and the MoSez/WSez heterojunctions the net carrier concentrations calculated from the CV- data agree with the values obtained from the resistivity and HaJ1 measuerements of the single crystals. The moderate photovoltaic effect of these devices (conversion efficiencies up to 0.6 *) led us to consider a heterojunction diode based on WSe2 in combination with a large gap window material. Therefore ZrS3 layers were growu on WSe2 substrates. Infact, the IV- characteristic of this junction shows soae photosensitivity (fig. 4) and exhibit a rectification ratio at IV bias of 20 or less. 1.0 ' 1 >/p- OS

RENT( m • or /

/ , -1.0 -0.5 0.5 y 4 VOLTAGE (V) // -o^ / / 0

* -i.o

Fig.4 Dark and illuminated (laser illumination) IV- characteristic of a ZrS3/WSef> hetero.junct ion.

The photocurrent (laser illumination hv = 1.96 eV) depends on bias and reaches saturation at several volts. The quantum efficiency for the laser light was approximately 0.5.

R. Spaeh, Fakultaet fuer Physik, Universitaet Konstanz, 1 B 4

LATTICE INSTABILITIES IN THE PSEUDOBINARY SYSTEM

J. M. Tarascon Bell Communications Research Inc., Murray Hill, New Jersey 07974

F. J. DiSalvo and J. V. Waszczak AT&T Bell Laboratories, Murray Hill, New Jersey 07974

The structure of Chevrel phases of formula MMo(X, (M = metal; X = chalcogen) can be

though of as a stacking of Mo6X, clusters with large cavities between them which accommodate the ternary metal atom. When M is large it is located on the 3 axis (structure I), and when M is a small cation it randomly occupies sites around the 3 axis

(structure II) leading to rhombedral angles (exr) smaller or greater than 90" respectively. The

crystallographic parameter at reflects the displacement of the ternary element from the 3 axis and it has been suggested'1' that ex, varies smoothly between structure I and II, and that large

values of at were associated with the temperature induced lattice instabilities of the type II

structures (Cu2Mo6S,, Fe2Mo6S,) from rhombohedral to triclinia However recent crystallographic investigations'2 >3) have also revealed phase transitions from rhomboedral to

triclinic in the type 1 phases such as EuMo6S, and BaMo6S,.

The present work was undertaken to further investigate these structure instabilities and

their effect on the electronic properties. The system InMo6S,_ISe1 was selected for this study (< because it has an unusual change in a.t of about 3° between InMo6S, and InMo6SC|. > This is the first series of compounds in which the composition and temperature dependence of the displacement instability of the ternary element could be studied.

The mixed compounds InMo(S|_tSex prepared by low temperature diffusion (500°C) of

indium into the channels of the binary phase MotS,__.Se, were characterized by X-ray, magnetic and superconductivity measurements. Room temperature X-ray studies show (Figure 1) that a, decrease sharply of about 3s around x = 6 and furthermore that the

hexagonal unit cell volume (vk) does not obey a Vegard's law and exhibits the largest deviation near x * 6. Powder diffraction patterns of samples, ranging from x = 5 to 7 by steps of 0.2 (Figure 2) indicate that for composition between x • 5.8 and 6.2 two rhomboedral phases with angles of 93* and 90* apparently coexist. Finally X-ray diffraction

patterns as a function of temperature (figure 3) reveal for instance that InMo#S2 4Se51

undergoes a phase transition. The low temperature phase with at of 93* transforms into a high temperature phase with a, = 90*, however in the intermediate temperature range, P 1 B 4

between 397K and 433K, these two rhomboedral phases with angles of 93" and 90° coexist. These results indicate that both temperature and chalcogen substitution induce the first known rhomboedral to rhomboedral structural phases transitions ever reported for the Chevrel phases. These transitions whose nature will be discussed are characterized by a large

discontinuous change in o, of about 2°. This region of instability in or may explain the lack of Chevrel phases with rhomboedral angles between 90.5" and 92°. We also found that the temperature at which these structural instabilities occur correlated perfectly with anomalies in the temperature dependence of both susceptibility (fig. 4) and resistivity and also with the onset of endotherranic peaks in calorimetry measurements.

(1) K. Yvon, Solid State Comm. 25, 327 (1978). (2) R. Baillif, A. Junod, B. Lachal, J. Muller and K. Yvon, Sol. State Comm. (1981), 40, 603. (3) R Baillif A. Dunand, J. Muller and K. Yvon, Phys. Rev. Lett. (1981) 47, 672. (4) J. M. Tarascon, F. J. DiSalvo, D. W. Murphy, G. Hull and J. V. Waszczak, Phys. Rev. B. (1984), 29, 172.

0.5 -970 D)

0.4

b 0.3

= 0.2

*O0 600 TEMPERATURE IK1

Figure 4. The temperature dependent susceptibility of several members of the series InMo^.jSe, are shown. The arrows distinguish between cooling or warming of the sample. FJgarel. Rhombohedral angle (a,), and hexagonal unit cell volume vk as function of selenium replacement for sulfur in InMo,S,_lSe1. IWO .00 {l0 *" 210 1 |.» 200 »0

•

82 27 i If] J 1*9.8 T • StT K hL • ,.ju.i

vt H , Jj *»6.2 i JLuJL i i .i J\. It

INTE I ;

X'6.8 .till 100 2*0 210 T.S30K (100) So i" T WO)OljfflO, l rM>v, I ULLLL j

Figure 2. X-ray powder diffraction patteras are Figure 3. X-ray powder diffraction panems as shown for several members of the InMo^Sj.^Se, function of temperature are shown for series. lnMo4S,«Se56.

800 Mounam *wnuc. MH IA-180 Mumv H* P 1 B 5

FIRST EXAMPLE OF (MojSe^J CLUSTERS IN SOLUTION

J. M. Tarascon Bell Communications Research Inc., Murray HDl, New Jersey 07974

F. J. DiSalvo, C. H. Chen, P. I. Carroll, M. Walsh and L. Rupp AT&T BelJ Laboratories, Murray Hill, New Jersey 07974

(1) Molybdenum cluster compounds MosXs containing either chalcogens (X = S, Se, Te) or halogens (x = Cl, Br, I) have been intensively investigated. These clusters are bound together either by intercluster molybdenum chalcogen bondssuch as in the Chevrel phases or

by bridging halogens such as in Mo(Clu. The different balance between the ionic and covalent components of the bonding is responsible for the difference in chemical behavior of these compounds. Indeed the solution phase chemistry of the halide clusters is well known,

while so far no similar chemistry has been reported for chalcogenide clusters such as Mo6S^~,

MojSff, Mo12Sf4 or (Mo3Se3-')i. The quasi one-dimensional MjMo6X6 compounds which

can be thought of as infinte chains of (Mo3X,)i clusters separated by the ternary element have a larger intercluster (e.g. interchain) distance than in the Chevrel phases 6.40A instead of 3.4A respectively. This considerably reduced intercluster interactions leads one to expect the solution chemistry of molybdenum chalcogenides to be easiest for these linear chain species. We have undertaken research in this direction, and here we report that some of the

one-dimensional M2MosX4 (M = Li, Na; X = Se, Te) compounds can be dissolved in highly polar solvents.

The polycrystalline lithium and sodium compounds (prepared by a low temperature ion- exchange reaction as previously described'3') were observed to spontaneously forms gel like masses when dropped into polar solvents such as dimethylsulfoxide or N-rnethylformamide. These gels can be diluted to produce free flowing liquids that look reddish or purple for the selenides and tellurides respectively, but also the dissolved solid can be precipilated from the solvent by adding salts or acids, leaving behind a clear solution. We will show that these conditions of swelling are strongly affected by the nature of the ternary element, its concentration in the starting material, and by the physical properties of the solvent (dipolar moment and dielectric constant).

The nature of the liquids obtained from the solid lithium molybdenum chalcogenides were examined by a variety of techniques. We show using optical microscopy, transmission electron microscopy and light scattering that some of these solutions contain individual chains (figure 1), yielding the first example of purely inorganic transition metal P 1 B 5

polymer solutions. The strong ability of these chains to be easily oriented by shear flow (drawing or painting) win also be presented. Polarized light as well as EPR and resistivity measurements were used to determine the orientation of these chains. For instance we found that solution flowing either through a flat thin tube (0.5 mm) or a small cyillary strongly polarized transmitted light (Figure 2).

(1) K. Yvon: "Bonding and Relationships between structure and physical properties in Cheviel-phase Compounds M,Mo6X, (M = Metal, X = S, Se, Te)" in current Topics in Materials Science, Vol. 3 ed. by E. Kaldis (North Holland, Amsterdam 1979). (2) H. SchSfer, H. G. Von Schmering, J. Tillack, F. Kuhnen, H. Wfihrle, H. Bauman: Z. Anorg. AJlgra. Chem. 353, 287 (1967). (3) I. M. Tarascon, G. W. Hull and F. J. DiSalvo, Mat. Res. Bull., 19, 915 (1984).

: ,^»;_ '•'/{£•-. ..!'• ;.:^rL >

Figure 1. Transmission electronit microscopy pictures collected for a solution of Li2MotSet with increasing dilution. 7he dilution increases by alphabetical order. P 1 B 5

Figure 2. The ability of the solution to polarize light is shown. Between a and b the polarizer has been rotated by 90".

B«« Commoncjnont flatorch GOO Mounun Avtnut. MH IA-180 Murray H*. NJ 07974 P 1 B 6 ON THE VANADIUM - TELLURIUM PHASE DIAGRAM P.Terzieff and H. Ipser Institut fur Anorganische Chemie der Universitat, Wahringerstrafie 42, A-1090 Wien, Austria E. Wachtel Max-Planck-Institut fiir Me tal If orschung, Institut fiir Werkstoff- wissenschaften, Seestrafle 92, D-7000 Stuttgart 1, F.R.G.

The vanadium-tellurium phase diagram is characterized by the appearance of a system of phases with NiAs-derivative structures between about 52.5 and 66.7 at% Te; additionally there exists a • monoclinic phase V.Te, with a narrow range of homogeneity. The first systematic X-ray investigation was performed by Grtfnvold et al. (1). Their results for samples quenched from 1273 and 1073 K are indicated in Fig.1 by the hatched stripes, where y. designates the undistcrted hexagonal NiAs-structure, *,'., is a monoclinic phase,

1600

800 -

600 -

400 -

60 »t%Tt 65

Fig.1 Partial phase diagram of V-Te (O,D thermal effects; + magnetic measurements; * thermodynamics (6); x, •, literature data 12), (4),(5)) P 1 B 6

and Y, is again a hexagonal phase with a Cd(OH)--type structure;

the crystal structure of Y4 was reported to be orthorhombic, however, it was shown by Rtfst et al.(2) to be actually monoclinic. The phase V-Teg was discovered by Brunie and Chevreton (3) and seems to be identical with the V2Te3 reP°rted bY <2'• ^e phase transition at the tellurium-rich end at low temperatures was first detected by Ohtani et al< (4), and the crystal structure of the monoclinic low temperature modification of VTe, was reported by Bronsema et al. (5). The same authors found also a sample with 65.8 at% Te to be hexagonal above the corresponding transition temperature which is in contradiction to earlier findings.

3.U

o 3 01

1100 1200 T (K)

Fig.2 Magnetic susceptibility of NiAs-type V-Te alloys PI B 6

In the course of the present investigation it was tried to combine the results of differential thermal analysis, magnetic measurements and X-ray measurements with the results of previous investigations (1-6) in order to clarify some of the phase relationships in this system. The most striking feature is the discovery of a phase transition with a maximum at 57-2 at% Te (corresponding to V,Te.) and 1180 K which was unknown up to now *). Both DTA and susceptibility measurements (see Fig.2) reveal clearly the existence of this transition, whereas no difference could be found in the powder pattens of samples quenched from temperatures above and below it.

(1) F. Grtfnvold, O. Hagberg, and H. Haraldsen, Acta Chem. Scand. _1_2» 971 (1958) (2) E. Rtfst, L. Gjertsen, and H. Haraldsen, Z.anorg.allg.Chem. 333, 301 (1964) (3) S. Brunie and M. Chevreton, Bull.Soc.Fr. Mineral. Cristall. 91, 422 (1968) (4) T. Ohtani, K. Hayashi, M. Nakahira and H. Nozaki, Solid State Commun. 40, 629 (1981) (5) K.D. Bronsema, G.W. Bus, and G.A. Wiegers, J.Solid State Chem. 5_3, 415 (1984) (6) H. Ipser, J. Solid State Chem. 54, 114 (1984)

*) After finishing this abstract a paper by T. Ohtani, S. Onoue, and M. Nakahira appeared confirming our results (Mat. Res. Bull. 19, 1367 (1984)).

P.Terzieff und H.Ipser, Institut fur Anorganische Chemie der Universitat, WShringerstraBe 42, A-1090 Wien, Austria P 1 B 7

CRYSTALLOGRAPHIC AND MAGNETIC STRUCTURE OF Tl-Fe-S COMPOUNDS 0. Welz, P. Oeppe, M. Rosenberg, Institut fur Experimentalphysik IV, Ruhr Universitat, D 4630 Bochum, H. Sabrowsky, Institut fur Anorganische Chemie I, Ruhr Universitat, D 4630 Bochum, and U. Schafer, Mineralogisches Institut der Universitat, D 5300 Bonn

Ternary sulfides of Thallium and Iron obtained in powder form by quenching from stoichiometric melts have recently been reported (1,2,3), namely monoclinic

TIFeS- (raguinite), and tetragonal TIFe 57 in the continuous range of phases l,3-x-l,8. The structures are built from parallel chains respectively layers of edge connected FeS^-tetrahedra, separated by Tl-ions. In order to learn about their magnetic nature and to improve on the description of the behaviour across the range of layered phases, we undertook a neutron-diffraction (SV7 spectrometer at FRJ-2 reactor in Julich) and MoBbauer effect study of these compounds. According to the MoGbauer spectra, monoclinic TlFeS2 is paramagnetic at room temperature, whereas at 4.2 K magnetic ordering occurs with an iron hyperfine field value of 16.4 T (Fig.l). From the onset of magnetic splitting the transition temperature was fixed at 190 ±10 K. In agreement to the interpretation as a transition to three-dimensional magnetic order of the Fe-chains, the neutron spectrum at 16 K shows additional magnetic diffraction peaks as compared to room temperature. From these the spin orientation turns out to be perpendicular to the chain axis, and with the additional assumption of antiferromagnetic order along the chains the orientation in a vertical crossection results to antiparallel (Fig.2). On the basis of the present spectra we were not able to determine the overall orientation with respect to the cell axes. The magnetic behaviour is similar to that observed on related alkali sulfides (4,5), but there the orientation in a vertical crossection is found to be parallel. In the case of the layered TIFe,£- phases the Fe-S-planes differ in Fe-occupation (65 to 90 55), thus rising the question of vacancy ordering. Indeed for TIFe. 5S_ from X-ray data (6,7) complete ordering, also known from Cs2MnjS4 (8) and Cs-Zn^S^ (9), has been inferred, which i3 characterized by the loss of equivalence of both tetragonal axes (Fig.3a). This is confirmed by our neutron P 1 B 7 spectra where split reflections for h*k#0 and additional peaks both correspond to this superstructure. For x=1.6 the splitting vanishes and different additional peaks appear, which in a preliminar analysis could be linked to the most simple ordering possible for this concentration (Fig.3b). In the range x»1.6 our samples show two concurrent phases of different c-axis with preference of the shorter axis for increased x, which in (7) after annealing had been obtained separately for x=1.6 and £1.65. The MSSbauer spectra of these compositions are paramagnetic at room temperature, but magnetically split at 77 or 4.2 K (Fig.4). Their structure at 4.2 K has to be interpreted in terms of contributions from environments with different numbers of Fe-neighbors. Thus the dominant hyperfine field value of 30 T may tentatively be related to three neighbors, whereas with four neighbors the Fe-moment seems to be lost completely. The comparatively simple spectrum for x=1.6 hints at a structure with all Fe being equivalent, and this is true indeed for the superstructure of Fig.3b. Moreover this allows to explain the different c-axes for x-1.6 by different types of stacking, e.g. identical and statistical. In contrast to the magnetically split MoSbauer spectra no three-dimensional magnetic ordering has been observed in neutron spectra down to 16 K. The MSQbauer splittings thus are a manifestation of twodimensional ' magnetic order only within the separate Fe-planes.

(1) K. Klepp, H. Boiler, Mh. Chem. 110, 1045 (1979) (2) M. Zabel, K.-J. Range, Z. Naturf. 34b, 1 (1979) (3) H. Sabrowsky, J. Mirza, C. Methfessel, Z. Naturf. 34b, 115 (1979) (4) P. Mliller, dissertation, Aachen (1980) (5) M. Nishi, Y. Ito, Solid State Comm. 30, 571 (1979) (6) M. Zabel, K.-J. Range, Rev. Chim. Miner. 17, 561 (19B0) (7) M. Zabel, K.-J. Range, Rev. Chim. Miner. 21, 139 (1984) (8) w. Bronger, P. Bottcher, Z. anorg. allg. Chem. 390, 1 (1972) (9) W. Bronger, U. Hendriks, Rev. Chir. Miner. 17, 555 (1980) P 1 B 7

-6-4-2 0 2 4 6 relative velocity / mm/s

Fig.l: MoGbauer-spectrum of TIFeSj Fig.2: (001)-projection of the at 4.2 K. crystallographic unit cell of TlFeS2 with orientation of magnetic moments.

I I I I

13 «'I.S loom = 16 I Lin

Fig.3: Vacancy ordering of TlFexSa for x=1.5 and 1.6. Hatched areas denote unit cells, space group designations correspond to identical stacking.

Fig.4: MoBbauer-spectra of TlFexSx for x=1.4, 1.6, and 1.8 at 4.2 K. -6-4-2 0 2 4 6 ralativ* velocity / mm/s PI B 8

CHANGE IN RAMAN SPECTRA OF (TaSe^I INDUCED BY THE PEIERLS TRANSITION A. Zwick and M.A. Renucci Laboratoire de Physique des Solides, Associe au C.N.R.S., 118, Route de Narbonne, 31062 Toulouse Cedex (France) P. Gressier and A. Meerschaut Laboratoire de Chimie des Solides, 2 rue de la Houssiniere, UU072 Nantes (France )

The chain compound (TaSe^JpI undergoes a Peierls transition at 263 K (1), associated with the formation of an incommensurate charge density of wave (ICDW) as is evidenced by superlattice spots both in X-rays (2) and electron diffraction

patterns (3). The present work provides an investigation of the lattice vibra-

tions of (TaSe[i)pI between 10K and U00 K by means of Raman spectroscopy, in view of studying possible changes induced in the spectra at the phase transition.

Single crystals were used for the Raman scattering experiments with typical dimensions of about 3 x 0.3 x 0.2 mm . The samples were freshly cleaved and immediately immersed in the exchange gas (He) of a cryostat for room and low temperature measurements. Temperature above 300 K were obtained with a specially cons- o tructed furnace. The Raman spectra were excited with the 51^5 A line of an A ion laser in the conventional bttckscattering geometry. The laser beam was focused onto the surface of the sample at nearly normal incidence, with polarization either parallel or perpendicular to the c axis. Analysis of polarization was performed on the scattered light. Care was taken to avoid local hpsting by utilizing a cylindrical focusing lens and a low excitation power level.

(TaSe,)„! crystallizes at room temperature in a. tetragonal structure which belongs to the IU22-S5? space group (k), with two molecular units per primitive unit cell. Among the 66 K = 0 phonons, 55 are optical Raman active modes. They are represented by the following irreducible representations of theSX point group :

IV, = 7 A1 + 7 B1 + 7 B2 + 17 E.

Therefore, there will be 21 non degenerate and 17 doubly degenerate Raman active lattice vibrations in the undistorted phase of (TaSer)pI. Figure 1 shows the high temperature polarized spectra of (TaSej^I for the three configurations

ei II eg [ c, Zi II eg // c and e^ \ es // c. "? 1 B 8

About fifteen structures appear in the frequency range from 20 to 300 cm" instead of thetO thev retically predicted, that implies maiy de- generacies. The structures will be studied in details in a forthcoming paper, but even now we confidently ascribe the highest frequency modes at 21k and 283 cm (room temperature wave numbers) to stretching mcdas of tightly bound Se-Se pairs, consistently with other selenides showing (Sep) pairing. 100 200 300 WAVENUMBER [cm-1) The temperature dependence of the unpolarized spectrum of (TaSe^pI for e. //c , in the 1 - Polarized Raman spectra ^00Kexcited range 10K-U00K,is depicted in figure 2. The scatterr.-i intensity has been corrected for phonon population effects and normalized to the integrated intensity of the peak at 27^ cm . Many structures display an anomalous behaviour when lowering down the tempe-

rature through T . The line at 10U Cm~ , for (TaS«t)2l instance, increases in intensity below T A=4880A « T(K] ms i 3 Besides, new lines show up close to and against \x 400 .0 —— some dominant peaks of the high temperature <5 A A AAJ AVL _ -' Al l /VJ2A 0 spectra, while diverging markedly from them. CO z l 1 \_ This behaviour is typified b;- the main peak u V> K jJv \ \225 at 185 cm (room temperature wave number) and iz 1 /\ the satellite peak, appearing on its high fre- < quency side. The intensity transfer between RA M the two lines below T is obvious in fig. 2. o ° 0 100 200 300 Figure 3 depicts the shift of the satellite WAVENUMBER (cm"1) line from its partner.

Actually, it is impossible to decide wether Fig. 2 - Temperature dependence of the unpolarized spectrum of or not the satellites lines are really new. (TaSe.)2I for % //c" , in the They could equally follow from k * "o modes range 1OK-UOOK . of the high temperature phase, or from~k=l5 low efficient Raman asodes, degenerate with more efficient on»s responsible for the main peaks of the undistorted structure. Ir. any case, these satellites peaks are undoubtedly activated solely in the CDW state.

We put together similar trends displayed by Raman spectra below the phase-transition temperature in other chain-structure compounds as TaS_ (5), vhere the formation of a commensurate charge density wave (CCDW) is ascertained by X-rays diffraction, and another tetrachalcogeni- de compound, (HbSe^) I , (6) even if the latter 200 undergoes a space groupe change with no yet TEMPERATURE CK) evidence for CDW state.

Fig. 3 - Shift of the satellite line from the main peak at 185 cm"1 versus temperature. The arrow indicates the phase-transition temperature.

(1) Z.Z. Wang, M.C. Saint-Lager, P. Monceau, M. Renard, P. Gressier, A.Meersmeut, L. Guemas and J. Rouxel, Solid State Commun. h6_, 325 (1983). (2) H. Fujishita, M. Sato and S. Hoshino, Solid State Commun. J+£, 313 (198U). (3) C. Roucau, R. Ayroles, P. Gressier and A. Meerschant, J. Phys. C 17,2993 (198U). — (4) p. Gressier, L. Guemas and A. Meerschaut, Acta Cryst. B 38, 2877 (1982). (5) S. Sugai, Phys. Rev. B 29, 953 (1983). (6) T. Sekine, K. Uchinokura, M. Izumi and K. Matsuura, to be published in Solid State Commun.

£>1 M.A.

Urn vutitl' Tank SafcjcU-.tr^ AU Swxt- ck t\ioii«wu 5* obi Tcutoo*- P 1 B 9

The pyrochlorea Pb-M^Sb, -0. _ (M111: Al, Sc, Cr, Rh)

C. Cascales and I. Rasines Institutes de Qulmica Inorganica Elhuyar, Serrano 113, 28006 Madrid, Spain.

P. Garcia Casado Facuitad de Farraacia. Universidad de Navarra. Pamplona. Spain.

J. Vega Dpto. de Quimica Inorgania. Facuitad de C. Quimicas. Universidad Complutense. 2S040 Madrid, Spain.

By heating mixtures of PbO and Pb Sb 0 , in vacuum at 1223-1273 X, Moisan et al. (1970) (1) prepared the rhombohedrally distorted pyrochlore Pb.Sb.O . Surchard et al. (1978) (2) obtained at 973 K a cubic pyrochlore, II r IV V i of an approximate composition Pb™ [ Pb. =Sb ] 0g __, which after heated it 1173 K gives rhombohedral Pb_Sb_O_. The present communication reports how

III Sb 0 •M III: A1 Sc to prepare pyrochlore oxides of composition P^-MQ = i 5 g = ( > > Cr, Sh) and their relevant crystal data.

Experimental

Pb.M Sb, _0.E (M : Al, Sc, Cr, Rh) were prepared from a mixture 2. 1.3 0.3

of analytical grade PbO, MO, and Sb-03, in the molar ratios Pb:M:Sb = 4:1:3, which were heated in air at 953, 1.073, 1.123 and 1.223 K. After each thermal treatment for 24 hours, the materials were quenched and ground. LECO 528-018 crucibles and a CHESA furnace were employed. Temperatures were read with a calibrated Pt-Pt/Hh '•hermocouple and were reliable to ± 5 K. X-ray diffraction peaks were recorded using '4(99.99 %) as internal standard, and Ni radiation in a Philips PW 1310/00 diffractometer and a 1050/25 goniometer provided with a curved LiF-crystal-focusing monochromator, at a scanning rate of 0.125' 28 itiin" . The unit-cell parameters were refined from the 28.values of the last ten reflections. X-ray intensity data were collected using a Siemens Kristalloflex 810 and a 0-500 goniometer provided with a graphite monochromator. The patterns were step scanned from 13 to 154»2e ,with increments of 0.02°2e. The counting time at each step was 4 sec. Background corrections were made by linear interpolation between regions judged to be fins background. The intensities were calculated with the program Lazy-Pulverix (3), and Eebye-Waller factors, Bj: 0 :0.a0, Pb: 0.01, Sb: 0.38, Al: 0.75, Sc: 0.68, Cr: 0.65, Rh: 0.44 A2, 1 B y

including Lorentz and polarization factors and corrections for anomalous disper- v v v sion. The discrepancy factors, R • (E|IQ -1^1)/ZIQ were computed after making

Results and discussion

The four samples were obtained as coloured powders, which produced good X-ray diffraction patterns, characteristic of cubic pyrochlores, spacegroup Fd3m (No. 227), Z = 8, and the a, V and D values shown in Table I. The best discrepancy factors, R, were obtained for Pb in 16(c), 0, 0, 0, origin at center, 3m, M111 and Sb at the ratio M :Sb = 1:3 distributed randomly in 16(d), 0.5, 0.5, 0.5, and 0 atoms at 48(f), x, 0.125, 0.125 and half 8(a) positions 0.125, 0.125, 0.125. The x values and observed interatomic distances are also included in Table I. The mentioned distances agree quite well with the sums of the Shannon (4) effective ionic radii if Pb(II) seven-coordinated is assumed, and the weighted average of six (c)-(f) and one (c)-(a) distances is taken as Pb-0 observed distance.

I 111 The deficient pyrochlores Pb2M* *Sb1 506 5 (M : Al, Sc, Cr, Rh) can be thought as the result of the substitution of 0.5 Pb(IV) by M(III) IV in cubic Pb_Pbn cSb. _0c __, with subsequent production of the corresponding oxygen vacancies.

References

(1) Hoisan, J.Y., Pannetier, J., tucas, J. C.R. Acad. Sc. Paris, t. 271, C (1970). (2) Buehard, G. RUdorff, W. 2. Anorg. allg. Chem. 447 (1978). (3) Yvon, R., Jeitschko, W. Parthe, E. J. Appl. Cryst. 10, 73 (1977). (4) Shannon, R.D. Acta Cryst., A32, 751 (1976). P 1 B 9

Table I. Structural data for Pb_M. _Sb- -0. _ (M Al, Sc, Cr, Rh) 3 O • 3

Al Se Cr Rh

a/A 10.3964(1) 10.5558(1) 10.4494(1) 10.4738(1) V/A3 1123.70(4) 1175.18(3) 1140.97(3) 1148.99(3) • 3 Dc/Mgm~ 8.45 8.17 8.46 8.70 X* 0.435 0.430 0.432 0.432 H 0.049 0.051 0.020 0.039 Pb-0 /A 2.603 o 2.61 2.596 2.603 Pb-0 Ik 2.61 2.61 2.61 2.61 M,Sb-0 /A 1.960 2,008 1.981 1.985 M,Sb-0 /A 1.964 2.016 1.984 1.996

origin at center, (3m). PI B lo

The oxides KgSbgO 2 (M - Y,Pr,Nd,S»,Eu,Gd,Tb,Dy,HofEr,Tm,Yb,Lu)

CM. Marcano and I. Rasines

Instituto de Quimica Inorganica Elhuyar, Serrano 113, 28006 Madrid, Spain.

It is known [l] that Nd^ can react with St>203 (Nd:Sb = 1:1,73) at 873-973 K in vacuum, to produce a cubic oxide, a = 11.000 A . More recently

M 0 it has been shown [2 ] that cubic crystals of 3S" 50.2 (M = La, Pr, Tb, Yb), o _ a from 11.027 to 10.721 0.003 A, S.G. 143m, can be obtained from mixtures of

Sb.O, and M0 by the hydrothermal method, at 720-S20 K and (1.2-1.5)10S Pa.

In this comunieation we wish to report the existence of the new oxides M,Sb_O._

(M = Y, Nd, Sm, Eu, Gd, Dy, Ho, Er, Tm, Lu). These oxides, along with those for which M = Tb and Yb, were studied by X-ray diffraction and spectroscopic techniques.

M3Sb50 (M a Y, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu) were prepared from mixtures of analytical grade M_0_ and Sb_O, (M:Sb = 3:5) heated in vacuum ( * 10~ mm Hg) at 963 K for 7 days in quartz ampoules, and quenched in liquid N_. Experimental details about the employed furnace and diffractometer can be found elsewhere [ 3 ] . The unit-cell parameters were refined from the 2 8 values of the last seven reflections. Infrared spectra were recorded using a Perkin Elmer 325 spectrophotometer and KBr disks.

M.Sb.O.. oxides were obtained as polycrystalline powders colored as indicated in table I. X-ray diffraction data and IR absorption bands frequencies are included in tables I and II respectively. The densities were calculated assuming Z = 4.

Cf JCPDS card No.26-111. PI Bio

The be observed reflections are allowed in the S.G. I43m (No. 127).

The measured a values for the compounds of Nd, Tb and Dy result comparable

to those previously reported [ 1 ] [ 2 ], specially in the fir3t two cases. The observed a values correlate well (r » 0.9998) to the Shannon effective ionic radii [ 4 ]for eight-coordination.

As for the IH absorption frequencies included in table II, the first band in the spectrum occurs for Pr at 685 cm" and is a strong band which shifts to higher frequencies (up to 695 cm •"", for the compound of Lu) as the formula weight increases. Thus, it is quite apparent that this strong band does not correlate with mass and has to be considered as an effect of lattice vibration. The six following bands, which occur in the region 500-275 cm" , also behave in the same manner. In all, the spectra of the oxides M Sb.O appear rather similar to those of the cubic M_0 compounds [ 5 ].

References

1. G. Adachi, M. Ishihara and J. Shiokawa, J. Less-comm. Metals, 32, 175 (1973).

2. Y.N. Venevtsev, R. Ch. Bychurin, S.Y. Stefanovich, V.V. Chechkin and

Kh M. Kurbanov, Ferroelactries, 45, 211 (1982).

3. C. Cascales and I. Rasines, Mater. Chem. Phys., 10, 199 (1984).

4. R.D. Shannon, Acta CrySt., A32, 751 (1976).

5. N.T. McDevitt, A.D. Davidson, J. Opt. Soc. Amer., 56, 636 (1966). P 1 B lo

TABLE I.- Color and X-ray diffraction data for the oxides with standard deviations in parentheses.

0 Color V/Ad D /Mgm"3 M ao/A

Y white 10.7546(1) 1243.89(3) 5.69 Pr green 11.0364(3) 1344.2(1) 6.03 Nd pale purple 10.9917(3) 1327.9(1) 6.15 Sm pale yellow 10.9202(5) 1302.2(2) 5.38 Eu white 10.8860(6) 1290.0(2) 6.46 Gd white 10.8557(4) 1279.3(1) 6.50 Tb gy y brown 10.8222(5) 1267.5(2) 6.59 Dy white 10.7872(4) 1255.2(1) 5.31 Ho pale yellcw 10.7558(3) 1244.3(1) 5.91 Er pale pink 10.7291(4) 1235.0(1) 7.00 Tra gray 10.7003(4) 1225.1(1) 7.06 Yb gy y brown 10.6746(4) 1216.3(1) 7.16 lu gy y brown 10.6557(3) 1209.9(1) 7.27

-1 TABLE II.- I.R. Absorption bands frequencies (cm ) for the oxides M.Sb 0 .

Ya 690 600 560 460 380 325 305 ?rb>C 685 585 540 425 _ 290 _ Nd° 685 585 540 425 365 305 2". 5 Sm 685 590 545 450 370 310 275 Eu 685 590 545 450 370 315 230 Gd 685 595 550 450 375 320 290 Tb 690 600 550 455 375 320 295 Dy 690 600 555 455 380 325 295 Ho 690 600 565 455 380 325 295 Er 695 600 565 465 380 325 295 Tm 695 500 565 465 380 325 295 Yb 695 600 565 475 385 33- 300 Lu 635 300 565 475 385 335 305 a one more absorption band at 270 cm" . -1 b further bands at 1005, 975 955, 900, 885, 855 and 835 cm c two additional absorption bands at 925 and 490 cm" . PI BU

A MOSSBAUER- AND X-RAY INVESTIGATION OF Zn-CONTAINING FAYALITE AND Fe-CONTAINING WILLEMITE T. Ericsson and A. Filippidis Department of Mineralogy and Petrology, Institute of Geology, University of Uppsala, Box 555, S-75122 Uppsala, Sweden

Five synthetic samples of the system Fe2SiO^-Zn2SiO^ were produced from oxide mixtures in an arc furnace using argon as protective gas. The melted products were homogenized at 1000°C in vacuum for two weeks (1). The material was crushed and small amounts of magnetic oxides (magnetite or franklinite) were removed under aceton using a hand magnet. Characterization of the obtained phases were performed using petrographic microscope-, X-ray diffraction- and microprobe techniques. The Zn-i'ich sanples (A, B, C) crystallized with willenute structure, while Fe-rich sample E crystallized with the more dense fayalite structure (Table 1). Sample D showed the coexistence of both phases, finely grained and inter- growthed (crystal size -3 pm) with each other, which excluded an individual microprobe analysis of the two phases.

Table 1. Composition and cell parameters of the investigated samples.

V(A3) Sample and Comp. Symm. a(A) b(A) c(A) form, unit

a A: Fe0.10Z'"',.90Si04 hex 13.943(1) 9.324(1) 87.21 B: Fe0.5CZn1.50Si04 -"- 13.955(1) 9.366(1) 87.76 C: Fe0.85Zn1.15SiO4 13.962(2) 9.416(3) 88.31 Di: . b -"- 13.980(4) 9.429(6) 88.66 re 31U c D l.35^'O.65 4 ortho 4.801(1) 10.471(1) 6.093(1) 76.58 E: Fe1.87Zn0.13Si04d -"- 4.820(2) 10.479(3) 6.083(3) 76.58

F: Fe2Si04 -'• 4.821(1) 10.478(2) 6.092(2) 76.93 a) Wiliemite structure, trigonal 83 , Int. Tabl. no. 148, hexagonal setting, Z = 18. b) Average composition of the coexisting two theses. c) Fayalite structure, ori:iorhombic Pbnm, Int. Tabl. no. 62, Z = 4. d) Synthetic fayalite earlier produced in our laboratory (2).

The Mossbauer data were obtained from powder absorbers at room- and elevated temperatures. For all sanples only one quadrupole split doublet, bEq,was obtained at -295 K. The fayalite structure contains two metal octahedral posi- P 1 B 11

tions, Ml having an inversion centre and M2 having a symmetry plane. The Ml site is smaller and more distorted: = 2.159 A compared to = 2.178 A (3). The willemite structure contains two tetrahedral_ metal positions of nearly equal size: = 1.95 A compared to = 1.96 A (4). The centroid shifts (CS; relative to a-Fe at -295 K) were -1.03 mm/s for samples A, B and C, but -1.13 mm/s for Dg and E at room temperature; a difference reflecting the different coordination numbers. At elevated temperatures (>500 K) iron gives significantly different Mbssbauer parameters at the two positions, with a similar behaviour for both studied structures (Table 2). The fayalite Ml position gives a lower AEn as well as CS compared to the M2 position, in agreement with size- and distortion-considerations (2). In the willemite-structure samples, CS is about equal for the two positions even at elevated temperatures, reflecting the equal sires of the two positions. In the fayalite structure the scatter in nearest neighbours is 0.105 A for Ml but 0.219 A for the M2 position having the largest AEQ. In the willemite structure the scatter is 0.007 A for Ml compared to 0.038 A for the M2 position (4). Thus we assign the larger AEQ to the M2 position (2n(2)tetrahedron in ref. 4) in the willemite structure; as well. Zn shows a preference for the Ml position in both structures. The Kp-value for the exchange reaction

Zn2+(M1) + Fe2+(M2) t Fe2+(M1) + Zn2+(M2) is defined as

XFe(Ml) • [l-XFs(M2)j/{XFe(M2) where X denotes the occupancies (Table 2). It is noteworthy that the Kp-values are rather similar for both structures inspite of the fact that the two positions are similar in size in the willemite unit but different in the fayalite structure. Thus it seems as if crystal field stabilization energies play the major rfile for the preferences, at least in the willemite structure. PI B 11

Table 2. Mossbauer parameters at different temperatures.

Ml M2 Sample T(K) CS W Int CS w Int h A -295 1.04 2.87 0.36 - 1.04a L.87 0.36 - - hex 695 0.77 1.38 0.34 0.44 0.75 2.24 0.49 0.56 0.78 B -295 1.03 2.84 0.35 1.03 2.84 0.35 _ hex 687 0.75 1.45 0.36 0.35 0.75 2.23 0.36 0.65 0.44

C -295 1.03 2.77 0.40 _ 1.03 2.77 0.40 _ hex 592 0.73 1.45 0.39 0.36 0.76 2.19 0.39 0.64 0.37

D2 -295 1.12 2.83 0.35 - 1.12 2.83 0.35 . - ortho 501 0.91 2.23 0.48 0.46 1.03 2.55 0.48 0.54 0.69b 674 0.81 1.74 0.44 0.43 0.89 2.26 0,44 0.57 0.47b

F -295 1.13 2.78 0.30 • 1.18 2.88 0.26 - - ortho 519 0.95 2.16 0.29 0.48 1.03 2.59 0.29 0.52 0.53b 629 0.83 1.84 0.35 0.48 0.92 2.31 0.3£ 0.52 0.53b c F 540 0.91 1.99 0.24 0.49 0.98 2.47 0.24 0.51 - ortho 673 0.83 1.76 0.24 0.49 0.90 2.25 p_.24 0.51 - a) Constrained values are underlined. b) Corrected for Int(Ml)/Int(M2) = 0.49/0.51 in Fe2Si04< c) Synthetic fayaiite earlier produced in our laboratory (2).

References

(1) A.G. Nord, H. Amersten and A. Filippidis, Am. Min. 67, 1206 (1982) (2) H. Annersten, T. Ericsson and A. Fili.pidis, Am. Min. 62, 1212 (1382) (3) I.D. Birle, G.V. Gibbs, P.B. Moore and J.V. Smith, Am. Min. 53_, 807 (1968)

(4) K.H. Klaska, I.C. Eck and D. Pohl, Acta Cryst. B34, 3324 U978) PI B 12

SOLID STATE CHEMICAL MODEL FOR THE SOLUBILITY BEHAVIOUR OF Co- Mn-CARBONATE SOLID SOLUTIONS H. Gams.iager. A. Fluch und W. Engelmann Institut fur Physikalische Chemie der Montanuniversitat, Franz-Josef-StraBe 18, A 8700 Leoben

Introduction: Recently the solubilities of the pure neutral transition metal carbonates, rhodochrosite, MnCO,, siderite, FeCO , spherocobaltite, CoCO , NiCO,, CuCO, and smithsonite, ZnCO-, have been studied by a sensitive potentiometric method (l). However, experimental data for solid carbonate solutions are still rather scarce. In fact, the various models of the solubility behaviour of carbonates in binary solid solutions so far proposed, are generally based on the properties of the end-members and some additional theoretical considerations (2,3).

Experimental: Solid solutions of CoCO,-MnCO- were prepared from aqueous acidic solutions of the respective chlorides by an auto- clave1 method (h). X-Ray powder patterns of the crystals obtained were recorded, and used to select samples which presumably consisted of a single phase. These products were chemically analyzed and further investigated. Using the lattice constants of the rhombohedral carbonates the distances from the centre of the cation to that of the anion, r, w^re calculated, and in the whole compositional range shown to obey Vegard's rule closely. Although the Co- Mn-carbonates studied were thus part of a homogeneous solid solution series, they dissolves congruently up to 80 % completion. If the solids do not change composition during the dissolution reaction (A)

C Mn C0 + 2+ 2+ °(l-x) j£ 3(s) * 2H »=—(l-x)Co + xMn + C02(g) + HgO (A) they can be treated thermodynamieally as one-component phases (5), the respective equilibrium constants being defined by equation (B),

2+ (l a) 2+ JC log KeqU) - log{[Co ] ~ [Mn ] pC0 + 2 pH} (B) P 1 B 12

Typical diasolution experiments are given, in the Figure.

-2.0

o o

i -3.o

o°

U.7 5.1 PH

Figure: Solubility constant of CoM ,Mn CO,.(50°C,ionic strength, 3 I a 1.0 mol dm (Ka)ClO,, Pco2 0.82 atm, x = 0.39). The straight line has been calculated with

log{[Co2+](l~x)[Mn2+]V 7.12.

At fixed composition in general log K r,..\ % remained constant vithin + 0.05 units and plots of versus pH fell on straight lines of the theoretical slope -2.0+0.1. P 1 B 12

Discussion: The equilibrium constants, log K / •. were used to calculate the excess free enthalpy of mixing,A . Ge . All data mix m were consistent vith a regular model according to equation (C)

A •_G" / RT = a xMTi.n (1-x ) (C) mix m MnCO, Mn(JO_ The parameter, a, agreed reasonably well vith that predicted by Lippmann's (6) simple electrostatic model. However, unlike to this model Co-Mn—carbonates reacted with fixed composition. Pre- liminary experiments with Ca/, sMn CO, samples indicate that these solid solutions dissolve congruently as well, although the solubilities of calcite and rhodochrosite differ by several orders of magnitude. This is an important practical aspect for leaching processes because it means that at low temperatures the dissolution behaviour of minerals may be quite different to that predicted from the thermodynamic data of the pur? end- members .

(1) F. Reiterer, Dissertation, M.U. Leoben 1980 (2) F.C.M. Driessens & R.M.H. Verbeeck, Ber. Bunsenges. Phys. Chem. 8^, 713 (1981) (3) F. Lippmann, Bull. Mineral. 105, 273 (1982) (k) H. Gamsjager & ?. Reiterer, Environment International 2, U19 (1979) (5) D.C. Thorstenson & L.N. Plummer, Amer. J. Sci. 277, 1203 (1977) (6) F. I.ipprcann, N. J"b. Miner. Abh. 139, 1 (1980)

'J. Gamsjager, Iastitut fur Physikalische Chemie der Montan- universitat, Franz-Josef-Strafle 18, A 8700 Leoben. P 2 A 1 r THE TERNARY SYSTEM ERBIUM-BORON-CARBON. ISOTHERMAL SECTION AT 1500°C. J. Bauer Laboratoire de Mgtallurgie et Physico-Chimie des MatSriaux, associe au C.N.R.S. n° 254, I.U.S.A., 20, avenue des Buttes de Coesmes, 35043 Retmes Cedex.

Introduction The existence of ternary borocarbides of all the rare-earth metals has been demonstrated during recent years. Ternary phase diagrams at high temperatures were proposed for Y-B-C (1), Gd-B-C (2), Eu-B-C (3) and Ho-B-C (4) systems. No relevant reference was found in a literature survey on the ternary Er-B-C system. The purpose of this investigation was to fill up this gap of information. Experimental procedure Mixtures of powders of commercially available elements of high purity (at least 99,9%) were compacted in stainless steel dies without the use of binders or lubricants. The pellets (1 g.) were arc-melted under purified argon on a water- cooled copper hearth. After melting the samples were crushed, one part was examined in the arc-melted and quenched state, the other part was wrapped in thin tantalum foil, placed in a molybdenum crucible and annealed under purified argon during 100 hours at 1500° C using a 25 kW 500 kHz high-frequency furnace. All samples were examined by micrography and X-ray technics. The ternary system The phase equilibria were established in the isothermal section at 1500" C. In addition to the already known phases (i.e. ErB2, ErBi., ErBu , ErBs6 , B13C2, ErC . ErisCi?, ErC2, ErB2C and ErBzCa) six new ternary compounds ErBC, EraBCs, ErjB2C3,

ErsB2Cs, ErsB2C6 and Eri5B2Ci7 were identified. ErBC and Er2BC3 present both low and high temperature modifications. In the binary Er-C system, a new carbide at the composition of Er5Cs was observed. The Er-B-C system, like the other Rare-Earth-Boron-Carbon systems, can be devi- ded into two parts : borides (carboborides) and carbides (borocarbides). The main difference -from a crystal chemical point of view- between these two classes is that the boride and carboboride lattices are formed by the boron and carbon atoms (three-aod two-dimensional covalent-bonded frameworks) and the metal atoms are located in the insterstitial positions, whereas the carbide and borocarbide lattices are formed by the metal atoms with the nonmetal atoms occupying the insterstices. Borides and carboborides appear when the Nbmetal/Metal ratio is > 2, carbides and borocarbides are situated in the more metal rich part of the diagram.

Three carboborides were identified in the ternary system. ErBzC (5) and P 1 B 12

Discussion: The equilibrium constants, log K , y. were used calculate the excess free enthalpy of mixing,A . Gex. All mix m •re consistent with a regular model according to equat:

" / RT

The j. -ter, a, agreed reasonably well with thf edicted by Lippmai. '6) simple electrostatic model. Hov unlike to this modt Mn-carbonates reacted with fixp jiposition. Pre- liminary ex ents with Ca/._ ,Mn CO, sas indicate that I J.—x) x j these solid t> ons dissolve congruentl- well, although the solubilities oi ite and rhodochrosiJ .ffer by several orders of magnitu 'his is an impor* practical aspect for leaching processes se it means at low temperatures the dissolution behav of miner jiay be quite different to that predicted from the -nody J data of the pure end- members .

(1) F. Reiterer, Dissertat" '. Leoben I960 (2) F.C.M. Driessens & R \ ck, Ber. Bunsenges. Phys. Chem. 8J., 713 (198'1 (3) F. Lippmann, Bul3 .ieral. 10£ ^ (1982) (k) H. Gamsjager & jiterer, Envir t International 2_, U19 (1979) (5) D.C. Thorst . & L.N. Plummer, Amer Sci. 277, 1203 (1977) (6) F. Lippr K. Jb. Miner. Abh. 13£, 1 (1.

E. Gamsjager, Institut fur Physikalische Chemie der Montan- universitat, Franz-Josef-Strafie 18, A 8700 Leoben. P 2 A 1

ErBzC2 (6) are typical "sandwich"-c"mpouads containing two-dimensional infinitesimal boron-carbon networks. ErBC contains infinitesimal zig-zag boron chains (as found in monoborides) with carbon atoms strongly bonded to it. The mutual arrangement of these chains gives arise to YBC (1), UBC (7) and ThBC (8) structures. ErBC shows at high temperatures (arc melted and quenched) the YBC-type structure which transforms after annealing at 1500° C to a new type of tetragonal symmetry but still unknown structure exhibiting a very long (46,6 A) c-axis.

On the isoplethe ErB2 - ErC2, separating borocarbides and carboborides(appears

on the carbon rich side E^BCj (topochemically ErBC + ErC2). The high temperature form of this compound shows I-centered tetragonal symmetry where the a-axis is nearly equal and the c-axis is twice as long as in ErC2. The low temperature form has a primitive tetragonal lattice and a doubled c-parameter. A single- crystal study on the high temperature form is actually in work. ErsB2C6 is isostructural with the superconducting lanthanum borocarbide LasB2C6 (9). Filling che octahedral voids on the c-axis of this structure with carbon atoms instead of carbon pairs gives ErsB2Cs. Substitution of carbons atoms by boron atoms in the ErisCi9 (10) structure gives a new ternary compound EruBzCn which is stable up to the melting point. lir3B2Ca shows a very complex powder pattern which could not be indexed. This phase is presumingly related to the monoclinic Tb.3B2C3 (11), containing C-B-B-C groups. Lattice parameters, drawings of the structures and the isothermal section of the ternary phase diagram will be presented at the Conference.

(1) J. Bauer, H. Nowotny, Monatsh. Chem., 102, 1129-1145 (1971) (2) P.K. Smith, P.W. Gilles, J. Inorg. Nucl. Chem., 29, 375-382 (1967) (3) K.A. Schwetz, M. Hor, J. Bauer, Ceramurgia Int., 5/3), 105-109 (1979) (4) J. Bauer, F. Vennegues, J.L. Vergneau, International Rare-Earth Conference, Zurich, Switzerland (1985) (5) J. Bauer, J. Debuigne, J. Inorg. Chem., 37_, 2473-2476 (1975) (6) J. Bauer, 0. Bars, Acta Cryst., B 36, 1540-1544 (1980) (7) L.E. Toth, H. Nowotny, F. Benesovsky, E. Rudy, Monatsh. Chem., 92(3), 794 (1961) (8) P. Rogl, J. Nucl. Mat., 73, 198 (1978) (9) J. Bauer, 0. Bars, J. Less-Common Met., 95, 267-274 (1983) (10) J. Bauer, J. Less-Coomon Met., 3£, 161-165 (1974) (ID P. Rogl, J. Hucl. Mat., 79, 154-158 (1979)

J. Bauer, Laboratoire de MStallurgie «t Physico-Chimie des MatSriaux, Institut National des Sciences Appliquees, 20, av. des Buttes de Coeanes, 35043 Senses Cedez - FRANCE. Tel (99) 36.48.30 poste 463. P 2 A 2

SIMPLE CHEMICAL WISHING : A MODEL TO ETPTrA™ THE MODULATED STRUCTURES WHICH APPEAR DURING THE CRTSTALLIZAHON OF SPUTTERED IRON-CAHBtM AMORPHOUS ALLOTS E. Bauer-Gross*, 6. Le Caer, C. Fraatz Laboratoire de MStallurgie (U.A. 159), Ecole des Mines - Pare de Saurupt 54042 NANCT-Cedex (France)

INTRODUCTION It is veil established that crystallization of amorphous metallic alloys can produce phases which are not observed by conventional methods of elaboration. It is che case for sputtered iron-carbon amorphous alloys F«IQQ« ** which crystallize in one (for x - 28 at.Z) or two steps [1,2] : we have already shown that a short-range ordered Fe_C_ phase [3,4] appears during the early state of crystallization when 28 < x < 50. The purpose of this paper is to give, by means of X-ray analysis, T.E.M. and Mossbauer spectroscopy, a structural identification of the carbides, called MSJJ, formed during the polymorphous crystallization (for x * 28) and during the second step of the crystallization for 18 < x < 32.

EXPERIMENTAL RESULTS Experimental procedure have been described elsewhere [2]. Electron micrographs concerning the polymorphous crystallization or the second step of the crystallization show very faulted microstruetures.

Electron diffraction patterns indicate that these microstructures are essentially formed from trigonal prismatic (T.P.) sheets as it is usually found in cementite or Hlgg carbides. But when the electron beam is perpendicular to the T.P. sheets, it is difficult to identify these carbides without ambiguity and it is necessary to obtain other informationsin the reciprocal space.

Figure 1 shews that, perpendicular to the T.P. sheets, there are always diffusion lines along which one can observe an evolution of the diffraction spots with varying carbon content. These modulated structures can be described by the theory of A. Janner, T. Janssen and P. de Wolff [5]. The diffraction patterns obtained during the polymorphous crystallization of the Fe-.CLgalloy indicates a commensurable carbide which is an Hagg type carbide with a random distribution of planar faults parallel to the T.P. sheets (figure la). Diffraction pattern of figure lb obtained from a Fe-QC_alloy indicates aa incaaaensurable carbide whose structure is derived from the Higg carbide structure and the diffraction pattern of figure Ic obtained frost a Fe-.C-. alloy indicates that there is an intergrowth of a likely faulted Fe.C and of the previous incoancnsurable carbides.

For the composition Fe^C,-, this inci.—umurable carbide has been investiga- P 2 A 2

ted by Moaabauer apectroacopy f6 ). The comparison with the spectrum of a synthetic

Higg carbide T71 indicatea that there are more FerXj.aud less FeI]; iron sites, the fraction of Fe_ sites being approximately the same. Fejjj belongs to four T.T.,

FeT and Fe to two T.P. (figure 2). For amorphous alloys with carbon content in the range 18-32 at.Z, the Moasbauer spectra and the broadening of the X-ray lines reveal the continuous character of the structure evolution from FegC to FejCj and beyond [».•»••««...» Figure 1 Electron diffraction patterns and corresponding key diagrams of S-r-r obtained on

(a) Fe72C28 conmensurate Fe c structure 5 2

Fe7OC3O inc ommensu ra te structure,

bl IA A V V V A A A. V V V A A A V V V

(a) (b) (c) Figure 2 Schematic illustrations of linkage of T.P. sheets in (a) Fe-C (Pnaa) ; (b) FesC7 and (c) hypothetical "Fe.C" which can deduce from each other by S.CITT * Only one half of the prisms are drawn. Roman numbers indicate the different iron sites Fe_, Fe._, Fe---. DISCUSSION AND CONCLUSION From these results, we can conclude that during the polymorphous crystallization or during the second step of the crystallization, T.P. sheets are formed. Depending on the local composition of the amorphous matrix and of ita homogeneity, these sheets are linked together to form commenaurable carbidea (Fe.C (fig. 2a), P 2 A 2

Fe.C. (Fig. 2b)) or incommensurable carbides or an intergrowth of these two types of carbides. The structure of Fe,C has been described by Andersson et al [83 as a result of a periodic h.c.p. lattice of iron atoms. This twinning, called simple chemical twinning (S.C.T.), creates prismatic interstices available for carbon atoms every four (11.2) planes. The modification of the twinning periodicity changes the way by which the T.P. sheets are linked and therefore the composition. If the S.C.T. is repeated with the sequence (4,3,4,3,..), we obtain FecC2 acc°rding to the formal reaction

In our case, Mossbauer results indicate that we can continue to transform

to FeI_I sites, reaching an hypothetical Fe-C structure (figure 2c) when all Fe_x sites are removed (sequence 3,3,3,3,..) which electron diffraction results indicate that it is not necessary to repeat S.C.T. regularly since we observe incommensurable carbides. H.R.E.M. would be a good technique to continue this study because in tempered martensites, internal planar defects of carbides have already been observed by A. Koreeda and K. Shimizu [9] and an intergrowth of Fe.C, Fe.C- and higher carbi-

des Fe. +,C has been found by S. Nagakura et al. CIO] ; an intergrowth o£ Fe_C and Fe.C. formed during the disproportionation of CO has also been reported by M. Audier et al. CM].

REFERENCES [1] E. Bauer-Grosse, C. Frantz, G. Le Caer and N. Heiman, J. non Cryst. Solids, 44, 277 (1983) [2] E. Bauer-Grosse, C. Frantz, RQ5, Wurzburg, 3-7 September 1984 [3] E. Bauer-Grosse, J.P. Morniroli, G. Le Caer, C. Frantz, Acta Met., 29, 1983 (1981) [4] E. Bauer-Grosse, J.P. Morniroli, G. Le Caer, C. Frantz, J. Physique, C9, 285 (1982) [5] A. Janner, T. Janssen and P. de Wolff, Acta Cryst., A 39, 658 (1983) [6] G. Le Caer, B. Lemius, J. WeIfringer, E. Bauer-Grosse, J.M. Dubois, M.R.S.- Europe , Strasbourg, Symposium II (1984) [7] G. Le Caer, J.M. Dubois, J.P. Senateur, J. Solid State Chenu, 22., 19 (1976) [8] B.G. Hyde, S. Andersson, M. Bakker, CM. Plug, M.O'Keepfe, Prog. Solid St. Chem., J2_, 273 (1979) [9] A. Koreeda, K. Shimizu, Proc. Fifth Int. Conf. H.V.E.M., Kyoto, 611 (1977) CIO] S. Nagakura, T. Suzuki and M. Kusunoki, Trans. J.I.M., 22_, n* 10, 699 (1981) Cll] M. Audier, P. Bowen and W. Jones, J. of Crystal Growth, 64, 291 (1983)

E. Bauer-Grosse. Laboratoire de Metallurgie (U.A. 159), Eeola des Mines - Pare de Saunipt, 5^CA2 RANCY-Cedex (Prance) P 2 A 3 TERNARY LANTHANOID MANGANESE CARBIDES WITH FILLED BaCd,, AND 11 Th-Zn17 TYPE STRUCTURES G. Block and W. Jeitschko Anorganisch-Chemisches Institut, Oniversitat Munster, D-44OO Munster, West Germany.

The investigations of ternary systems of the lanthanoids with manganese and carbon have resulted in several new ternary compounds. They were prepared by melting the elemental components in a hit,h—frequency furnace with subsequent annealing in closed silica tubes. The crystal structures of two new compounds were determined from single-crystal X-ray data. LaMn.,C_ crystallizes in space group l4./amd with the lattice constants a = 10.4134(4) S, c = 6.7293(4) 8, V « 729.72(5) 83, and 2 = 4 formula units per cell. The final residual is R - 0.013 for 503 F values and 21 variable parameters. The positions of the metal atoms correspond to those of BaCd.. (1). The carbon atoms are located in octahedral voids formed by four Mn (1.91 8) and two La (2.74 5) atoms. The carbon positions are occupied to only 74.5 +_ 1.4 %. The La atoms have 22 Mn and 4 C neighbors. Only one of the three different Mn atoms has a C contact (Fig. 1). The new compound PrMn.-C, was found to be isotypic with LaMn.-C,

Fig. l. Coordination polyhedra in Ladn C, . P 2 A 3

Th« compound Pr-Mn,_C-_ crystallizes with a rhonbohedral cell, space group R3m with the lattice constants a - 8.8714(7) 8, c - 12.783(2) 8, V « 871.2 83 and Z-3 formula units in the hexagonal cell. It was refined to R - O.O23 for 25 variables and 414 independent structure factors. The structure can be derived from that of Th_Zn,_ (?) with the Pr and Hn positions corresponding to those of Th and Zn. The carbon atoms fill octahedral voids formed by a rectangle of Mn atoms (Mn-C distances of 1.94 S and 1.9S 8) and two Pr atoms at 2.57 S. The ideal composition with all octahedral voids filled is Pr.Mn.-C-. The refinement of the occupancy parameter of the C positions showed that they are filled only to 57 *_ 3 % in the crystal picked for the structure determination. The Pr atoms are Located in coordination polyhedra formed by three C atoms at 2.57 8, 19 Mn atoms (at distances varying between 3.19 and 3.43 8), and one Pr atom at 3.95 8. The four different Mn atoms have coordination numbers 12, 13, and 14 with no or one C neighbor, one, two or three Pr neighbors, and between 9 and 13 Mn neighbors (Fig. 2). La Mn C, is isotypic with Pr,Mn C, .

(1) M.J. Sanderson and N.C. Saenziger, Acta Cry3tallogr. 6_, 627 (1967). (2) E. S. Makarov and S. I. Vinogradov, Sov. Phys. Crystallogr. j_, 499 (1956).

o Fig. 2. Coordination polyhedra in Pr.Mn.^C, . Only one half of the structure (between z » 0 and 1/2) is shown. P 2 A 4

NEW SPUTTERED Nb-Nv FILMS WITH HIGH NITROGEN CONCENTRATIONS : ELABORATION AND PROPERTIES R. CABANEL, J.C. JOUBERT Laboratoire de Genie Physique, ENSIEG (ERA 836 CNRS), BP 46, 38402 SAINT MARTIN d'HERES, FRANCE J. CHAUSSY, J. MAZUER* Centre de Recherches sur les Tres Basses Temperatures, CNRS, BP 166X, 38042 GRENOBLE Cedex, FRANCE (laboratoire associe a l'Universite Scientifique et Medi- cale de Grenoble).

INTRODUCTION Niobium nitrides films have been widely studied in the past. Much attention has been paid to the super-conducting properties of the 6-Nb-N phase. Interest on this phase is not vanishing at present as emphasized by recent investigations on the upper critical field (1) and the use of Nb-N films for Josephson integrated circuits (2). Early work on the influence of nitrogen concentration (3) show that when the ratio x = N/Nb increases up to 1.5, superconductivity disappears and the resistivity at ambient temperature drastically increases. Films with an amorphous structure were prepared elsewhere (4) and they were shown to be of interest to secondary thermometers between liquid helium temperature and ambient temperature. The sensitivity was demonstrated to be good over this total temperature range, but the resistances of such sensors was very high. It seemed to us that it might be of interest to establish quantitative relation between the electronic properties of the films and the nitrogen concentrations obtained under different growth conditions. Our preliminary results are reported in the following. SAMPLE PREPARATION Our films are prepared by D.C. magnetron reactive sputtering of a pure Nb target. The sputtering apparatus was specially designed in order to ensure a close control of the sputtering parameters. Such a control proved to be necessary to prepare films which characteristics were as reproducible as possible. As our aim was to increase the nitrogen concentration, we used a pure nitrogen atmosphere instead of a mixture of nitrogen and argon, as it is commonly used to obtain the face- centered cubic 5-Nb-N phase. The magnetron is supplied by a controlled D.C. current source which gives a steady adjustable current.

Five parameters are of interest : the temperature Ts of the substrate holder, the nitrogen pressure p, the distance D between the target and the substrate, the voltage U and current I applied to the magnetron. Obviously all these parameters are not independent. When the values Ts, d and I are fixed, the sputtering process can then be driven in two ways by controlling either the voltage or the pressure. In either case, we act on the aperture of an electric regulating valve (type RME 010 from BALZERS) to maintain either p or U = constant. During all our experiments, the substrate holder is cooled at Ts ^ 290 K by a water flow. The substrates made of sapphire are pressed on to this support to ensure a good

H J. MAZUER is also with the laboratoire d'Electrotechnique, ENSIEG (LA 255-CNRS). P 2 A 4

I (A) O(cm) 6 A 6 a .6 S • 6

C 100 .2-

0.00 0.05 0.10 JOO . pressure (mbar) deposition rate (A.nvT1) Ftg. j ; Variation of the resistivity Fig, g :• Variations of the resistivity ratio with nitrogen pressure. Experimen- of various NbSx with the deposition tal conditions are given in the insert. rate. The arrows show the decrease of the pressure during the growth process when If is regulated at a constant value. thermal contact ; but we do not control up to now the temperature of the surface of the substrates. That temperature probably is greater than Ts due to radiation of the plasma. EXPERIMENTAL RESULTS AND DISCUSSION A great number of samples have been prepared under different sputtering conditions. As a preliminary criterium, we measure the resistivity ratio (R.R) = R(77 K)/R(300 K) between liquid nitrogen and ambient temperature. One can see on figure 1 that all the samples exhibit a (R.R) > 1, a feature of a semiconductor-like behaviour. (R.R) can vary as much as four orders of magnitude when PN2 is increased from ^ 10-2 to *\. 10"! mbar. The samples were prepared according to the two methods previously described. When p = constant, U only varies by a few percent. It hardly reaches 15 % when the sputtering is operated very near the limit of the plasma stability. On the contrary, when U is constant, the pressure decreases regularly during the process to a value down to 50 % of the initial value. Of course the decrease is a function of the sputtering duration but also of the pressure. On figure 1 the pressure decrease.when significant, is shown by arrows. The corresponding magneton voltage was U ^ -700 V. It is clear that the nitrogen pressure is a very sensitive parameter with respect to R.R. On the contrary no correlation could be observed for R.R. with the voltage U. It is to be noticed that the nitrogen pressure is not the only parameter to drive R.R. The distance D and the current I also interfere. Experiments are in progress to determine their influence. Measurements of the thickness of some samples were made with a Talysurf. The thickness varies between .45 and 2.6 urn. It is thus possible to know the resistivities and also the deposition rate. The resistivities at ambient temperature vary in the same way as R.R. Figure 2 emphasizes the influence of this deposition rate on the resistivities. High R.R. values are obtained at low rate which agrees with the increase of the probability for the niobium atoms to collide with the nitrogen molecules or ions. This probably results in a higher nitrogen concentration in the films leading to the large R.R. and resistivities. This high P 2 A 4

nitrogen concentration in NbNx films is confirmed by measurement of back-scattering of ct-particules at 2 Mev. A value of x = 2.7 corresponding to 73 % nitrogen at. and 27 i niobium at. is observed for a film with a R.RD n in4 and a resistivity at 300 K of 3.5 ft. cm. To get information on the,composition and structure of our films, we studied the density and X-rays diffraction. The densities were obtained from the differences between the weights of the substrates before .00 40.00 60.00 80.00 100.00 and after sputtering. As previously shown 29 (degrees) (3), the density of sputtered NbNx decreases from 8.5 g.cm-3 at x = 0 to ^ 6 g.cm~3 at Fig. 3 : X-Eay diffraatogram (FeKa) x = 1.5 We measured densities as low as of a NbNx sample having RR ^ 10^ ^ 3.4 g.cm~3. Such a low value can only and a resistivity ^6.4 Q.am. be explained by a very high nitrogen concentration as it has been previously demonstrated by a-particules back-scattering (x = 2.7). We also performed X-rays diffractograms which all exhibit the same features as those we show on figure 3.

One recognizesa large peak of diffusion (from 20 to 45°) characteristic of an amorphous structure and a narrow peak at 52.5°corresponding to the 200 reflection of the f.c.c. 8 NbN. The presence of this only one observed reflection is typical of a marked orientation effect of the growth process. Preliminary measurements of the optical reflectivity of some samples did not reveal any reflectance edge between .3 and 2.7 \m, by opposition of what has been observed on TiNx, HfNx and ZrNx (5.6). As expected we noted that the reflectivity is decreasing with the electrical conductivity. Further experiments are in progress to determine the variations with temperature of the conductivity and Hall coefficient between 1 K and 300 K. REFERENCES (1) M. ASHKIN,, J.R. GAVALER, J. GREGGI , M. DECROUX , J. Appl. Phys. 5£, 1044 (1984). (2) J.C. VILLEGIER,. J.C. VELER, IEEE Trans. Mag. MAG-19, 946 (1983). (3) A. AUBERT, J. SPITZ, Le vide, .U£. ! (1975). (4) J. CHEVALIER, J. BAIXERAS, P. ANDRO, Revue de Physique Appliquee, 14, 663 (1979). ~~ (5) B. KARLSSON, R.P. SHIMSHOCK, B.O. SERAPHIN, J.C. HAYGARTH, Solar Energy Mater., 7., 401 (1983). (6) P. GRAVIER, G. CHASSAING, A. AUBERT, Proc. Vile Conference on solid compounds of transition elements, Grenoble (1982).

R. CABANEL, Laboratoire de Genie Physique, ENSIEG (ERA 836), BP 46, 38402 SAINT MARTIN O'HERES, FRANCE P 2 A 5

UNUSUAL ELECTRICAL BEHAVIOUR ASSOCIATED WITH THE SEMICONDUCTOR- METAL PEASE TRANSITION-IN MIXED PHASE VANADIUM OXIDES

A. Clark, M.C. Lovell and D.L. Tunnicliffe

School of Electrical Engineering- and Science Royal Military College of Science Shrivenham, Swindon, Wiltshire SN6 8LA U.K.

We present evidence ->f electrical switching in a range of sputtered thin film vanadium oxides associated with the semiconductor-semiconductor phase transition (SSPT). The influence of the mixed phase nature of the film confirmed by x-ray diffraction is important.

Figure 1 shows a typical log (current) V temperature plot for a mixed phase of V5O13 and VO2 in the temperature range 70-360K. Two first-order phase transitions are evident, the first at about 170K due to vg0j3 and

the second at 340K due to V02.

Figure 2 is a log (current) V temperature plot for this sample showing the effect of varying the applied voltage. The transition decreases from 170K to 140K as the voltage is increased from 0.5V to 6V. The quoted transition temperature for polycrystalline thin film VgOjj samples vary widely (1). This variation seems therefore to depend not only on sample preparation but on the voltage used to measure the electrical transition.

In contrast to the previous sample, in a mixed phase of VgOj^ and VO2 there is a mutual effect between the compounds. For VgO^, we observe a transition which is semiconductor - semiconductor in nature (rather than semiconductor-metal) at about- 170K because of the influence of

semiconducting V02 below its transition temperature. The magnitude of the VO2 transition is affected by the presence of the VgO^!2). This sample exhibited a switching phenomena illustrated in figure 3. The switching takes place between a high resistance state (HRS) below the transition temperature of VgO^j and a "supercooled" low resistance state (LBS) which is an extension of the high temperature semiconducting state, derived in this case from the presence of VO2.P). This phenomenon is characteristic of non-linear systems far from equilibrium such as the tunnel diode. The observed instability is associated with such behaviour. P 2 A 5 Fig. 1. Current - temperature plot for a mixed phase of and VO2.

Pig. 1.

-11- -

140 160 180 200 "'300 320 340 360 T(K)

Fig. 2. Current - temperature plot showing voltage dependence in VgO Fig. 3. Switching phenomena observed in

Fig. 2. Fia. 3.

80 100 120 140 160 180 TIK) P 2 A 5

This sample also shows a voltage dependence of the transition temperature analogous to our observations of

An explanation of the switching behaviour has been given by analogy with the ballast resistor-(4), The observed voltage dependence can be explained by reference to figure 4, the Inter-relation of the applied voltage and the temperature dependent resistance. In the neighbourhood of the SSPT graphs of rate of heating v temperature for two different applied

voltages V£>V1 are shown. Thus the Intersection of the high temperature and low temperature regions of each curve which define the transition, occurs at different temperatures. The higher voltage shifts the transition teimierature to a lower value.

Figure 4.

RATE OF HEATING

XX TEMPERATURE

Fig. 4 Influence of Joule heating on phase transition temperature.

References:

(1) J.M. Honig and L.L. van Zandt, Ann. Rev. Mat. Sci. 5_, 225 (1975) and references therein.

.£2) C.H. Griffiths and H.K. Eastwood, J. Appl. Phya. 45_, 2201 (1974) .

(3) A. Clark, M.C. Lovell and D.L. Tunnicliffe, submitted to J. Phys. C (Solid State Physics).

(4) R. Landauer, Phys. Today, Nov. 1983, P.23.

M.C. Lovell, The Royal Military College of Science Shrivenham, Swlndon, Wiltshire, SN6 SLA, O.K. P 2 A 6

THE INFLUENCE OF STOICHIOMETRIC DEVIATIONS ON ELECTRICAL PROPERTIES OF TITANIUM OXIDES H. Gruber and E. Krautz Institut fiir FestkBrperphysik, Technical University, Petersgasse 16 A 8010 Graz

Among the different binary systems of transition metals the system titanium-oxygen has found special interest because different large and small phases of homogeneity with different crystal structure exist which allow to investigate in a larger scale the influence of deviations from stoichiometric compositions for different titanium oxides TiOx with 0

o as u> o O/Tl RATIO X—

FIG. la,b- Specific resistivity dependence on (b/Ti] ratio x P 2 A 6 sing concentrationsxfor the titanium oxide phase (0

Ti?0 caused by an oxygen layer ordering in the substructures. In the phase of TiO again an increase in the resistivity occurs more stronger for 77 K than for 273 K. Fig. 1b shows the influence of deviations from stoichiometric compositions in much smaller phases of homogeneity for the corundum phase Ti203+6 at 293 K and the rutil phase TiO2_6 at 273 K. Here the highest specific resistivity of these semiconductors are reached at the exact stoichiometric compositions and the lowest temperatures. At 4.2 K the specific resistivity surmounts values of 6 13 >10 i2.cm for Ti2O3 (5, 6, 7) and >10 i2cm for TiO2.

O/H RATIO «-

FIG. __. 2_.._.__ Hall coefficien. t R H FIG. 3 Thermopower dependence dependence on jp/Tf] ratio on [O/Ti] ratio

In Fig. 2 our measurements of the Hall coefficient RH are shown for the hep and fee phases of TiOx with x<1.25. Just at the phase boundary RH changes its sign being positive for the hep phase and negative for the fee phase. Ru gets a positive maximum near Ti^O and a negative maximum inbetween the hep and fee phase. The changes of the thermopower related to copper at 273 K are shown in Fig. 3 for the hep and fee phase of TiOx with x<1.3. With increasing solution of oxygen in pure titanium a change of the thermopower from positive values to small negative ones takes place in the region x<0.1 and a continuous increase to negative values *-13 pV/K in the fee phase of TiOx 0.7

Fig. 3. Highest positive values of thermopower are observed for

Ti203 up to +400 pV/K at 293 K. The measurements of the relative magnetoresistance &.Q^/QQ show remarkable differences for the different TiOx phases in dependence on mangetic induction and temperature as shown in the Kohler diagrams of Fig. 4a and 4b.

FIG. 4a,b KohTer diagrams for relative magnetoresistance dependence on magnetic induction The positive thermopower for pure a-Ti indicates predominant electrical conductivity by holes whereas the negative Hall constant is exhibited by the much higher mobility of the electrons in the s-band. With increasing oxygen content in the binary system titanium-oxygen the mechanism of electrical conductivity depends on the increase of polarization effects of the charge carriers in the lattice.

References (1) P. Ehrlich, 2. anorg. Chem. 247, 53 (1941) (2) E. Krautz, Werkstoffe der EliTFrotechnik, TH Aachen 1953 (3) H. Gruber and E. Krautz, phys. stat. sol. (a) 6S, 287 (1982) (4) H. Gruber and E. Krautz, phys. stat. sol. (a) 75", 511 (1983) (5) C.N.R. Rao and K. Rao, Phase Transitions in SoTTds, McGraw-Hill Publ. Co., New York 1978 (6) J.B. Goodenough, Progr. Solid State Chem., Vol. 5, Ed. H.Reiss, Pergamon Press, Oxford 1971 (7) N.F. Mott, Metal-Insulator Transitions, Taylor & Francis, Ltd., London 1974 P 2 A 7

FCC OXYCAHBIDE PHASE OF SCANDIUM AND OF YTTRIUM: COMPOSITION LIMITS P. Karen. V. Brozek, B. H6dek Department of Inorganic Chemistry Prague Institute of Chemical Technology, 166 28 Prague

Scandium and yttrium metals are known to possess the hop structure. The ccp arrangement of the metal atoms, however, can be stabilized if the octahedral vacancies are at least in part occupied by nonmetal atoms of a suitable size and valency. Carbon stabilizes the fee structure by forming disordered ScCC-Di „) or P o \ x j.—x YCC-Di „) ' where x is 0.3 * 0.6. The presence of carbon and oxygen in the equiatomic ratio leads to carbide-oxides, Sc~OC ^' and Y-OC^' '. A variable composition of Sc(Q,(J) has been suggested^' ''; however, the existence of a homogeneity region in the ternary system where a single Sc(0,CX3) or YCO.CJU) phase occurs can be expected. Preparation of Starting Mixtures of Sesquioxide, Metal, and Car- bon and Their Thermal Treatment Pellets of the homogenized and ground mixtures were sintered in a ZrB, crucible in a SiO2 ampoule under Ar (10 kPa) (X •»• C mixtures were treated in a vacuum of 10"^ Pa). For yttrium, the temperature was 900 or 1200°C, reaction time 24-48 h. For Sc, the temperature was 900°C, and the equilibrium composition was obtained across the sample bulk within 750 h. Analysis of Reaction Products The Y and bonded C content of the samples from the Y-O-C system was determined by elemental analysis, the ^o^v ^» an^ "^2 content was determined by X-ray diffraction analysis. The composition of the fee phase was established by balance. Because of the great influence of the chemical analysis errors on the final result, the combined diffractographxc and hydrolysis method of analysis was applied to the SC-OJ-C system" . This approach is based on the fact that the intensity ratio of pairs of X-ray reflections for odd and even hkl. vary considerably with the Sc(0,Cj3) phase composition. In a different manner the phase composition is related with the hydrogen-to-hydrocarbon (predo- minatly CH^) ratio, or strictly speaking, the {HJ/[C[ ratio in the gaseous hydrolysis products of the carbide-oxide. The phase P 2 A 7 composition then is given by the point of intersection of the corresponding lines in the Sc-O-C ternary diagram. With regard to the effects of the texture, absorption, and temperature factors, the reflection pairs for hkl 331 and 420 were used. The theoretical intensities were calculated making allowance for the different angles of diffraction for the variable a parametr of the cubic phase. The content of additional components (Sc, ScgO,, Sc,5C,g) was determined by quantitative X-ray diffraction analysis by using the internal normalization method, by comparison with the theoretical intensities of the phases present. The error of the diffrac- tographic and hydrolysis method of determination of the Sc 0 C2 phase composition (x+y+z=l) due to inaccuracy in the intensity measurements is not greater than -0.005 for x. The error of determination of HP by GLC results in errors in y and z not exceeding -0.005 within the unfavourable region of the oxygen-poor phase with the [HJ/[C] ratio about 6. Results The results are shown in Figs 1 and 2. While a continuous homogeneity region was found for scandium, two separate regions seem to occur in the Y-O-C system, a hitherto unknown carbide-oxide Fig. 1 SCnO starting mixtures, Sc1(0,CO1 the directions Homogeneity of their phase Region recomposition fee Sc^O.C.n^

Sc 0,1 0,2 0,3 SG2C 0,5 P 2 A 7

Fig. 2 starting mixtures, th direction of phase Homogeneity reoompo sition fiegions

Y C 0,1 0,2 0,3 2 0,4 YC2 phase exhibiting complex X-ray diffraction patterns appearing within the central region. The two elements differ basically in that the carbon-richest carbide of scandium is Sc,,-C,n, which is not true-of Y. The formation of Y,rC,g has not been observed within the Y-O-C system, and the Y(O,Cfl) cubic phase could be in equilibrium with YC,. The formation of M(O,CX3) phases is possible in general only by reaction of the oxide and carbon with the partici- pation of the metal; the carboreduction of the sesquioxide itself limits strictly the composition of the fee intermediate to a region very close to M?OC. Hence, it can be claimed that formation of free metal is noL involved in the oxide carboreduction mechanism.

(1) H.Eassaerts, H.Nowotny, G.Vinek, F.Benesovsfcy: Mh.Chem. 98,460 (1967) (2) F.H.Spedding, K.Gschneidner, A.H.Dsane: J.Amer.Chetn.Soc. 80, 4499 (1958) (3) M.Atodi, M.Kikuehi: J.Chen.Phys. £1, 3863 (1969) (4) B.Hajek, P.Karen, V.Broiek: J.Less-Common Met. 98, 245 (1984) (5) C.E.Holcombe, D.A.Carpenter; J.Aser.Ceram.Soe. 6£, C-82 (1981) (6) B.Hajelc, P.Karen, V.Broiek; Collect. Czech. Cheat. Conmun. 49, 936 (1984) (7) H.Auer-Velsbach, H.Nowotny: Mh.Chea. 92, 198 (1961) (8) H.Jedlicka, H.Nowotny, F.Baneaovsky: BE.Chem. 102, 389 (1971) (9) P.Karen, R.Krai, V.Broiek, B.Hadek: Proc. 39th Cohf. Czech. Chem. Soc, Olooouc August 30th - September 2nd, 1983, p. 40 P 2 A 3

C HYDROLYSIS OF CARBIDES OF MANGANESE lto^C2 and Ua23 6 P. Karen and B. Hajek Department of Inorganic Chemistry Prague Institute of Chemical Technology, 166 28 Prague

Carbides of manganese differ from those of the neighbouring d-transition metals by their hydrolyzability by water and humidi- ty.1 'The carbon atoms in the MnyC-j, Mn^Cg* Mn-jC, Mn^C^, and Ifcu-Cg carbides ' are so situated within the Mh atom prisms that C-C bonding interactions are not conceivable. Still, their hydrolysis does not give rise7 '8 }' to a single hydrocarbon, methane, accompanied by hydrogen according to the carbide stoichiometry. Since no data where the composition of the gas phase corresponded to the carbide stoichiometry were found in the literature dealing with the hydrolysis of carbides of manganese (except for Mh,C ), we studied the hydrolysis of MnyC^ ' . A mixture of hydrogen, saturated hydrocarbons in a concentration-decreasing sequence, and a similar sequence of olefins in trace concentrations resulted from the hydrolysis. Based on a comparison with uranium carbides UC- with various carbon deficit, we suggested a model for the hydrolysis reaction where active adsorbed hydrogen (H*) plays the leading part; this hydrogen partly recombines to Hg and

partly reacts with the CH^ formed by the hydrolysis of the C1 groups in the carbide structure. The products are hydrocarbon ra-

dicals and additional molecular Hg, which increases the H2 content of the gas mixture above that given by the carbide composition

(and by the presence of C-defects in UC1-X). The resulting H2 content obeys the relation x (1 + m) ([H]/[C] - 4)/([H]/[c] - 2) (1) H,2 where [H] / [c] is the amount-of-substance H-to-C ratio in the gas (this is a measure of the hydrolysis reaction stoichiometry) and m is the fraction of H* (from total H*) that has reacted with CH,. For ifayC^ and ^i-x* m is °*^ " 0.65. Clearly, so high values can occur for carbides whose composition permits the formation of CE. in appropriate amounts with respect to H*. For 50% H* to be able to react with CH., the corresponding amount of the latter must be present, hence H/C<6. Thus, m should decrease with increasing [H]/[C] ratio (for Ita^y, H/C * 2x/y). While for HrXjCg (H/C = 5) ? 2 A 8 all H* could theoretically react with methane, for Mng-jCg (H/C = 7.67) no more than 27% H can react so. Results The following two hydrolyaable phase-pure (by X-ray) carbides were prepared: JfacC2 (S.G. C 2/c, a * 1167.4-0.2) pm, b = (458.35^0.12) pm, c = (509-5-0.1) pm, /3 = (97.71-0.01)°, Z = 4) C S#G Fm and Mn23 6 ^ * 3m, a » (1059.46-0.04) pm, Z = 4). u'.on3tirred MncC, dust of mean particle size 15 pm in water under argon reacts completely at 60°C within 3h, Mhg-jCg, within 24 h. According to the reaction halflives, at 20°G the hydrolysis of both carbides is 10 times slower. The composition of the gas products of hydrolysis is given in Table I. Table I. Hydrogen and hydrocarbon concentrations in the gas mixtures after the hydrolysis of IfacCg and Mno^C^

Concentration, % vol.

Component 20 °G 60°C 2O°C 60 G

Hydrogen 53.8 51.0 70.3 67.5 Methane 28.7 34.1 25.5 30.1 Sthane 15.3 13.1 3.64 2.07 Ethene 0.128 0.080 0.043 0.017 Propane 1.63 1.40 0.683 0.261 Propene 0.016 - 0.013 0.014 i-Butane 0.052 0.037 0.011 0.004 n-Butane 0.267 0.130 0.055 0.020 Butenes 0.050 0.018 0.006 0.001 Butadienes? 0.022 0.006 0.004 0.001 i-Pentane 0.022 0.005 0.005 0.0006 n-Pentane 0.040 0.004 0.006 0.001 >C6 0.021 0.023 0.008 0.007 f H] / [C] 5.00 5.04 7.666 7.672 m 0.613 0.53 0.086 0.043

The results agree with the concept of m decreasing for carbides with heigher H/C ratios; the decreasing trend, however, is even more pronounced than as corresponds to the decrease in the probability of selection of CH+ and H* pairs from the assumed ratio of the starting components of the hydrolysis reaction; e.g.,

CH4:H* should initially be 3:11 for Mn23C6 and 3:2 for VfojC^. This departure from the theory will be caused by the high K* con- P 2 A 8

centration emerging from the hydrolysis of Mng^Cg, due to which a high fraction of CH4 radicals will be immediately terminated by trimolecular collisions with H*. The termination will be favoured by higher temperatures, as evidenced by the lower abundance c C~- and higher hydrocarbons in the mixture obtained at 60°C. J LJ The fact that the m values for MnyC^ " and for Mn^C2 (with CH,:H* = 1:1) approach each other closely indicates that the termination plays a major part if the amount of H* on the reacting carbide surface exceeds that of CH.. Other factors such as the structure and the associate cage effect (preferential recombination of radicals cf the same kind near the site of their formation) , the chemical bond nature, and the surface affinity for the adsorption of H* and other radicals will naturally contribute as well. Experimental The carbides were prepared by heating pellets of manganese and carbon mixtures under argon (1 kPa) in a Ta casing at 900°C for 600 h. About 0.1 g of sample was hydrolyzed with water in vacuum in a 10 ml ampoule equipped with a silicone rubber stopper. GLC analyses on a HP 5840A instrument; carrier gas He, TCD. columns: DC 200 on Chromosorb P and molecular sieve 5A. Calibration with a mixture of a natural gas standard and hydrogen. The hydrolysis rate was measured from the gas volume at a constant pressure.

(1) S.Hilpert, J.Paunescu: Ber. Deut. Chem. Ges. 46, 3479 (1913) (2) M.Rouault, P.Herpin, M.Fruchart: Ann. Chim. (Paris) 5., 461 (1970) (3) J.-P.Bouchaud: Ann. Chim. (Paris) 2, 353 (1967) (4) K.Kuo, L.E.Person: J. Iron Steel Inst. 178, 39 (1954) (5) A.L.Bowman, G.P.Arnold, E.K.Storms, N.G.Nereson: Acta Cryst., Sect. B 28. 3102 (1972) (6) N.I.NovikT J.N.Taran: Izv. Akad. Nauk SSSR, Neorg. Mater. 13, 1013 (1977) (7) W.R.Myers, W.P.Fishel: J. Amer. Chem. Soc. 6J, 1962 (1945) (8) L.T.Domeshevieh, B.V.Fenotschka, S.F.Gordienko, I.I.Timo- feeva, S.M.Karalnik, A.V.Koval, T.Ya.Kosolapova, in: "Karbi- dy i splavy na ikh osnove" (Ed. G.V. Samsonov).Kaukova Dumka, Kiev (1976), p. 44-49 (9) B.H&jek, F.Karen, V.Broiek: Collect. Czech. Chem. Commun. 48, 2740 (1983) (10) 77Bro£ek, B.Hdjek, P.Karen, M.Matucha, L.Zilka: J. Radional. Chem. 80, 165 (1983) (11) B.H6;jek"7 P.Karen, V.Broiek: Collect. Czech. Chem. Commun 49., 793 (1984) P 2 A 9

THERMAL EXPASSIOH STUDIES ON THE SUBCARBIDES OF TEE (SOUP V-VI TRANSITIOH METALS

B. LSnnberg Institute of Chemistry, Box 531, S-751 21 Uppsala, Sweden

Introduct ion

The thermal expansion behaviour of V_C, Mo pC, WgC and TapC has been studied. Several modifications of these carbides exist. The difference lies in the different ordering of the carbon atoms. All the carbides are disordered at high temperatures. At lower temperatures V~C and MOpC have the orthorhombic ^-Fe-U type structure while WpC has the e-Fe^Il type structure and TagC the anti-Cdlp type structure. The aim of the present work was to study the thermal expansion behaviour of these modifications and to investigate whether there is any connection between carbon ordering and thermal expansion.

Experimental

The samples were prepared by arc-melting mixtures of metal and carbon powders under an argon atmosphere. All samples, except VLC, were subsequently heat- -treated in evacuated and sealed silica capsules at 1QU8 K for 1k days. The average linear thermal expansion coefficients were obtained by measuring the lattice parameters versus temperature using a focusing high-temperature X-ray powder diffraction camera (1) with CuKot radiation. The specimens were heated in vacuum (10 Pa) and the temperature measured with a Pt/Pt-10j5Rh thermocouple. The samples were examined from room temperature to 1300 K.

Results and discussion

The variation of the average linear thermal expansion coefficient with temperature is shown in figs. \-h. It is seen that the hexagonal modifications display a similar thermal behaviour. The thermal expansion along the c-axis is larger than along the a-axis, except for Ta2C where the reversed thermal behaviour is obtained. Fig. 1 shows the thermal expansion behaviour of the V2C modifications. Their crystal structures are closely related since the orthorhombic modification can be regarded as a distortion of the hexagonal modification. It is seen that the anisotropy increases markedly as the hexagonal modification transforms into the orthorhombic modification. In the basal plane there is a decrease in thermal expansion along the orthorhombic b axis as compared to the thermal expansion along the hexagonal a axis while the expansion along the orthorhombic c axis is increased. The thermal expansion perpendicular to the basal plane is also increased. P 2 A 9

The same thermal behaviour is displayed by the Mo_C modifications (fig. 2), which crystallize in the same space groups as VgC. The anisotropy is, however, smaller than for VgC.

A similar orthorhombie distortion of the hexagonal lattice has also been found for the lov-temperature modification of HbpC^). The distortion is larger than in VpC and MOpC, vhich is also reflected, in the larger thermal anisotropy for

Nb2C(3).

There jLs an increase in thermal expansion in the direction where the metal-metal distances are decreased and a decrease in thermal expansion in the direction where the metal-metal distances are increased. This shows that the metal-metal interaction is a factor of minor importance to the thermal anisotropy. It seems more probable that the carbon atoms and their ordering scheme play a more important role in determining the thermal anisotropy of these carbides.

The thermal expansion of W,C (fig. 3) was measured on as arc-melted samples. 'ih° disordered modification is expected and thus the same thermal behaviour as for hexagonal V2C and Mo-C.

As mentioned earlier, the thermal expansion along the c-axis is lower than along the a-axis for Ta2C (fig. U). The same thermal behaviour was obtained for as arc-melted samples as for heat-treated samples, even though they display a different type of ordering. Possibly the high-temperature modification was not (*) retained in the arc-melted samples

As can be seen from the figrres, the thermal behaviour changes considerably as the carbide transforms from one modification to another. It seem thus probable that the rearrangement of the carbon atoms will effect the thermal expansion be- baviour of the carbides.

References

(1) G. Hagg, N.-O. Ersson, G. Rudenholm and B. Sellberg, J. Appl. Crystallogr. .12.; 221 (1979). (2) K. Yvon, H. Novotny and R. Eieffer, Mb. Chem. £8, 31* (1967). (3) B. Lonnberg and T. Lundstrom, submitted to J. Less-Coanon Met. (198U).

* A neutron diffraction study on WpC and Ta^C is in progress. P 2 A 9

oti 1

1000 T(K) 500 1000 TOO

Fig. 1. Variation of the average Fig. 2. Variation of the average linear thermal expansion coeffi- linear thermal expansion coefficients with temperature for hexa- cients with temperature for hexagonal and orthorhombic VJ2. gonal and orthorhombic MOgC. hexagonal VgC hexagonal M02C orthorhombic — orthorhombic Mo [_ perpendicular to the basal I perpendicular to the basal plane plane | parallel to the basal plane I I parallel to ^ae basal plane

6 a.io or a«io6 (K"') 10

500 • ' • KXX) T(K) 500 1000 TOO

Fig. 3. Variation of the average Fig. U-. Variation of the average linear thermal expansion coeffi- linear thermal expansion coefficients with temperature for WpC. cients with temperature for

Bertil Lonoberg Institute of Chemistry SIEMENS

Siemens your partner for diffractometry Automatic powder Automatic single cristal diffractometer D 500 diffractometer AED 2 • Position sensitiv proportional counter • Eulerian cradle diffractometer for high speed diffractometry • Supporting computer • Computer system • Graphic system • Color graphic system with software • Control and structure determination DIFFRAC11 software • Compact medium frequency • Compact medium frequency generator KRISTALLOFLEX® 710 H generator KRISTALLOFLEX 710 H

For complete details on Siemens diffractometer, from both, an analytical and investment point of view, write to: Siemens AG, Infoservtce 213-162 P. O. Box 156, D-8510 Fuerth

Siemens AG Osterreich • P. 0. Box 83 i A-1211 Vienna P 2 A lo

PREPARATION AND CHARACTERIZATION OF NEW OXYNITRIDES WITH PEROWSKITE STRUCTURE R. Marchand, F. Pors et Y. Laurent Laboratoire de Chimie Minerale C, L.A. 254, Avenue du General Leclerc, F. 3S042 Rennes Cedex.

Introduction of nitrogen into the anionic network of an oxyde leads to an increase of the negative charge. As a consequence, it is necessary to change an element with an oxydation degree by another one with a higher oxydation degree into the cationic network. So, if we consider the baryum titanate BaTiO3 which has got a perovskite structure type, the substitution of some oxygen by nitrogen has to be made with substitution of baryum by an element which can be trivalent or more. We can also change titanium by an element five or more. These substitutions can be made together or one by one. New oxynitrides with perovskite structure type have been prepared and characterized with ABO2N formula (A = Ca, Sr, Ba - B = Ta, Nb). There are two ways to prepare these compounds : - either introducing nitrogen as nitride or oxynitride by solid state reaction. - or nitriding an oxygenated mixture in a reducing atmosphere. We used ammonia which presents the advantage to be at the same time nitriding and reducing.

As an exemple BaTaO2N is prepared by reaction of ammonia flow with a stoechiometric mixture of barynm carbonate and tantalum oxyde at the temperature of 1000°C. ' This product is characterized by chemical analysis and thermogravimetric, magnetic and radiocristallographic studies. It is not possible to determine if it exists an order between oxygen and nitrogen by X-ray diffraction. A neutron diffraction study has been made, because by this way it is possible to see the difference between oxygen and nitrogen (bO = 0.58 10~12 cm - bN = 0.94 10~12 cm). The time of flight neutron diffraction method associated with profile analysis refinement has been used as experimental technique. The figure 1 shows the experimental and calculated diffraction spectra. P 2 A lo BR Tfi 02 N 5000, 28 = 90 DEC

tooo. sperme OBSERVE SPKTIC CBLCULE oE

3000.

2000.

1000.

100. l50. • aOoV" 250. 300. 350. 400. ' t|50. "500. 550i GOO. ' 650. "7007. NUHERO OE CRNflL.

Neutron time-of-flight diffraction spectra. Tick marks below the profile indicate the calculated line positions.

The unit cell is cubic : Space Group Pm3m - R profile = 0.0426 for 74 independent reflexions. Oxygen and nitrogen are statistically distributed on the 3c position. We have not observed an order-desorder transition until the temperature of liquid helium.

This research completes the studies we had already done about the L712AIO3N compounds which possess a K2NiF4 structure type which can be described as a bidimensionnal perovskite (1, 2).

(1) R. MARCHAND, C.R. Acad. Sci., 282C, 329, 1976. (2) R. MARCHAND, R. PASTUSZAK, Y. LAURENT et G. ROULT, Rev. Chim. Min., 19, 684, 1982. P 2 A 11

X-BA7 DIFFRACTION AND DIFFUSE SCATTERING OF MyC-j SINGLE CRYSTAL CARBIDES J.P. Morairoli, M. Khachfi, M. Gantois Laboratoire de Genie Metallurgique (IT.A. 159 OTRS) Ecole des Mines - Fare de Saurupt, 54042 NANCT-Cedex (France) A. Courtols Laboratoire de Mineralogie et Cristallographie (ERA 162 CNRS) Universite de Nancy I, B.P. 239, 54506 VANDOETJVRE-les-NANCY Cedex (France)

INTRODUCTION Previous electron microscopy and diffraction studies on various types of M-C, carbides [1,2,3] indicated that the structure of these carbides can be described by the orthorhombic structure proposed by Fruchart [4]. This structure, with a b/a ratio which equals ST, can be considered as ordered with respect to an "average" hexagonal disordered structure. It exhibits three orientation variants separated by antiphase or twin boundaries.

The possibility of synthetizing Cr_C_ single crystals and extracting (Fe.Cr, V)_C. single crystals from high chromium content cast irons has allowed us to perform an X-ray study of these two types of carbides. Experimental results thus obtained are compare'** rith those found earlier and information about the stacking of planar defects is presented.

SAMPLES AND EXPERIMENTAL RESULTS Pure Cr-C- single crystals are elaborated from powder by induction melting. (Fe,Cr,7)-C. carbides are extracted by anodic dissolution from white cast irons of composition (wt Z) 2.8 7.C, 15.2 ZCr and 3.1 TV. In both samples needle-shaped single crystals with an average length of about 1 am are suitable for rotating and equi-inclination Weissenberg experiments. Rotation is performed about the c axis and 0, 1st, 2nd and 3rd level layer photographs are recorded using Ka., and KOL, radiation. The 0 level layer photographs are absolutely identical for Cr-C- and (Fe,Cr, V) C_ carbides. These patterns show perfect hexagonal symmetry and can be interpreted using the Fruchart's orthorhombic cell with a b/a ratio which strictly equals /T. Since only fundamental reflections typical of the "average" structure are present, these patterns do not allow making a distinction between the three orientation variants (Figure la).

Careful examinations of the reflections of the 1st level layer photographs for both carbides (Figure 1 b,c,d) reveal the real orthorhoabic sjranetry. In addition to fundamental reflections, the patterns contain superlattice reflections which prove without ambiguity the simultaneous presence of the three P 2 A 11

Ca) Cr_C, and (Fe,Cr,V)_C,

(b) (c) (Fe,Cr,V)7C3 (d) (Fe,Cr,V)7C3 W Fundamental reflections, • • • Superlattice reflections belonging to tbe 3 orientation variants. O A Extra spots at 1/2 and 1/3 due to the ordered stacking sequences.

Figure 1 : Sections of the zero (a) and first (b,c,d) level layer Weissenberg photographs of Cr-rC^ and (Fe.Cr,V)jC^.

11100) "Q-x-q-x« - x Q x D.

Figure 2 : Symbolic descriptions of the stacking sequences of the non-faulted orthorhombic variants (a and b) and of the four stacking sequences observed in heavily faulted (Fe,Cr,V)_C- carbides (c, d, e and f)r P 2 A 11

'orthorhombic orientation variants. The relative proportion of each varies from one crystal to another and can be easily determined by intensity measurements of the superlattice reflections. Moreover 1st layer photographs of (Fe,Cr,V)_C, carbides also exhibit diffuse scattered streaks crossing the superlattice spots and forming stars of David distorted by the Wei3senberg geometry (Figure 1 c,d). These stars of David, which have been previously observed by electron diffraction [2], indicate the presence of numerous randomly distributed {1100} planar defects. They are twins and antiphase boundaries with fault vector R • a,b or a+b (where fault planes and vectors are given with respect to the "average" structure) separating orientation variant domains. These patterns also exhibit extra spots located on the streaks between neighbouring superlattice reflections at half or/and one third of their distance. Several arrangements are observed depending on the crystal under examination. This indicates that many of the {1100} planar defects are not randomly distributed in the crystal but preferentially adopt typical arrangements. Consequently, the stacking sequences of the non-faulted crystals which can be described either by ++++ or (Hagg's notation [5]) (Figure 2 a and b) are changed into the following sequences : +-+-+- ; -++-++ ; +—+— and —H-—H- (figure 2 c,d,e,f). They are the simplest sequences and they correspond to the most heavily faulted arrangements. Therefore, X-ray experiments show that on the scale of the whole crystal, these sequences actually occur with a relatively high frequency in domains distributed throughout the crystal. These sequences and other more complex ones have been recently reported by Dudzinski et al [6] but they were observed on an electron diffraction scale (a small area of a single crystal).

CONCLPSIONS This X-ray study of Cr_C, and (Fe,Cr,V)_C, single crystal carbides confirms previous X-ray and electron microscopy and diffraction studies. It indicates that when very numerous defects are present, interactions between them probably occur and produce typical stacking sequences. In this study, four simple stacking sequences are observed at the scale of a whole crystal.

[1] E. Bauer-Grosse, J.P. Morniroli, 6. Le Caer and C. Frantz, Acta Met., 2£, 1983, (1981) [2] J.P. Morniroli and M. Gantois, J. Appl. Cryst., \6, 1, (1983) [3] J.F. Morniroli, E. Bauer-Grosse and M. Gantois, Philos. Hag., 48, 311, (1983) [4] R. Fruchart et A. Kouault, Ann. Chim., £, 143, (1969) [5] G. Higg, Archiv. For Kemi. Miner. Gcol.» 16B, 1, (1943) [6] W. Dudzinski and M. Kovalski, 3rd Int. Conf. "Carbides, borides and nitrides", Poznan (Poland), oct. 1984

J.P. Morniroli, Laboratoire de Genie MStallurgiquc (U.A. 159 CHBS) Ecole des Mines - Pare de Saurupt, 54042 NANCT-C«dex (France) P 2 A 12

DIFFRACTION STUDIES OP ELECTRON DENSITY, LATTICE VIERATIONS AND INTERDIFFOSION IN SOME TRANSITION METAL CARBIDES V. Valvoda Charles University, Ke Karlovu 5, 12116 Prague 2, Czechoslovakia

Studies of the above mentioned physical effects were carried out using X-ray and neutron diffraction on powder samples. Electron densities in TaC QQ(1) and TiC og(2) were determined from X-ray intensities corrected for preferred orientation and surface roughness effects which were experimentally investigated separately. The multipole expansion method of Kurki-Suonio (3) or the difference Fourier synthesis were used for the description of the atomic charge or crystal charge distribution deformation due to bonding. The charge transfer from the Ti-atoma to the C- -atoms is seen in (110) plane of TiC (see Fig.l) as well as an accumulation of the charge in the space between atoms.

c Ti C \ i

"\\

( 6)) (Co ) J

^—• f -> f Ti

Fig.l. Deformation electron density map in the (110) plane of TiC. Contour interval 0.4 eX~^. Negative contours are broken lines, zero being the first broken line. (This figure was kindly computed and drawn by Dr Peter Blaha from TO* Wien using our experimental data.} P 2 A 12

-fries vibrations represented by thermal root-mean-aquare displacements (r.m.s.d.) /u } ' of atoms or Debye characteristic temperature ©„ were determined from the temperature dependence of integrated intensities of X-ray reflections. The small values of

Tab.l. compound (2)

TiC.96 0.045 810 0.064 TaC.98 0.046 475 0.061 thermal displacements of atoms in the two carbides presented in Tab.l reflect the strong interatomic bonds in the crystal lattices. The values of total r.m.s.d. (also given in Tab.l) were determined from the angular dependence- of intensities of X-ray reflections at rooE temperature and they indicate the existence of random static displacements of atoms in the carbides under study. This fact is supported by neutron diffraction on the TiCa< - sample (see = when tile Fig.2) from where <(u )£otal °'°60 * strong reflections are used only (they are represented by circles in Fig.2).

NEUTRON DIFFRACTION

0 1 2

Fig.2. Wilson plot of neutron diffraction intensities for TiC ac where Icj[£c are the intensities calculated for the lattice in rest, 9 is the Bragg angle a !X wavelength

From this experiment we may also conclude that the observed decrease of intensities of the first two reflections found by X-ray diffraction (as compared with the calculated values for free spherical atome) cannot be explained by systematic static displacements of the Ti-atoms outwards the carbon vacancies and should be connected with the charge redistribution due to bonding. P 2 A 12

Interdiffusion plays an important role in preparing mixed carbide a and in sintering proceaaea generally. Homogenization of TaC and HfC powder blends was investigated after annealing at 19OC°C by means of X-ray diffraction. The diffraction band between the 220 reflections of both components, which is formed during the interdiffuaion process (see Fig.3), indicates that TaC is much faster contamined by hafnium atoms than HfC by tantalum atoms. The degree of interdiffusion and diffusion coefficients were determined using the modified Rudman's method (4). The influence of grinding and differences between the bulk and surface diffusion were observed.

29 en- 28 29 BC)-—

20 h 200 h

28 29 28 29 en—

Fig.3. The diffraction band 220 of HfC and TaC homogenized at 1900°C for various times. (1) V. Valvoda, phys.stat.sol.(a), £i, 133 (1981) (2) V. Valvoda, P. 5apkovd, phys.stat.sol.(a), Sit 203 (1984) (3) K. Kurki-Suonio, Aeta CrySt., 121, 379 (1966) (4) E. Delhez et al., J.iEater.Sci., 12, 1671 (1978)

Or. 7. Valvoda, Charles University, Ke Karlovu 5, 12116 Prague 2, Czechoslovakia P 2 A L3

CRYSTAL STRUCTURES AND MAGNETIC PROPERTIES OF LOW-DIMENSIONALLY BRIDGED TRAN- SITION METAL HEXACYANO COMPLEXES D. Babel, M. Witzel and J. Pebler Fachbereich Chenrie und Sonderforschungsbereich 127 der Universitat, Hans-Meer- wein-StraBe, D 355o Marburg

The transition metal ions of octahedrally coordinated M(CNU units in most solid hexacyano compounds are either isolated or threedimensionally bridged by their ligands. Intermediate between these types of isolated and framework structures, which are represented on the one hand by e.g. slpasolites CSgLiMfCNU

(1) and on the other by Prussian blue-related compounds like KFe2(CN)g (2),

Fe4[Fe(CN)g]3 • xHgO (3) or Mn3[Co(CN)g]2 -xHgO (4), we have found a linear cyano-bridged chain structure in the octahydrated compounds NMe^MnM(CN)g -8^0, M( 111) = Mn, Fe (5). In addition to the isostructural Co(III) chain compound we report now two further intermediate structures: A layer structure adopted by the tetrahydrate

NMe4MnCr(CN)g • 4H«0 and a spacious framework structure observed at two new

"trihydrates", NMe4Mn2(CN)g • 3,2H20 and NMe4CdFe(CN)g • 3,2H20. The latter contain the same kind of layers as the tetrahydrate, but interconnected in the third dimension in a "dilute" way, schematically shown in Fig 1. Fig 2 illustrates the MnNg coordination at the connecting sites and the trans-dihydrated MnN^HjOJg octahedra within the layers, as also found in the tetrahydrate structure.

MF.n Mn^rC^W^ Mo*. 32

Fig. 1: Two- and threedimensional Mn-NC-M{III) bridging in NMe4MnCr(CN)g-4H20 and NMe4Mn2(CN)g-3,2^0 P 2 A 13

Fig. 2: 2112 103.7 MnN5 and 4(2)2 coordination of the

2t«.S Mn(II) cations by surrounding Mn(CN)g3- anions and water of n m NMe^Mn Mn (CN)5 • 3,2 HjO hydration.

The cell parameters of the compounds mentioned are listed in Table I along with some averaged interatomic distances. Details of the structures and the results of magnetic (6) and Mossbauer studies performed at related compounds will be discussed.

Table I: Crystal data of hydrated hexacyano complexes, exhibiting cyano- bridged chain, layer and spacious framework structures.

S.G. a b c (pm) (pm) (pm) (pm) (pm) (pm)

NMe4MnCo(CN)6 8H20 P4/n 2 Io62.2 Io46.2 189.3 22o.l 218.4 NMe4MnCr(CN)6 4H20 Pnma 4 1491.0 785.5 1723.5 2o6.7 221.1 22o.2 NMe4Mn2(CN)g -3,2^0 14 lo 1653.2 1728.8 2ol.7 220.4 222.1 NMe4CdFe(CN)g-3,2H20 14 lo 1650.3 1733.8 193.6 228.0 233.7

(1) B.I. Swanson and R.R. Ryan, Inorg. Chem. 12, 283 (1973 and ^3, 1681(1974) (2) J.F. Keggin and F.D. Miles, Nature 137, 577 (1936) (3) F. Herren, P. Fischer, A. Ludi and W. HSlg, Inorg. Chem. ]£, 956 (198o) (4) A. Ludi, H.U. Glide! and M. Riiegg, Inorg. Chem. £, 2224 (197o) (5) H. Henkel and D. Babel, Z. Naturforsch. 39b, 88o (1984) (6) W. Kurtz and D. Babel, Solid State Comnun. 48, 277 (1983)

Prof. Dr. D. Babel, Fachbereich Chemie der Philipps-Um'versitat, Hans-Meerwein-Stra8e, D 355o Marburg P 2 A 14

X-EAT. SUXSSICS7 SPECTBOSCOPIC STUDY 07 SOME TEMSHION k.T.ffiuiuw't'H ^J 0ALCIU1C GERUAITAXB GAHETBTS G. JaaioieJc. H. D^blcowska, A. D^bkowaJd and B. Sobczak Institute of Physics of Polish Academy of Sciences, 02-668 Warszawa

The single crystals with the CaiXpGeÔ general formula, where'X = Or, Mn, Fe and Ga, have been the subject of our investigations. There have been grown from different fluxes. They have been grown from the high temperature solutions by the slow cooling method. The fluxes and temperature conditions used have been similar to those established by Mill /1/ for these materials. Single crystals grew up to 7mm x 7ram x 6mm in the case of CaiMruGeÔ-p* 6mm x 6mm x 6mm in the case of Ca-.Cr;> do Gez?0 /,,iz) and'5mni x 5mm x 5mm in the case of Ca-jFeoGej-O-,.o d y td. Cooling rate was about 1.5deg/h for all the cases. The chemi- cals used were all "p.a" grade. The morphology of the crystals was revealed by the SEM techniques. The observations have been performed on the(iio) oriented natural surfaces, mainly. The concentrations of the elements forming the investigated crystals were measured using the Electron Microprobe Analyser /EPMA/. The observations of the valence band structure changes due to the type of the ion in the octahedral position in the garnet lattice were the aim of our investigations. The X-ray emission spectroscopic study of the valence band structure were carried out using one-crystal spectrometers of the Johansson type with LiP /interplane spacing d=2.015.8/ and EAP /d=13.06S/ crystals in the EPMA /JXA-50A/. Particular attention was paid to the features of CaKpc* âKf-i z» ^pci

XKp-, ,, Gal^ , GaL a , GeLo< and Gel.^( lines. Kpc line corresponds to the MJY —* K transition, K^ , line corresponds to the MTT T-PT—"• K transition, L« and La, correspond to and tr the Mjy. 7 • I»TTT U-vj—""^II a^isitions, respectively. The chemical bonds of Ca, X and Ge with the oxygen are realised by the 4-s—, 3d- and 4p-orbitals. The valence band of the CaaXgGe-O^p is touched by the 3p-electrons of P 2 A 14

calcium, and 3d-electrons of the remaining cations,which compose these garnets. She intensity distributions, structure and some parameters of the analysed lines in these garnets were found to be distinctly different froa those of the standards, Ca?o ssxd pure metals. For example, the position of I^r of Oa on the energetic scale was shifted in respect to this line position for CaEp crystal in the short wavelenght direction, up to about (3.0 - 0.2) e7, when Ixltn. For remaining Xr* ions the observed shift does not exceed the value of (2.0 £ O.i)eV. The position of CaKa* -, changes slightly. But characteristic or s is, that for Ca,Mn2Ge,0,|2 y *sls a shift appears in short wavelenght direction in respect to the standard whereas for Cr- , F-e- and Ga-ions the oposit shift was detected. The half

-width /A E1/2/ of CaKp1 ^ line increases from (2.7 £ 0.1)eY

for GaF2 to (3.0 £ 0.1) eV for the garnets, except for the gar- was net with manganese. In this last case no changes of -A E^ ,2 observed. These values and these mentioned below are cot decon- voluted with the half-widths of the spectrometer windows. For the garnets with Cr, Mn and Fe it was detected, that the distance between Kgc and So,, , lines decreases about (2.2 £ O.^eV in comparison with these line separation for the standards. The A Bi/2 of EPi 3 line ciiar*Ses for manganese only. It decreases from (12,5 - 0.5) eV for metal to (11.5 - 0.5) e7 for garnet. Furthermore a satellite appears in the distance of 14eY from K^ 1 line in the long wavelenght direction. The "satellite can be characterised by the(22 - 2)percentages of the sain maximum. The half-widths of 1^ and L^ lines for gallium changes from (2.1 * O.i) eV and (2.6 £ O.i) eV for standard to (2.5 £ O.-j^Yand (2.7 £ O.i) eV for the garnet, respectively. Similar changes were observed for the ^GacO^ comPound« The positions of the L-lines of germanium change slightly betweene the standard and garnets. In each case a shift occurs in the short wavelenght direction. The largest one /(O.8£o»i)eV/ appears for the garnet with manganese ions. For this garnet one can observed strong difference in the intensity relation L. in comparison with the other garnets and with the P 2 A standard. The SPMA analyses confirmed that in the case of CaiMnoGexO-p *ke concentration of germanium is less than this arising from the formula and that the manganese concentration is larger than that. But decrease of germanium is larger than the increase of manganese concentration. In the case of the other garnets no deviations from stoichiometric composition were registrated. The obtained results suggest that in the ca- Ge 0 arne s se of Ca5ifci2 3 i2 S * manganese ions occupy not only octahedral sites /2/ but also the tetrahedral ones. That means that the correct formula for the investigated garnets with manganese ions should be Ca^M^Mn Ge, O.^* The precise estimation of the x parameter is difficult. But one may be say that it does not exceed the value of 0.4. On the other hand the parameter x is smaller than 0.2, most probably. The appe- arenca of the "satellite* of Ka* -> is caused, most probably,by 4+ * "*" the presence of Mn ions in the garnet lattice. A simple model which takes into acount the lattice constant changes versus the ionic radius of the ions in the octahedral sites /2,5/ leads to the conclusion, that effective ionic radius of manganese is 0.622. The literature gives the values 0.645§ and 0.540$ for !tfn*+ and Mn +, respectively. Calculated effective 4+ radius corresponds to about 20^ of Ma . That neans this value corresponds to the x parameter equal about 0.4,in our notation. The authors of the present work beljcre that their investigations permitted to revealed the new interesting aspect of the real structure of the CazHn^GtezO^ garnets. /1/ B.T. Mill , Kristailografia 19 ,5, 1057, /197V /2/ S.Seller, Zeitschrift fttr KrTstallogaphie, Bb, 125,S,1967 Z.A.Kazej, B,7,Mill and W»I.3olcolov, Journal of Experimen- tal and Theoretic Physics, 24, 4, 229 /1976/

Dr Gabriel Jasiolek Institute of Physics.Polish Academy of Sciences al.Lotniiow 32/46, 02-668 Warszaua, Poland P 2 A 15

MECHANISM OF PHOTOCHROMISM IN CADMIUM SILICATE GLASSES

Peter G. Perkins Department of Pure 8nd Applied Chemistry University of Strathclyde Glasgow G1 1XL Scotland

Photochromism is an important property occurring in some materials

and indeed some commercial 'glasses' are photochromic. In the latter

cases one finds crystals embedded in the glassy matrix. Our interest

is however centred on those glasses which are intrinsically photochromic and one such is cadmium silicate. Here the glass undergoes reversible darkening on exposure to light. It is important to understand the origin of the phenomenon and to this end we have mounted a theoretical attack by means of band structure and cluster calculations.

We have previously found that a linear silicate chain gives a good account of the electronic structure and density of states of a glass and in the present work we have constructed a CdSiO, chain. The basis sets used were the valence orbitals plus 3d for Si. For Cd, test calculations showed that the 4d (filled) orbitals were of no essential importance. Accordingly we have a valence basis set for this atom of

5s, 5p.

The band structure method was a self-consistent technique including orbital overlap. The band structures are plotted automatically and various density of states curves (total d.o.s., joint d.o.s., partial d.o.s.) can be generated.

The valence bands calculated are curved at high energy but, near the band gaps, are rather flat. They consist mainly of oxygen non- P 2 A 15

bridging states. Of more interest is the first conduction band.

This is generally broad but has a flat section near the gamma point.

The band has strong Cd (5p) character near T but assume more silicon

character near the middle of the T-X line. The band gap is near 7eV

and this is lowered to around 3eV by elecfron-hole interaction. The

latter value is in good accord with the experimental absorption energy.

Transition across the band gap causes a movement of electrons

from O(nbr) to Cd(5b). The charge on Cd in the ground state polymer

is - +0.5 and hence the excited state causes the Cd atom to become more nearly neutral. Experimental evidence from e.s.r. shows that the unpaired electron in the darkened state is associated with an oxygen

(nbr) atom and not with the Cd atom. This concurs with the present model.

An electron in the conduction band can be trapped by the defects inherent in the glass structure and, whilst in this state, further transitions in the conduction band can be made. There can be continuous absorption due to the wide conduction band and hence a darkened glass is predicted.

Relaxation of the electron to the valence band can be brought about by thermal excitation which transfers electrons along the conduction band to the silicon states.

When copper atoms are doped into the glass then the 3d levels fall into the region of the higher valence states of the model.

Hence the presence of Cu enhances the valence band states and so increases the intensity of the first transition. Hence copper should enhance the photochromism. When iron is doped into the structure the

3d levels are situated in the band gap. They thus provide a site for P 2 A 15

electron-hole recombination and moreover they 'absorb' the ion energy transitions. Hence iron should quench the photochromisni. Both these effects are observed.

1. A. Breeze, P. M. Magee and P. G. Perkins, Phil. Mag. B. 1979, 40, 105. P 2 A 16

TETRAMERIC SILVER(I)TRI-tert-BUTQXYSIL/«CIHIQLATE W. Wojnowski , M. Wojnowska, K. Peters, E.rM. Peters and H.G.van Schnering Max-Planck-Institut fiir Festkarperforschung, D-7000 Stuttgart 80

The formation of salts of trialkoxysilanethiols with alkali-, alkaliearth- and heavy metals: Cd, Hg, Pb, Ag, is already known (1), however, not their crystal structures. These compounds are of interest for the chemistry of metal sulfides as well as for bioinorganic chemistry. The bulky tri-tert-butoxysilanethiolate- ligands restrict the metal-to-sulfur coordination. Another feature is the proto- lytic stability of the Si-S bond in this substituent. The tetrameric silver(I)tri-tert-butoxysilanethiolate Q] has been obtained according to the following equation:

4 Ct-C4HgO)3SiSH + 4 AgNO3 = [(t-C^HgO)-SiSAg]4 + 4 HN03 Compound 1 crystallizes as colourless triclinic plates (space group PT, a=1769,7, b=2012.8, c=1266,8 pm, o=119.63°, 8=82.22°, y=95.08°; 2=2, R=0,067 from 76S6 observed reflections hkl).

Fig. 1 a)permanent address: Institute of Inorganic Chemistry, Technical University PL-80-952 Gdansk P 2 A 16

A view of the molecule is shown in Fig. 1. The alternating Ag^S4 framework with two-bonded silver and three bonded sulfur atoms is nearly plane. The mean values of relevant bond distances and bond angles are listed in Tab. 1. Most important is the significant deviation of the S-Ag-S bonds from a colinear configuration C172.3°) resulting from a displacement of the Ag atoms towards the center of the Ag.S. ring (Ag-Ag=313.5 pm). This seems to be a common effect in comparable molecules of Cu(I), Ag(I) and Au(I) (2) and can be interpretet as a strong indication for high order direct Ag(I)-Agd) interactions.

Tab. 1 Bond distances [pm] Ag-S 238.3(3) 0-C 136.5(15) S-Si 211.9(3) Si-0 155.3(9) Ag-Ag 313.5(1) Bond angles [°]

Ag-Ag-Ag 90.3(1) S-Ag-S 172.3(1) Ag-S-Ag 81.8(1) and Ag-S-Ag 82.8(1) Ag-S-Si 98.5(1) and Ag-S-Si 94,1(1) Si-O-C 147.8(9) and Si-O-C 152.9(10) S-Si-0 106.9(4) and S-Si-0 111.6(5) O-Si-0 108.9(5)

29 The Si-WR spectrum of 1 in CDC13/TMS solution is shown in Fig. 2. fflj . ., • •eaa -6&5 -68.6 -68.7 -6ftB Fig. 2 6[ppm] P 2 A 16

The splittings around S(Si)—68.54 ppm are remarkable and can be explained by couplings of silicon with silver nuclei, not removed by broad band proton de- coupling. 2 evaporates in vacuum at 170°C without decomposition. In F.D.-mass-spectrum the molecular ion M m/e=1548, with the expected isotopic pattern, was found.

In E.I. mass-spectrum the [M-(C4Hg0),SiS]-cation (Fig. 1, a-split) is dominant

(4H relativ intensity) over the [M-(C^HgO)3Si] cation (Fig. 1, b-split), (51), + but M is also present (5!). No fragments [(t-C4Hg0)3SiSAg]n, n=1,2,3, were found.

(1) W. Wojnowski, Wiss.Hefte Techn.Uhiv.Gdansk, Chem. 22, VTZ, 1 (1971) (2) S. Gambarotta, C. Floriani, A. Chiesi-Villa, C. Guastini, J.Chem.Soc, Chem.Commun. 1983, 1087, 1156 and 1304.

Prof .Dr. W. Wojnowski, Max-Planck-Institut fur Festkoperforschung, Heisenbergstr. 1, D-7000 Stuttgart 80 P 2 A 17

CALCULATED ANGLE RESOLVED PHOTOEMXSSION SPECTRA FOR TiC(lOO) J. Kedinger and P. Weinberger Institut fur Technische Elektrocbeioie, Tec.hnische UniversitSt Wien Getreidemarkt 9, A-1060 Wien, Auitria E. Winsaer and A. Neckel Institut far Physikalische Chemie, TJniversitit Wien, WShringerstraBe 42, A-1090 Wien, Austria

Classification: CMD 8, Photoetnission In relation and based on selfconsistent FLAPW film potentials for TiC(lOO) (!) the angle resolved photocurrent for unpolarized Ne-I and He-I radiation is calculated within the one-step model (2,3). For off-normal emission good agreement with experiment (4) is only achieved by using a realistic surface potential.

(1) E. Wimmer, A. Neclcel and A.J. Freeman, Phys.Rev.B, in print (2) J.B. Pendry, Surface Science 57, 679 (1976) (3) L.Z. Johansson, C.G. Larsson and A. Callenas, J.Phys.F 14, 1761 (1984) (4) A. Callenas, L.I. Johansson, A.N. Christensen, K. Schwarz and J. Redinger, Phys.Rev. B27, 5934 (1983) P 2 A 18

CALCULATED ANGLE RESOLVED PHOTOEMISSIOH SPECTRA OF NOtlSTOICHIOMETRIC TiN^C100) J. Redinger, 6. Schadler and P. Weinberger Institut £Sr Technische Elektrocheoae, Technische UniversitSe Wien, Getreidemarict 9, A-1060 Wien, Austria J. Klima and A. Neckcl Inscitut fur Physikalische Chaoxs, Universitit Wien, WahringersCrafle 42, A-1090 Wicn» Austria

Classification: CMD 8, Photoemiasion the theory of angle resolved phocoemission for statistically disordered alloys (1) is extended to complex lattices. Based on scattering amplitudes from a recent KKR-CPA calculation for TiN- ., (2) normal emission spectra for the (100) surface are calculated for various photon energies. The theoretical spectra are compared to experimental data for TiNQ „, (3).

(1) P.J. Durham, J.Phys.F II, 2475 (1981) (2) J. Klima, G. Schadler, P. Weinberger and A. Neckel, J.Phys.F subm. (3) L.Z. Johansson, A. Callenas, P.M. Stefan. A.M. Christensen and K. Schwarz, Phys.Rev. B24, 1883 (1981) P 2 B 1

ENERGY-BAND-STRUCTURE STUDY OF THE NbN(IOO) SURFACE

A. Callends and L.I. Johansson Department of Physics and Measurement Technology, Linkb'ping University, S-581 83 Linkoping, Sweden A.N. Christensen Department of Chemistry, Aarhus University, DK-8000 Aarhus C, Denmark K. Schwarz Institute of Technical Electrochemistry, Technical University of Vienna, A-1060 Vienna, Austria

The results of angle-resolved photoemission measurements and of a linear augmented-plane-wave (LAPW) band-structure calculation carried out on NbN are reported. The photoemission experiments were made on a NbN(100) single-crystal and the results were interpreted by using the calculated band-structure and assuming direct-transition. This allows a detailed study of the electronic structure, identification of individual bands and determination of band dispersions, and similar studies have earlier been made on some other transition- metal nitrides and carbides (1-4), but not on NbN. Earlier photoemission experiments on NbN (5) have been made angle-integrated and provided therefore only information about the total density of occupied states.

A photoemission system equipped with a differentially pumped He resonance lamp was used for these experiments which were performed under UHV conditions, base pressure less than 1-10 torr. The NbN(100) surface wes cleaned in situ by sputtering with argon ions followed by annealings. The cleanliness of the sample was checked with Auger electron spectroscopy.

In Fig. 1 angle-resolved electron energy distribution curves (EDO recorded from the NbN(100) surface using unpolarized He I (21.2 eV) radiation are shown. The spectra are recorded at normal electron emission using two different incidence angles, 8^-, of the radiation. A polarization effect is seen in Fig. 1 which together with symmetry selection rules (6) makes it possible to identify the symmetry of the initial-state bands. Peak A, at about -3.4 eV in Fig. 1, can be associated with emission from a Ac initial-state band and peak C, at about -6.9 eV, as originating from a t^ Initial-state band. Looking at the band-structure calculation shown in Fig. 2 and by using the direct-transition P 2 B 1

1 1 1 . 1 i NbN(100) 40 NORMAL OMSK* hv-I L2«tf

••

•

•>*

V V. a,-

V, C B " i;

1 1 1 i 1 -8 -6 -4 -2 0 WITIAL-STATE ENERGY (eV) Fig. 1 Angle-resolved He I spectra from NbN(100). model, peak A can be interpreted as arising from transitions between the initial- -15 state band of Ag symmetry at about 4.4 eV below the Fermi level and the final-state TAX band of A. symmetry at about 16.8 eV Fig. 2 Energy bands of NbN calcu- above the Fermi level, (E- - 4.4 eV) lated along the r*X synme- •*• 4 (E + 16.8 eV) (see dashed arrow try line using the LAPW 1 f method. in Fig. 2). Peak C in Fig. 1 corresponds to transitions between the initial-state band of& 1 symmetry at -7.3 eV and the final-state band of ^ symmetry at 13.9 eV above the Fermi level, &1 (Ep - 7.3 eV) * a, (Ep + 13.9 eV) (see dashed arrow in Fig. 2). The polarization dependence observed for the peak at about -5.0 eV (labeled B) associates this peak to transitions from an initial-state of Aj symmetry. However, when using the direct-transition model and the band- structure calculation shown in Fig. 2, there is no corresponding &y initial- P 2 B 1

state available. We believe that one-dimensional density-of-state effects may give rise to this structure and that it may reflect the location of the rela-

tively flat portion of the Aj band extending around the X4' point.

Our results indicate that most of the structures observed in the photoemission spectra from the NbN(100) surface can be explained in terras of direct transitions. The experimental results and the calculated band-structure of NbN will be discussed in more detail in a forthcoming publication (7) where results on VN will also be presented.

REFERENCES

(1) J.H. Weaver, A.M. Bradshaw, J.F. van der Veen, F.J. Himpsel, D.E. Eastman, and C Politis, Phys. Rev. B 22, 4921 (1980). (2) A. Callena°s, L.I. Johansson, A.N. Christensen, K. Schwarz, and J. Redinger, Phys. Rev. B 27, 5934 (1983). (3) L.I. Johansson, A. CallenSs, P.M. Stefan, A.N. Christensen, and K. Schwarz, Phys. Rev. B 24, 1883 (1981). (4) A; Callenis, L.I.Johansson, A.N. Christensen, K. Schwarz, P. Biaha.eand 0. Redinger, Phys. Rev. B 30, 635 (1984). (5) W.K. Schubert, R.N. Shelton, and E.L. Wolf, Phys. Rev. B 24, 6278 (1981). (6) J. Hermanson, Solid State Coranun. 22, 9 (1977). (7) A. CallenSs, L.I. Johansson, A.N. Christensen, K. Schwarz, P. Blaha, and J. Redinger, to be published.

A. CALLENAS Department of Physics and Measurement Technology. Linkoping University, S-581 83 Linkbping, SWEDEN P 2 B 2 VACANCY INDUCED CHANGES IN THE ELECTRONIC STRUCTURE OF TITANIUM NITRIDE

P. Herzig1, J. Redinger2, R. Eibler1 and A. Neckel1 ^Institut far Physikalische Chemie, Universitat Wien, Wahringerstrasse 42, A-1O9O Wien ^Institut fur Technische Elektrochemie, Technische Universitat Wien, GetreidemarJct 9, A-1O6O Wien

1. Introduction Like many other transition metal carbides and nitrides, titanium nitride exists in a wide range of homogeneity, TiNx (0.61^x.£l.OO) [1]. The vacancies are localized primarily on the aitcogen sublattice and a certain amount of short-range order can be deduced from electron diffraction measurements for x-values below 0.75 [2], Whereas the theoretical treatment of short-range ordered vacancy structures is a formidable task, the treatment of the two limiting cases of totally disordered vacancies on the one hand and of long-range order vacancy structures on the other hand is less demanding. We present a calculation of the electronic density of states (DOS) and charge density of the hypothetical ordered

TiN0<75. A muffin tin KKR CPA calculation of the DOS of substitu- tionally disordered TiNx will also be presented at this conference by Klima et al [3].

2. Method A self-consistent APW band structure calculation was performed for the hypothetical ordered compound Ti N3CJ3 (Figure 1). Eacn vacancy is octahedrally sur- 4 rounded by six Ti^ l atoms, which have T.M only four nitrogen atoms as nearest neighbours. The Ti'"' atoms are surrounded octahedrally by six N neighbours, as in stoichiometric TiN, A similar structure was already assumed for the calculation of the electronic structure O • of ordered TiCQ 7S [4J. A Hedin-Lundqvist exchange poten- Fig. 1. Unit cell for tial was used for the APW calculation. hypothetical ordered TiNo.75 P 2 B 2

3. Density of states of ordered TiN0>75

Figure 2 shows the DOS of TiN and ordered TiN0>75. The first and second peak, formed mainly by nitrogen 2s ("s band") and nitrogen 2p states ("p band"), show reduced width and intensity in the substoichiometric nitride because fewer nitrogen 2s and 2p states are present. The DOS in the energy region between -0.203 and +0.420 Ryd with respect to the Fermi level EF has mainly Ti-3d character ("d band"). In the case of TiN0#75 the DOS in che "d band" is distinctly increased below Ep. In this energy region two supplementary peaks appear with an appreciable amount of local partial s and p DOS, respectively, in the vacancy sphere. These two peaks, which cen therefore be attributed to so-cailed -1.0 "vacancy-states", have been observed -as E(Hyd) experimentally in the XPS spectra of Fig. 2. DOS for TiNand TiN at about 2 eV below E [5,6], x F TiN0.75 in states per Ryd per unit cell per spin (...: vacancy sphere)

Figure 3 shows the main local partial DOS components for ordered TiN0 75. In the "p-band" not only the N-2p but also the Ti^4J-3d DOS is greatly reduced, 100 indicating less p-d bonding in g{CE) the substoichiometric nitride. The main contributions to the two EF vacancy peaks come from Ti'^'-3d I1 1 J L states of dz2 and (dxz,dyz) sym- r .'•, metry. The additional vacancy peaks at the bottom of the "d • / as 10 band" compensate partly for the E(Ryd) lower intensity in the "s" and "p Fig. 3. Local partial DOS for band" leading to a decrease of TiN0.75 in states per unit cell the Fermi level compared with per spin ( : N-2s, -.-.: N-2p, the stoichiometric nitride. : TiI4]-3d, ...: Ti[6]-3d) P 2 B 2

4. Bonding mechanism In a coordinate system with the z-axis pointing from the Ti^4! atom towards the vacancy, the Ti^-d^ states have no nearest nitrogen neighbours in the +z and -z. directions and can therefore form p-d a bonds with only four N neighbours. Corre- spondingly, the number of p-d a bonds is reduced in the "p band" of

TiNQ 75. The bonding interaction of the dz2 orbitals of the neighbouring Ti [4] atoms leads to a net charge of s-symmetry in the vacancy sphere which is octahedrally surrounded by the Ti1 J atoms. This bonding mechanism can be tentatively classified as a Ti-Ti d-D-d a bonding interaction. Figure 4 shows the charge density in the (lOO)-plane between the vacancy, two Ti'4' atoms and an N atom for the state of symmetry X^ at 0.159 Ryd below EF.

Fig. 4. Charge density for Fig. 5. Charge density for state X^ of TiN0 75 in the state of 7c in the (100)-plane (in electrons/A*3) (lll)-plane (in electrons/A )

In the substoichiometric nitride the number of possible p-d 4 TT bonds involving Ti' ^-dv and dv_ states is reduced. These f A. 1 states can form bonds between the six Ti l*J atoms octahedrally surrounding the vacancy, thus stabilizing the metal octahedra. Figure 5 shows, as an illustration, the charge density in the 4 (lll)-plane between three Ti' ^ atoms for the state of symmetry X3 with an energy of 0.098 Ryd below EF. These octahedral d-d a bonds result from our calculation without the necessity of postulating a new kind of hybrid orbitals, as w*s done by Andersen and Satpathi [7] in the case of NbO. P 2 B 2

[1] S. Nagakura and T. Kusunaki, J.Appl.Cryst. LO, 52 (1977) [2] J. Billingham, P.S. Bell and M.H. Lewis, Acta Cryst. A2_8, 602 (1972) [3] J. Klima, G. Schadler, P. Weinberget and A. Neckel, tbis conference [4] J. Redinger, R. Eibler, P. Herzig, A. Neckel, R. Podloucky and E. Wimmer, J.Phys.Chem.Solids, in press [5] H. Hoechst, R.D. Bringans, P. Steiner and T. Wolf, Phys.Rev. B2_5, 7183 (1972) [6] L. Porte, L. Roux and J. Hanus, Phys.Rev. BJ2£, 3214 (1983) [7] O.K. Andersen and s. Satpathi, in "Basic properties of binary oxides", p. 21, edited by A. Oominguez Rodriguez, J. Castaing and P. Marquez, Publicaciones de la üniversidad de Sevilla (1984) P 2 B 3

ON" THE ELECTRONIC STRUCTURE OF

J. Klima(+) (1), G. Schadler (2), P. Weinberger (2) and A. Neckel (1)

(1) Instituc fiir Physikalische Chetnie, University of Vienna, Austria (2) Institut fUr Technische Elektrochemie, Technical University of Vienna, Austria (+) Permanent adress: Faculty of Mathematics and Physics, Charles University, Prague, Czechoslovakia

The electronic structure of substoichiometric TiNx is studied (1) in terms of the Korringa-Kohn-Rostoker (KKR)-Greens Function and the KKR-Coherent Potential Approximation. In accordance with recent XPS studies, peaks in the density of states (DOS) for the substoichiometric systems are found in the vicinity of the minimum of the DOS for the stoichiometric system. Such peaks are found for a single vacancy case, as well as for the concentrated vacancy case. In Fig.1 the density of states for TiN (full line) and TiNx (broken line), x- 0.75, are shown. In the substoichiometric system the Fermi energy (E^) is at lower energies than in the stoichiometric system.

glE) 20

10 •Em

to 10 al'io

* 9 C.ry 3 * .1 * 7 M Jt

Fig. I Fig.2 P 2 B 3

Concomitantly the DOS at Ep is about 10Z higher, indicating an increase in the linear coefficient of the specific heat for the subetoichiometric system. In Fig.2 th« main components of the local DOS'a to the total DOS are shown for the stoichiometric (full line) and the substoichiometric (broken line) system. It can ba seen that vacancies induce an additional

peak in the gap of the e (r.2)-like local DOS for Ti. The nature of the additional states in the minimum between the predominantly p-like non-metal and the d-like metal subband, can be seen from Fig.3, where the local DOS for a single vacancy in TiN (full line) is compared with the corresponding quantity for the concentrated vacancy case (broken line). The single vacancy case shows a well-pronounced virtual bound state. In both cases the main contribution to the local vacancy DOS is s-like. Fig.4 shows the condition for the occurence of a virtual bound state for a single vacancy. In the minimum of the DOS for the stoichiometric system the imaginary part of the scattering path operator (1) is very small. If there is a steep slope for the real part of the scattering path operator in this energy regime, the energetic position of the virtual bound state (E ) is rather insensitive to the scattering potential of the vacancy. This situation seems most likely to be the case also for other substoichiometric refractory phases.

* *

Fig. 3

(1) submitted for publication to J.Phys.F. P 2 B 3

•8 E, ry .A. ^

/ -.5

Fig.A

Financial support by the Hochschuljubilaumsfonds der Gemeinde Wien is gratefully acknowledged. P 2 B 4 ENERGY BANDS AND CALCULATED OPTICAL CURVES OF CoO. J. Hugel and C. Carabatos C.L.O.E.S. lie du Saulcy 57045 METZ CEDEX FRANCE

Abstract A band structure calculation using well localised orbitals obtained in a self-consistent manner has been performed on CoO. The localisation of the orbitals allows the treatment of both the valence and lowest conduction bands by the LCAO method. The detailed formulation of the crystal potential is given elsewhere (1). The energy band results are shown in figure 1 along the principal symmetry directions.The 2p band overlap slightly the 3d band in agreement with tht results of Eastman and Freeouf (2). The present band structure is an improvement in comparison with the structure obtained by Mattheiss (3) mainly because the energy gap between the 2p and 3d bands has been suppressed.

r A * Z W Q L A r Z KSX AfWtnrr w*s Figure 1. Energy band results and density of states for CoO. P 2 B 4 The imaginary part of the relative dielectric function e (at) has been computed in the electric dipolar approximation using the present band calculation. The separate evaluation of the various band to band transitions allots the assigment of the main peaks of the curve as can be seen on figure 2. The first peak at 5.6 eV is the most important and corresponds to the absorption edge. It cones exclusively from the transitions between the 2p and 3d bands.

Energy («v)

Figure 2 Imaginary parte* (w) of.the complex dielectric function together with the various band to band transitions.

The knowledge of the imaginary part of the dielectric function allows the obtention of all the optical constants. Figure 3 shows the calculated reflectance spectra. The position of the peaks as well as their magnitude are in good agreement with the experimental data (4). P 2 B 4

10 20 30 40 Enargy UV|

Figure 3. Calculated reflectance spectra (full curve) compared with experimental results of Powell and Spicer (1970) (dotted curve)

References 1) Huge! J., Carabatos C, Bassani F. and Casula F. 1981 Phys. Rev. B24 5949 2) Eastman D.E. and Freeouf J.L. 1975 Phys. Rev. Lett. 34 395 3) Mattheis L.F. 1972a Phys. Rev. B5 290 1972b Phys. Rev. B5 306 4) Powell R.J. and Spicer W.E. 1970 Phys. Rev. B2 2182 P 2 B 5

Self-consistent band structure and superconductivity in MoN and WN

Mei-caun Huang Department of Physics Xiamen University Xiamen, Fujian, China

and

T. Oguchi and A.J. Freeman Department of Physics and Astronomy Northwestern University Evanston, Illinois 60201 U.S.A.

Abstract

The total energy, self-consistent band structure, total and partial density of states (DOS) and Fermi surface sections on symmetric planes of

irreducible Brillouln zone for MoN and WN crystals with B1-type (NaCI) structure are calculated by the scalar relativistic linearized muffin-tin orbital (LMTO) method in the atomic sphere approximation (ASA) based on local density functional theory (LDF) with von Barth and Hedin's form for the exchange-correlation energy.

On the basis of our self-consistent band structure calculations, the electron-phonon interaction coupling constant \,and superconducting transition temperature, T , are calculated by using McMillan's strong coupling theory and Gaspari-Gyorffy's rigid muffin-tin approximation. The high predicted value of TQ ( 40 fC) should be taken to indicate that the perfect MoN (and WN) should be an excellent conductor.

The effects of the non-metal elements and fy>n-^*tal~vana^fj'i^a °n the superconductivity in both MoN and WN will also be discussed based on our theoretical results.

* This work was supported by the APOSR and supported partly by the Science P 2 B 5

Fund of the Chinese Academy of Sciences.

••also Department of Physics and Astronomy, Northwestern University, Evanston,

Illinois 60201 U.S.A. Oie wichtigste Investition für den Aufbau der Zukunft ist Vertrauen. Auch im Bereich der Technik ist erfolgreiche Zusammenarbeit nur auf der Grundlage gegenseitigen Vertrauens möglich. Als weltweit führendes Unterneh- men auf dem Gebiet der Vakuumtechnologie genieflt Leybold-Heraeus den Rufeines Partners, auf den in allen Situationen VerlaS ist Wirksame Investitionen zur Sicherung des Fortschritts leisten wir auch im Bereich der Aus- und Weiterbildung. Mehr als bisher ist dia Vermittlung von praxisorientierten Kenntnissen eine wichtige Voraussetzung in einer sich immer weiter technisierenden Wert BnegründlicheSchulung unserer Anwender gehört deshalb zu unserem Serviceprc- gramm. Die Seminare tragen mit dazu bei, daß quali- fizierte Fachkräfte auch den Anforderungen von morgen gerecht werden. LEYBOLD-HEHAEUS GES.M.B.H., Vakuumtech notogie FavoritenstraSe 35. A-1040 Wien P 2 B 6

PHCTCELECTRCN SPECTROSCOPY OF SCANDIUM NITRIDE LAYERS. L.PCRTE. Laboratoire de Physico-Chimie minerale, CNRS LA116, Universite LXCN I, 69622, VILLEURBANNE, FRANCE.

Mono-carbides, nitrides, and oxides of the first transition metal (TM) groups form a class of compounds closely related in their properties. Cne remar- -kable feature of these compounds concern their ability to cristallize with im- -portant deviations to the ideal stoichiometry. However while the non-metal defective B1 structure is the very cannon case with d and d TM, there is no convincing evidence of such defective structure in the case of d TM. In this work scandium nitride layers of different stoichiometries are analysed by X-ray and Ultra-violet Photoelectron Spectroscopies (XPS and UPS). Core level peaks and valence band structures are reported for the stoichiometric and non-stoi- -chianetric ScN layers. A main goal is to bring an experimental evidence for the existence -or not- of a substoichianetric ScN phase. Scandium nitride la- -yers were elaborated by the reactive sputtering technic. Metallic scandium is Ar- sputtered in a nitrogen gas atmosphere and scandium nitride layers with stoichiometry depending on the nitrogen partial pressure deposit onto a tanta- -lum foil substrate. XPS and UPS measurements were recorded using conventional unmonochromatized MgK^ radiation and Hel-Hell lines, respectively. The XPS ana- -lysis revealed no trace of contamination of the scandium nitride layers. The composition of the layers was obtained using the intensity ratio between the N1s and Sc2p core peaks. In figure 1 the experimental XPS and UPS valence band of stoichicntetric ScN are compared with the density of states (DOS) calculated by Neckel et al(1I The structure at about 15 eV binding energy (BE) originates from nearly pure N2s states while the valence band structure around 5 eV BE is built up from nitrogen 2p states which are strongly hybridized with scandium 3d states. The DOS falls to a minimum corresponding to the completion of these hybridized p-d states, then rises again due to states having mainly metal d character. In stoichianetric ScN the eight valence electrons are just acccmodated with the completion of the hybridized N2p-Sc3d band. Then ScN should be a semiocnductor if there is an energetic band gap between the p-d valence band and the metal d conduction band. A rough band gap value of 1.5 - 2 eV can be deduced from the low BE edge of the experimental UPS valence band. This is in the range of the 2.1 eV found for the optical band(2). The theoretical COS does not predict P 2 B 6

such a band gap. The DOS reproduces well the XFS/UPS spectra except that the ex- -perimental valence band and N2s band have BE 2-3 eV higher than calculation predicts. It is however noticeable that by simply shifting the DOS by the value of the band gap the agreement is far better. Figure 2 shows the Sc2p-N1s core level obtained for non-stoichiometric scandium nitride layers, for stoichicmetric ScN and for pure Sc. In the non- stoichicmetric layers the Sc2p, - photopsaks evidence the presence of two che- -mical entities. Par each layer the experimental Sc2p spectra can be resolved in two Sc2p doublets, one with BE close of the Sc metallic values, another with BE close of the stoichicmetric ScN values. The intensity of the latter increases with the nitrogen content. This result demonstrates that non-stoichiometric la- -yers are not nitrogen deficient nitrides of formula ScN but do correspond to mixtures of metallic Sc and stoichicmetric ScN. It appears that there is no creation of nitrogen vacancies in ScN as it is the common case in others TM com-

-pounds i.e. TiNx, ZrNx, TiCx- In these TM compounds removing a metalloid atan leads on one hand to sup- -press metal-metalloid interactions, on the other hand to create metal-metal interactions. The competition between these two kinds of interactions can explain the origin of large vacancy concentration if it results in a lowering of the to- (3 4) -tal energy of the compound ' . The most important difference in chemical bon- -ding between ScN and TiN results fran the increase of the strenght of the metal- -lic bonding for the latter. It could explain why deficient nitrogen scandium nitrides are not obtained. Removing a nitrogen atom in ScN would be energetical- -ly less favourable than in TiN, because the covalent N-Sc interactions and the covalent N-Ti interactions are nearly similar while the metallic Sc-Sc interac- -tions are distinctly weaker than the metallic Ti-Ti interactions. (1) A. Neckel, P. Rastl, R. Eibler, P. Weinberger, K. Schwarz, J. Phys. C: £, 579, 1976 (2) J.P. Dismukes, W.M. Yim, V.S. Ban, J. Cryst. Growth 13/14, 365, 1972 (3) L.M. Huisnan, A.E. Carlsson, CD. Gelatt Jr., H. Ehrenreich, Phys. Rev. B22 991,1980 ' (4) P. Pecheur, G. Toussaint, E. Kauffer, Phys. Rev. B1£, 6606, 1984 P 2 B 6

'\HeI ;' \ Sc2p \ f __ -. N1s

/ \ .. '. jell / H- ScN 1

intens i •

P r g1(E) i* j n /ij|\ i= »•"/ V

15 10 5 0 V Binding Energy (eV) £- 0 27 / •• / ^ •

Fig.1. High: experimental ScN Sc metal / ': valence band spectra obtained fron Hel (h)= 21.2 eV), Hell (h} = 40.8 eV) and MgK (hi = 1253.6 eV) radiations.

LowiLCAO partial density 410 405 400 395 of states for ScN, from Ref. (1). Binding Biergy (eV) s; —p; d. Fig.2, Core level spectra for Sc metal, ScN and nitrogen deficient scandium nitride layers. P 2 B 7

ON THE ELECTRONIC STRUCTURE OF STATISTICALLY DISORDERED TiCx P. Marksteiner and P. Weinberger Institut fur Technische Elektrochemie, Technical University of Vienna A. Neclcel Institut fiir Fhysikalische Chemie, University of Vienna

The density of states (DOS) of TiC- _- was calculated by ;he Korringa-Kohn- Rostoker Coherent Potential Approximation for complex lattices (I). The influence of the vacancies in the carbon sublattice on the electronic structure is very similar to the influence of nitrogen vacancies in substoichiometric TiN (1). The total DOS (?ig.l) has no longer a pronounced minimum between

states of mainly C-t]u (p-like) and Ti-e and t_ (d-like) character, as is the case in stoichiometric TiC (2). Thus, contrary to the case of TiN (1), the DOS at the Fermi energy is drastically increased, since in stoichiometric TiC the Fermi energy and the minimum of the DOS coi.icide (2).

TiC07S total DOS

Fig. 1

10-

0.3 cu 0.5 Ot 0.7 EIRyd) A comparison of the DOS at E~ for stoichiometric TiC (1), for an artificially ordered Ti^C- O (3) (O denoting a vacancy), and the present KKR-CPA results for TiC_ ., with experimental linear coefficients of tbo specific heat is shown in fig.2. The decomposition of the DOS into its main contributions is shown in figs.3-5. The DOS in^the region around the Fermi energy is dominated by states located in the titanium sphere and, above all, in the vacancy sphere. The vacancy DOS shows two pronounced peaks of aj (s)-character, and only minor t]u (p-like) contributions. This is in contrast to recent APW calculations for an artificially ordered Ti,C,D, where the second P 2 B 7 vacancy peak has predominantly p-character. Y (mJ/molK2)

Experiment

2- APWC33 APWC21

60 70 80 90 100 %C Fig. 2

TiC07C carbon sphere local partial t1u -DOS

Fig. 3 J0"

oj ai as ac a.'

TiCa76 titanium sphere 30 local partial e,- and t2g - DOS

Fig. 4 „.

a* as a* a; e<*r«) P 2 B 7

vacancy sph«r» 30- local partial a,,- and t^-COS 0,« r-, 0*7 aw an 0,70 0,71 0.72 0,73

_^ • — «tu Im-c 20-

-010-

10-

-020-

0.3 as as a? E(Ryd) -030 Fig. 5 E* \

-O40- \ s. R.T

Fig. 6 The peak at 0.685 Ryd corresponds to a well pronounced virtual bound state: The condition OO \ 11 Re rr° (1)

- where f is the scattering amplitude of the vacancy, f is the CPA oo scattering amplitude of the nonmetal sublattice, and Re ra. is the real part of the cell-diagonal ai -like scattering path operator - is fulfilled at the resonance energy EV B, as is shown in fig.6. The term on the right side of equation (1) is shown in dotted lines. At E the imaginary part of the scattering path operator Im r°° has Ig a minimum, its absolute value is, however, much higher than the corresponding value in TiN (I). Support by the "Hochschuljubilaumsstiftung der Stadt Wien" is gratefully acknowledged.

(1) J.Klima, G.Schadler, P.Weinberger, A.Neckel, J.Phys.F, 1985 (in print) (2) A.Neckel, P.Rastl, R.Eibler, P.Weinberger, K.Schvarz, J.Phys.C 9, 579 (1976) (3) J.Redinger, R.Eibler, P.Herzig, A.Neckel, R.Podloucky, E.Winner, J.Phys.Chem.Solids, 1985 (in print) (4) R.Caudron, P.Costa, J.Castaing, Solid State Camun.8, 621 (1970) (5) J.Landeajnan. Thesis, University of Strasbourg (1985) P 2 B 8

REACTIVE RP AMD DUAL IOET BEAN SPOTTER DEPOSITION OF Ti, Zr. AMD Hf HITRIDES D.S.Yee, K.SchwarZt J.J.Cuotao, J.M-E-Harper IBM Thomas J.Watsor. Research Center, Yorktown Heights, New Ifork 10593, USA * Permanent address: Institut fur Technische Elektrochemie der Technischen Universitat Wien, A-1O6O Vienna, Austria H.T.G.Hentzell, and B.O.Johansson Lir.kopir.g Institute of Technology, Linkoping, S-581 33, Seeder.

The nitrides of titanium, zirconium and hafnium have been prepared by reactive (radio frequency) RF and dual ion beam sputter deposition as described before. The mono-nitrides TiN, ZrN and HfN are prepared in a partial pressure of nitrogen in argon or krypton. They are golden colored, hard, good metallic conductors, high Tc superconductors, and good barriers to alumi- num diffusion similar to the properties previously reported. Films prepared from Zr and Hf in 100% nitrogen are straw colored, transparent and insulating. The composition shows a stable ratio at N/metal = 1.33, and exceeds this value under some conditions. The Ti films prepared in 100 % nitrogen are bluish and conducting, with a composition up to N/Ti = 1.25. An ordered defect structure containing 25 % vacancies of the metal sublattice is suggested for the insulating phase at the composition N/metal = 1.33. The suggested structure is denoted

Zr3N4 and is obtained by removing one of the 4 Zr atoms in the NaCl structure as shown in Fig.l. Electronic band structure calculations performed by the augmented spherical wave (ASW) method support the existence of an insulator for this defect structure. The metal vacancies have 3 the effect of making Zr3»4 an octet insulator, since it has 32 valence electrons (3 times 4 from Zr plus 4 times 5 from S) which can just occupy all the states of the 's* and 'p-bands' of nitrogen (8 states per nitrogen, i.e. 32 in total). P 2 B 8

Fig.l: Proposed crystal struc-

ture for Zr3N4- In terms of the MaCl structure the center Zr atom has been removed and this void is octahedrally surrounded by 6 N; the remaining Zr (open circle) and the other N (at the corner) keep their first shell coordination as in the NaCl structure.

The importance of this insulating phase stems from the fact that in combination with the superconductor ZrN it creates the possibility of building a common-lattice Josephson junction device with ZrN/Zr^N^/ZrN. in this case the superconductor and the insulator have essentially the same crystal structure (besides the defects) and the same constituent elements, i.e. it is isostructural and isochemical. This is not the case in the usual materials for Josephson junctions where lead-alloys or Nb are used for the superconductors separated by insulating oxides.

(1) J.H.E.Harper, J.J.Cuomo, and H.T.G.Hentzel, Appl.Phys.Lett. 43_, 547 (1983) (2) A.R.Williams, J.Kubler, and C.D.Gelatt Jr. Phys.Rev. B_19, 6094 (1979) (3) K.Schwarz, A.R.Williams, J.J.Cuomo, J.H.E.Harper, and H.T.G.Hentzel, Bull.Am.Phys.Soc. 29, 3 (1984) P 2 B 9

ELECTRONIC STRUCTURE OF TRANSITION METAL DISILICIDES K.Schwarz Institut fiir Techr.ische Elektrocheraie der Technischen Universitat Wien, Getreidemarkt 9, A-1O6O Vienna, Austria V.L.Moruzzi and F.M.d'Heurle IBM Thomas J.Watson Research Center, Yorktown Heights, New York 10598, USA

The transition metal / silicon interface is of special interest in many semiconductor devices. Various silicides can form at this interface. In this connection the observation was 1 made that the compounds Co2Si and Ni2Si have very similar crystal structure and lattice parameters and that they form solid solutions. Also, the disilicides CoSi2 and NiSi2 both crystallize in the CaF2 structure and both have about the same lattice constant of 5.43 A. The monosilicides, CoSi and NiSi, however, behave completely different: they differ in crystal structure and do not form solid solutions, but separate into layers. In order to understand this qualitative difference, for example in terms of solubility, their electronic structure has beer, investigated. The augmented spherical wave (ASW) method of Williams et al.2 has been used to obtain the self-consistent band structure results. These calculations are based on the local density approximation for treating exchange and correlation. There are already several bonding studies of transition metal silicides, for example the work by Weaver et al. In the present paper we shall focus on the disilicides NiSi2 and CoSi2. In the ASW method the unit cell is divided into regions according to the atomic sphere approximation (ASA). With the ASA the CaF2 structure can best be described by adding an "empty sphere" (i.e. a site with nuclear charge Z»0) for the interstitial region at the other tetraheccal position opposite to the Ca position. Thus in the ASW calculation the CaF2 structure consists of an fee lattice with a basis of four 'atoms': two equivalent fluorine positions, the Ca position, and the void (empty sphere); these three types of sites are used to decompose the density of states into site projected contributions.

Fig.l shows the band structure of CoSi2 as a representative example and Fig.2 the corresponding site- and t-projected densities of states (DOS) for both CoSi2 and NiSi2- P 2 B 9

V ' ,1 1.0 mm-

i ''•.t 0.5 -

••<.

0.0 -

W Q L A X Z V/ K

Fig.l. Energy band structure of CoSi2 in the CaF2 structure as obtained by ASW calculations.

The DOS of these two compounds look similar, but for the Ni compound the width of the d bands is smaller and the Permi energy lies higher than for the Co compound. In terms of a binding mechanism the following observation should be made: i) the Si-3s and Si-3p bands overlap as is aparent from Fig.l and the partial DOS of Fig.2; ii) there is a strong Si-p M-d interaction, where M represents a transition metal (Co,Ni); iii) the M-d DOS is mostly below the Fermi energy EF; this is more pronounced for Ni than for Co (Fig.2); iv) Ep falls around a minimum going from CoSi2 to NiSi2 and in a crude model Ep would go through this minimum in the case of an alloy; since Eg, in a minimum tends to lower the total energy, our result would indicate that the disilicides will readily alloy. P 2 B 9

-15 -10 -5 0 5 -15 -10 -5

Energy (in eV relative to EF) Energy (in eV relative to Ef )

Fig.2. Site and Jl-projected densities of states of CoSi2 and NiSi2.

This last observation is qualitatively different from the mono-silicides where a pronounced peak exists in the DOS which is occupied in the case of Nisi, but which remains unoccupied in the case of CoSi.1 Attempts to alloy or form solid solutions with the monosilicides will tend to place EF on this peak in the DOS and increase the total energy. Since this is unforvorable, the monosilicides should not alloy. Our findings correspond directly to the observed solubility differences in the mono- and disilicides of Co and Ni.

(1) D.D.Anfiteatro, V.R.Deline, F.M.d'Heurle, V.L.Moruzzi, and K.Schwarz, Thin Solid Films (in press). (2) A.R.Williams, J.KUbler, and C.D.Gelatt Jr. Phys.Rev. Bl^, 6094 (1»79) (3) J.H.Weaver, A.Franciosi and V.L.Moruzzi, Phys.Rev.B29, 3293 (1994) P 2 B lo

ENTHALPY gWKCTS IN THE Gd-Ni INTEHMETALLIC COMPOUNDS C. Colinet, A. Pasturel. Laboratoire de Thermodynamique et Physico-Chimle M&tallurgiques, ENSEEG, B.P. 75, 38402 Saint Martin d'Hereg Cedex, France. K.H.J. Buschow, Philips Research Laboratories, 5600 JA, Eindhoven, The Netherlands.

Introduction. Prediction of enthalpies of formation of alloys, AH, has been the object of considerable attention recently. Semi-empirical schemes employing model parameters as well as electron energy band approaches have been used with semi- quantitative success (1-5). Recently Pasturel et al (6,7) proposed a model for calculating enthalpies of formation of disordered solid alloys involving transition metals within a tight-binding scheme for the d band. Employing a physically plausible set of d band parameters, we were able to predict enthalpies of formation at any composition for transition metal alloys (8). This kind of tight-binding approach was also applied in rare-earth- transition metal alloys to calculate the enthalpies of formation of RNi- compounds (9). To check our theoritical predictions over all composition range in a rare-earth-transition metal alloy, we have determined the enthalpies of formation of the Gd-Ni intermetallic compounds using a calorimetric method.

Experimental results The Gd-Ni compounds were prepared by arc melting using purified argon gas. The purity of the starting materials was 99.9 % for Gd and 99.99 % for Ni. After arc melting the Ni-rich samples were vacuum annealed (4 w - 900°C). Lower annealing temperatures (700-800°C) were applied for the Gd-rich samples. For this purpose the samples were wrapped in a Ta foil and sealed into evacuated quartz tubes. After vacuum annealing it was verified by X-ray diffraction whether the annealing treatment has resulted in single phase samples. The enthalpy of formation of the Gd Ni compound corresponds to the in n reaction : P 2 B lo

i ts value is deduced from the dissolution enthalpies of the compound and its components in liquid aluminium using the relation :

H 9 + n 9 * *i ° ^."V ' " n where 0 (x » Gd, Ni, Gd Ni ) is the heat of dissolution of the elements or of x m n the compound at infinite dilution in aluminium. The enthalpies of dissolution of the Gd Ni compounds in liquid m n aluminium are obtained using the previously described calorimetric method .298 (10-12), they are given in Table I. Experimental values of AlC derived frsm these data for the seven intermetallic Gd—Ni compounds are listed in Table I. These values have been plotted as a function of Ni composition in Fig.l. In the two phase regions the enthalpy is indicated by means of the straight lines connecting the data points.

Table 1 : Experimental data of the solution enthalpies 0 obtained from solution experiments of Gd, Ni, Gd-Ni compounds from 298 K in liquid Al at a temperature T. The data listed for the pure elements Ni and Gd are those reported elsewhere by Chatillon-Colinet et al (12) and by Pasturel et al (13) respectively. The compositioposition range^^ considered in the solution experiments is x^x^. and x- < 00.55 at 55» . Q°" andd AAHH. are expressed per moll of atomt s (s : standardd deviation). T

Material TK Number of Q" s AH dissolutions (kJ/mol) (kJ/mol) (kJ/mol) Ni 948 17 - 119.75 2.01 Gd 966 9 - 143.09 0.72

Gd3Ni 999 10 - 118.01 1.25 - 18.2 GdNi 999 8 - 93.83 0.85 - 36.3

GdN.i2 995 9 - 88.50 1.67 - 38.5

GdNi3 995 10 - 86.65 1.00 - 37.6

Gd2Ni? 995 9 - 88.23 0.89 - 35.4 GdNi 995 10 - 90.93 1.67 - 31.3

Gd2Ni1? 995 8 - 100.77 0.89 - 20.0

Discussion. The enthalpies of formation have been calculated within a tight-binding scheme for the d band (7). The input parameters of the model are the number of d electrons of the elements, the relative position of their atomic energy levels and their bandwidths. Using the previously reported values of these parameters (8,9), the enthalpies of formation of the Gd-Ni compounds' are calculated and P 2 B lo

XNi Figure 1 : Enthalpies of formation of the Gd Ni compounds : experimental values, calculated values.

compared with the experimental results in Fig.l. The asymmetric 4H. curve observed in the Gd-Ni system is well represented. We have previously shown (7) that such a behaviour may be attributed to the large difference between Gd and Ni d-band parameters.

( 1) A.R. Miedema, P.F. de Chatel and F.R. de Boer, Physica 100 B, 1 (1980). ( 2) D.G. Pettifor, Solid State Commun. 28, 621 (1978) and Phys. Rev. Lett. 42, 846 (1979). ( 3) CM. Warma, Solid State Commun, 31, 295 (1979). ( 4) R.E. Watson and L.H. Bennett, Phys. Rev. Lett. 43, 1130 (1979) and Calphad 5, 25 (1981). ( 5) J. Van der Rest, F. Gautier and F. Brouers, J. Phys. F5, 2283 (1975). ( 6) A. Pasturel, P. Hicter and F. Cyrot-Lackmann, Solid State Commun, 48, 561 (1983). ( 7) A Pasturel, C. Colinet and P. Hicter, Acta Met., 32, 1061 (1984). ( 8) C Colinet, A. Pasturel and P. Hicter, Calphad, to be published. ( 9) A. Pasturel, C. Colinet, C. Allibert, P. Hicter, A. Percheron-Guegan and J.C Achard, Phys. Stat. Sol. (b) 125, 101 (1984). (10) J.C Mathieu, F. Durant and E. Bonnier, I.A.E.A. Vienna 1, 75 (1966). (11) N. Jeymond, D. Landaud, M. Legardeur and A. Pasturel, Thermochim. Acta 55, 301 (1981). (12) C. Chatillon-Colinet, H. Diaz, J.C. Mathieu, A. Percheron-€ue'gan and J.C Achard, Ann. Chim. (Paris) 8, 657 (1979). (13) A. Pasturel, C. Chatillon-Colinet, A. Percheron-Gu6gan and J.C. Achard, J. Leas-Common Net. 90, 21 (1983).

C. COLINET - Laboratoire de Therraodynamique et Physico-Chiaie MStallurgiques - LTPCM ENSEEG - B.P. 75 - 38402 SAINT MARTIN D'HERES CSdex - France.. P 2 B 11

HYBRIDIZATION EFFECT ON EXCESS ENTROPY OF TRANSITION METAL BASED ALLOYS A. Pasturel, C. Colinet, P. Hicter. Laboratoire de Thermodynamique et Physico-Chimie M6tallurgiques, ENSEEG, B.P. 75, 38402 Saint Martin d'Hires, France.

Over the last decade the binary alloys with a strongly non ideal mixing behaviour have been widely investigated in so far as these strong interactions lead to peculiar physical properties ; for instance transition metal-metalloid alloys A B (A = Fe, Co, Ni, Pd and B = Si, B, P, Sn, Al) are characterized by exceptional negative thermodynamic data of mixing (1) and by important evolution of the magnetic moment as a function of composition (2). The most popular interpretation of these experimental data assumes the existence of chemical complexes or "associates". Moreover there is no information on the mechanism stabilizing the associates and their use is full of objections. Evidently a more fundamental description should start form the electronic structure of these alloys and could allow a better understanding of their behaviour.

The major physical effect in these alloys is the covalent bonding between the d states of the transition metal and the sp states of the non transition metal ; when two atoms are brought together their states hybridize with one another to form bonding and antibonding states. That is in fact the well known interaction that leads to molecular bonding and to band broadening in solids. This hybridization causes a strong depression in the sp-states density at energies near the top of the d band of the transition metal (3). This depression acts like a gap, preventing metalloid states from passing through the Fermi energy and thus holds the number of sp electrons constant (2). This peculiar shape of the density of states allows to determine the critical composition x* for which the Fermi level is located in the pseudo-gap. The number of states, located below the pseudo-gap are given by : 1Ox + n V - 8p (D where n is the constant number of sp states below the pseudogap.

Spin-polarized band-structure calculations indicate that n should be equal to sp 1 for f.c.c. alloys, and 0.8 for b.c.c. alloys (2). P 2 B 11

The number of valence electrons i3 given by :

av ™ A " a where Z and ZQ are the valence of the A and B elements respectively.

The critical composition is then given by equalizing equ.(l) and (2) :

x* = B ~"SP (3)

10 + Zg - ZA

Recent theoretical approaches have shown that the exceptional negative entropies of mixing in transition metal based alloys are mainly due to the electronic and magnetic contributions of entropy. Indeed the electronic contribution AS is proportional to the density of states at the Fermi level and for a transition metal, due to its high density of states, this contribution is important ; when it is alloyed with metalloid, there is a change in the electronic entropy and at the critical composition, for which the Fermi level is located at the top of the d band of the transition metal, the d-electronic contribution becomes zero. So the electronic contribution to the entropy of mixing at the critical composition is given by :' AS . = - x» S . . (4) el el,A where S . is the electronic entropy of the A transition metal. Beyond x*, there is a linear change in electronic contribution ; in the composition 0 < x < x*, if we assume that the d density of states of the transition metal decreases linearly with metalloid composition, we also obtain a linear evolution of the electronic contribution to the entropy.

A similar approach can be used for the magnetic contribution to the entropy of mixing ; according to the Friedel's result (4), the atom-averaged moment in Bohr magnetons is :

"av = 2 N+ " Zav <5> and with the fact that the number of sp electrons is constant, it becomes advantageous to write equ.(5) in the form : - V ,Bt)- ZB with N •» 5 for F«, Co, Ni and N. _+» 0 for metalloid. d,A d,B P 2 B 11

At the critical composition x*, the atom-averaged moment becomes zero and the magnetic contribution is then given by : AS = - x* S . (7) mag mag.A where Sma g .A.is the magnetic entropy of the A element,

As conclusion, the present model provides a determination ~? the critical composition for which the Fermi level is located in the pseudo-gap and accounts successfully for the large negative AS values observed for this composition.

(1) R. Hultgren, P.D. Desai, D-.T. Hawkins, M. Gleiser and K.K. Kelley, 1973, Selected Values of the Thermodynamic Properties of the Elements, American Society for Metals (Metals Park, Ohio). (2) A.P. Malozemoff, A.R. Williams and V.L. Moruzzi, Phys. Rev. B 29, 1620 (1984). ( 3) D. Mayou, D. Nguyen Manh, A. Pasturel, F. Cyrot-Lackmann, J. de Physique (to be submitted). (4) J. Friedel, Nuovo Cimento, Suppl. to Vol. Ill, 237 (1958).

C. COLINET - Laboratoire de Thermodynaaiique et Physicc—Chi«ie Metallurgiques - LTFCM - B-P. 7R - 3ftdTW> RITiyr MAPTTM n'HUHES r.ftA*-x . F-rnnr-.m. P 2 B 12

STRUCTURAL TRANSFORMATION AND SuPERCONDU'TnVlTT IN THE ZrCRt^ Pd ) SYSTEM R. Kueticzler Institut de Physique, O.L.P., C.N.R.S., L.A. 306, 67084 Strasbourg, France R.M. Waterstrat Health Foundation, Washington, D.C. 20234, U.S.A.

The structural transformation often observed in alloys with a CsCl type structure is associated with various properties like shape memory phenomena and has therefore been the subject of a wide activity. But many questions are still open even for the archetype TiNi. The concurrence of structural transformation and high-temperature superconductivity has stimulated extensive investigations both experimental and theoretical on these two phenomena. This is particular true for the A., compounds. The alloys with a CsCl type structure are good candidates for the study of martinsitic transformations as a function of concentration and V-Ru has already been largely studied (1). It has been shown in particular that the electronic specific heat coefficients reach high values near the cone;ntration corresponding to the appearance of the structural transformation. Therefore it has been proposed that the structural transformation is electronic in nature and results in a large drop in d-electron densiry of states in going from the cubic to the tetragonal phase. Alternatively, on the basis of sound™ velocity measurements in particular, it was suggested that the high T , superconducting transition temperetare, is caused by the softening of the lattice. A recent theoretical paper reconsiders for the A., compounds the explanation of the high Tc (2). The CsCl-type pseudo-binary system Zr(Ru, Pd ) presents interesting proper- A^X X ties probably related with a structural transformation. We have undertaken a study of this system and its properties. Here we present low temperature specific heat results in the range of temperature 1.4 - 20 K over the whole concentration range 0

(1) C.W. Chu, S. Huang, T.F. Smith, E. Corenzvie, Phys. Rev. J5U, 1866 (1975). (2) C.C. Yu and P.W. Anderson, Phys. Rev. B29, 6165 (1984).

R. Kuentzler Institut de Physique 3, rue de l'Uaiversiti F 67084 Strasbourg P 2 B 13

LOCAL STBDCTDBE IN THE COMPUTER SIMULATED TRANSITION METAL- METALLOID ALLOYS A. Jezierski Institute of Molecular Physics, Polish Academy of Sciences, Smoluchowskiego 17/19, PL-60-179 Poznan, Poland

In the last decade numerous compute-: models of dense random packing of hard spheres were constructed, namely by Bernal (1), Finney (2), Boudreaux (3), Fujiwara and Ishii (4), and by Gaskell (5). Many of the radial distribution functions are in satisfactory agreement with the observed ones. In transition metal - metalloid alloys, the topic of the present work, the i^mjortant question is that of the existence of chemical short range order. Gaskell (S) developed a model for glass using trigonal prisms.. When he applied it to the amorphous alloys of PdSi and FeB he obtained good agreement with experimental data. Boudreaux and Frost (7) had analyzed large computer generated models simulating metallic glass alloys and they found that in an amorphous alloy of FeB over 50Z of Bernal holes have shapes of a dihedron or a trigonal prism. It it the aim of the present work to analyra the local structure in binary transition metal - metalloid alloys using computer-simulation models. Our model consists of a cluster of 200-300 atoms which are distributed inside a sphere and should represent the binary alloy; for the distribution of the atoms we use two different algorithms. In model 1 the Bennett (6) algorithm is applied. Model 2 is constructed by spherical clusters consisting of a metalloid atom at the central site and various nearest neighbours; the radius of the cluster vas equal to 1.87 d (where d ;.s the diameter of a metal atom). The coordinates of atoms in both models were chosea randomly. The type of atom added to the existing atoms, metal or iietalloid, was determined such that the ratio of metalloids to the total number of atoms was not greater than the average concentration c (which we assued to be c « 0.2) of metalloids in the alloy. After such clusters had been generated a nearest neighbour analysis was made for each atom not eX the surface. The average number of the neighbours for metal atoms and for metalloid atoms was found for both models. The probability distributions of the number of nearest neighbours are presented in Fig.1-3 for both uodels. la these P 2 B 13 aalculations we have fixed the ratio iu diameters between metalloid and metal to 0.72, a value which should correspond to the FeP amorphous alloy.

Model I Model 2

* 11 13. IS

Fig.I Fe atoms around Fe atoms

Model 1 Model 2

Fig.2 P atoms around Fe atoms P 2 B 13

Model 2 Model 1

n e s * 9 11 13is

Fig.3 Fe atoms around P atoms

From a statistical analysis we can deduce

9.03 (8.93) and NFeP - 1.98 (1.73), respectively. NFeFe

(]) J.D. Bemal, Nature, J£8, 910 (1960) 2 1. Finney, Proc. R^sSc. A319, 479 1970) 3 D.S. Boudreaux, Phys. Rev. B18, 4039 1978) 4 T. Fujivara and Y. Ishii, J. Phys. F 10 190 (1960) S ld (5) P.H. Gaskell, J. Non-Crysty . f,f,^§2727 ^§'(((^^972 (6) C.H. Bennett, J. Appl. Phys. 43, n ' ^ f9gI) (7) D.S. Boudreaux and H.Jl .P Frosth 4,3 Phys 2727(. Rev'97. B23^ , 1506 U9»i)

tititutfof Molecular Phy.ics, Polish.Ac«d«nyof Sci.nces, Smoluchowskiego 17/19, PL-60-179 Pozn«n, Poland P 2 B 14

SUBLIMATION, DIFFUSION. MELTINQ, BO I UNO : UNIFYING MODEL FOB METALUC AND SEMIMETALUC ELEMENTS AND FOR SOME INTERSTITIAL COMPOUNOS (CARBIDES, NITRIDES)

J. Wach

L.A.R.I.G.S.. Laboratoire M. Letort. C.N.R.S., Laboratoire assocle a runlverslti de NANCY I - B.P. 104 F - 54600 VILLERS-LE5-NANCY

This thermodynamics! model Is baaed on a new concept which allows the free energy of formation ot point detects (vacancies and self-interstltlals, and their clusters) to be linked with their migration energy, and lit particular to their transfer energy into the bulk from where they are formed (kink of ledge or dislocation line) via Generation ot equilibrium point detects by diffusion the superficial terraces or the regions around from kinks, atomic key- the core ot the dislocations (Fig. 1, 2, 3). positions In Volmer halt- crystal As an example, let us take two vacancies L (Lacunes, fig. 2), one In the ledge LM, the other one In the terrace Lj. Between these two types ot equilibrium vacancies, the barriers for exchange (Fig. Generating mechanism of a 3) are linked to their formation energies. bulk vacancy by step by step I.e. the transfer energy of L^. which transfer from a converts Ly Into Lj, has to be equal to the formation energy of Lj. Reciprocally, the migration enirgy of Lf, which converts Lj Into L/u, has to be equal to the formation energy of L/y. This exchange law expresses In fact «."» conservation of energy In dynamical equilibrium.

The enthalpy E$ and the entropy Sg ot sublimation are two fundamental key-data. Energy diagram tor the proposed mechanism ot They correspond Indeed to two extreme generation of vacancies with atomic transfers (Fig. I). On on* hand. the new concept P 2 B 14

they oorraapond to tha tranatar lumps ot kink atoms (atomaa da decrochement. D) Into tha gaa phaaa In equilibrium with tha crystal. On tha othar hand, thay corraapond to tha tranatar of any bulk atoms to these hay positions D. by concomitant step by step exchange with vacancies. Thus, all the tree energies ot formation and migration ot the

point detects In the bulk, at the surface The dlffuslvlty spectrum for (terrace and ledge) and along the grain metals (8) boundaries (near and along dislocation lines) are expressed In terms of Eg and Sg

CSS = SLtSM * 25,5 <± 4.5) + 2,3 (* 0.5; & 28 (± 5) e.u.. S. and SM being the vaporization and melting entropies (1)}. The theoretical expressions show that the two diffusion mechanisms (by vacancies ar ai aa and by dumbbells) are equivalent, at all temperatures. There is a good fit with the experimental diffusion coefficients In bulk, at grain boundaries and surface of metals, as In liquids (Fig. 4, 5, 6). At malting temperature Ty, the sources of point defects become activated at each atomic vibration. Therefore, we mtst have Surface self-dtffuslon tor Eg * 2S$TM - 567"M. In good tgraement FCC and BCC metals (11) with experimental values

The modei provides all th» empirical rules correlating activation energy Q tor diffusion with melting temperature of metals

(Fig. 4-6, 8). tor bulk (Q =• 3STM) (7), for 8 grain boundaries (Q • 21T^f) (8,9) and liquids (0 a TTfJ {S-9} and for surfaces from low temperatures (Q - 14T/^) to very high ones, near melting (Q - 58T^) (8-11). This unifying model can be generalized to seml-metalllc elements and to interstitial compounds : for monocarbides tor example (Fig. 9-10). the knowledge of the partial heats of sublimation allows the determination of all the semi-empirical rules (12) working tor carbon and metal self-diffusion data. Bulk telf-dltfusion in Carbi- des (12). TpQ are Congruent Mononitrldes (Fig.11) (13) are under scope. Melting Temperatures

(1) K.A. Gschneidner Jr

Solid State Physics. £6 (1964) 275 IK w (2) W. Shockley r^t| JM •»_*••» «;• «•;• W).H •»• :» Report of the 9th Solvay Conference, it.t r%J «j »Jtua ^i«j M . if l.t «4 ; tM Ut T4 IJM . I.* 431 (1951); Ft. Stoops. Brussels tn^iW/^ lj*** lir* *.«•** " i.m** ' •.«#•* i.w' 10

(3) J. Friedel VTK 17^ n.t a* \^ nj it.i 'Dislocations', Perg. Press (1984) (4) D. Kuhlmann-Wilsdorf

Phys. Rev., A 140, 5 (1965) 1599 !.f. »rai- *..»- r«wr. JJI. *»(• i (5) R.M.J. Cotterlll and Coworkers J.Phys.. 36 (1975) C2-35 Phil. Mag.. 27 (1973) 623 § i 33 (1974) 229 and 245 nun 32 (1975) 1283 ; 36 (1977) 453 (6) T. Nlnomlya J. Phys. Soc. Jpn.4±(f) (1978) 263 (7) Y. Adda and J. Phlllbert 'La diffusion dans les soildes', 510 Press Univ. France (1968) (8) N.A. Glosteln In 'Diffusion', ASM. Ohio. (1973) (9) P. Gulraldenq 11 J. Phys., 36 (1975) C4-201 (10) G. Neumann and W. Hlrschwald Z. Phys. Chem.. NF 8± (1972) 163 (11) H.P. Bonzel In 'Surface physics of materials', vol.11. Academic Prwsa. NY. (1975)

PROPERTIES AND ELECTRICAL CONDUCTIVITY 0? THANSXTICW METAL OXIDES V.B. Lazarev and I.S. Shaplygin Kumakov Institute of General and Inorganic Chemistry, Akademy of Science, Leninsky prospakt 31, Moscow, 117071, USSR

Metal oxides belong to one of the oldest and largest classes of inorganic compounds. Transition metal oxides have been extensively studied in the last years as they are the base of new materials for the production of ceramics, refractories, catalysts etc. In the present work properties and electrical conductivity of single and double transition metal oxides are presented. 3d-element monoxides have been extensively studied and their properties have been described in reviews (1,2). Monoxides with rocksalt structure TiO and VO for 0.75 tx £ 1.30 and NbO for 0.98 4x6 1.02 exhibit metallic xx „ conductivity in the homogeneity region and have resistivities of 3.1x10 , -3 -5 1.9x10 , and 2x10 ohm.cm respectively- at 300 K. Below 0.7 and 1.5 K

TiO. _0 and NbO .- respectively are superconductors. The other 3d-element monoxides are insulators or semiconductors. Manganese monoxide MnO has a 9 15 resistivity in the order of 10 - 10 ohm.cm and an energy gap E = 3.2 eV. Below 118 K it shows antiferromagnetism. Iron and cobalt monoxides are semiconductors. For defect Pe .0 the resistivity may have values ranging 5 8 from 10 to 10 ohm.cm in dependence on the type of cation sublattice ordering. CoO has cation vacancies too. It is a p-type semiconductor; below 293 K antiferromagnetic ordering has been found with a distortion from cubic symmetry. The resistivity of pure CoO has values ranging from 10 to 10 ohm.cm. Stoichiometric pure NiO is an insulator with j> = 10 ohm.cm. and E * 3.9 eV. It is antiferrom«;netic with the Neel temp' *ture at 523 K. Diamagnetic PdO is a p-type semiconductor with C = 10 - 100 ohm.cm (depending on admixture level) and E - 0.05 - 0.10 eV.

0 nd V exn t Sesquioxides Ti2 3 * 2°3 iki semiconductor-to-metal transitions. The high-temperature phases are metallic, while all three forms of Kh?0^ are p-type semiconductors with resistivity values of 60 - 300 (a-form),

7-30 (B), and 130 ohm.cm (y ); Ea« 0.20, -0.1, and 0.16 eV respectively. All forms of Rh_O, are diamagnetic. Rutile-type dioxides have many interesting properties too. Among than aze insulators, semiconductors and metals. A nuraber of properties of transition metal dioxides and their «n«rgy diagrams have b««n reported in (2). P 2 B 15

TABLE

Properties of some double oxides with perovskite-like structure

298 X Shortest distance Oxide Symmetry ohm •cm M-M in octahedra Magnetic

MO6, A properties

LaTiO, cubic 8x10~4 5.55 TIPC1.34) j 15 3 LaCoO, item 2xl0" 4.41 TIP 3 rhombohedr. 5xlO~3 4.42 TIP LaNiO3 LaRuO, orthorhombic 4.2xlO~3 5.49 AF-P(2.49) 3 BaCoO, hexagonal 6.6x10"" 2.38 TIP J 3 BaNiO, item lxio" 2.416 TIP 3 BaNbO, cubic lxio" 2.040 •> J BaMoO, item lxio"3 2.020 TIP J 3 2) 3 BaRuO, hexagonal 2.8xlO" 2.55 -5.60 ) TIP BaOsO, item I.OXIO"2 2.602)-5.653) TIP J 3 2) 3 BalrO, item 1.2xlO" 2.60 -5.75 ) TIP CaRuO, orthorhomb ic 5.5xl0"4 5.35 AF(T = 110 K) CalrO, item 4.5xl0"4 5.40 TIP CaOsO item 4 8x10"3 5.45 TIP

4 Sr.iv o? item 4.0xl0~ 5.50 P(T = 160 K) C 4 SrIrO3 •tern 5.2xlO" 5.40

3 •} SrOsO3 6.9xl0" 5.45 CaVO, cubic 2.2xlO"4 — TIP CakoO, orthorhombic lxio"3 - TIP SrMoO, cubic lxlO"3 TIP SrCrO. item TIP

TIP - temperature-indapendent paramaqn<:f.ism, AF - antiferromagnetism, F - ferromagnetism, P - param&gnetism, in parenthesis effective magnetic 2) moment in Bohr nagnatons. above 940°C. within cluster. between clusters. P 2 B 15

Double oxidas of transition elements have also interesting chemical and electrical properties. In the Table double oxides of transition metals with perovskite-like structure exhibiting metallic conductivity are given. Metallic conductivity of the double oxides was found for other structural groups too, namely for stoichiometric or defect pyrochlores (for instance Bi.Ru_O_, Pb.Re O,, T£_Rh,0 , Ca.Os 0_), alkali molybdenum, tungsten, Z z I Z Z b Z Z 1 Z Z I rhenium and ruthenium oxide bronzes, spinels, monoclinic double oxides like Li-RuO , and so on. Very interesting phases are catalytic active La_CuO and La.NiO and their solid solutions because of the metallic conductivity of La CuO. at 50-950 K (experimentally) and of \AJ&i.Q above 780 K. In the solid solutions semiconductor-to-metal transitions were found. Seme of the phases are anti- ferromagnetics because of a superexchange i_i Cu -o -Cu chains which are located in the xy plane. Some theoretical models to explain the electrical properties of oxides.are discussed in the paper too.

(1) V.B. Lazarev, V.G. Kraso'; and I.S. Shaplygin. Electrical conductivity of oxide systems and films. Moscow, Nauka, 1979, 168 pp. (in Russian). (2) V.B. Lazarev, V.V. Sobolev and I.S. Shaplygin. Chemical and physical properties of simple metal oxides. Moscow, Nauka, 1983, 239 pp. (in Russian).

Prof. Lazarev V.B. Kurnakov Institute of General and Inorganic \3Mmistry, Leninsky prospekt 31, Moscow, 117071, USSR. SPIN-GLASS BEHAVIOUR IN CONCENTRATED CHROMIUM SPINELS. P3 Al E. Agostinelli and D. Fiorani ITSE-CNR, Area della Ricerca di Roma, CP 10 — 00016 Monterotondo Stazione (Roma) - ITALY

Spin-glass behaviour has been observed in the system Zn Cd Cr S x 1-x 2 4 for the composition range 0.3

is an antiferromagnet (T =18K, ©p = 18K) while CdCr S is a ferromagnet (T =85K,©f=l52K). Zn Cd _Cr?S represent a very peculiar system because in this case the spin-glass state is not a consequence of the disorder induced by the dilution. The frustration (contradiction of interaction) is the only one responsible for one existence of the spin-glass state because the system is a fullv concentrated magnetic material. The spin-glass state is a result of the presence of competing nearast neighbours ferromagnetic interactions (J >0) and antiferromagnetic interactions with higher order neighbours (J?< 0). The J /J ratio changes with the composition x and finally for J = -J minute fluctuations should be sufficient to destroy the magnetic long range order and lead to the spin-glass type effects. We observed a spin-glass behaviour in the composition range 0.3

EXPERIMENTAL DETAILS Powdered samples have been sinthetized by direct reaction of the corresponding sulfides at T » 900°C (4 days) in sealed quartz tube. X-ray diffraction shows that the samples are composed of single phase spinel. A.C. susceptibility measurements ( H <* (—) ) ac dH II *o were performed at V = 198 Hz using a mutual inductance bridge.- The P 3 A 1 measuring coil consists of a primary and two identical, connected in opposition., coaxial secondary coils. The insertion of a magnetic sample in one of the two secondary coils induces an e.m.f. to which a voltage across the mutual inductance M and a calibrated resistance can be opposed. At balance M is proportional to the real part of the susceptibility. Low field (H = 400 Oe) D.C. susceptibility measurements ( Kri = M/H) were performed with a Fara- day balance after zero field cooling ( *A ) as well as after cool Zr C —• ing under the measuring field ( K__).

RESULTS AMD DISCUSSION Three different regions of composition x have been observed with different magnetic behaviour. For 0^x^0.3 ferromagnetic intera£ tions are dominant. When a fraction of Zn ions is introduced in CdCr S lattice the ferromagnetic structure appears strongly perturbed . The A.C. susceptibility curve ( in phase component ft ) (Figl) shows a large maximum centered around 60 K, which should

A. X-0.3 20- l\ • « '. •* •• * • e 10 - •* *

50 100 T(K) Fig.l - Thermal variation of (V = 198Hz) for x = 0.3 A.C. reflect the presence of ferromagnetic regions (for x=1.0,T =85K) of different dimensions. At lower temperature a strong peak is ob served indicating the occurrence of the spin-glass freezing (T = 16.6K for x = 0.1; T =15.8K for x=0.3). For 0.3 0.7 the data indicate an antiferromagnetic region. A large maximum is found for x=0.85.(Fig.3) centered at T=18OK, indicating a residual perturbed antiferromagnetic order with respect to x=1.0

(TN=20.0 K).

F.C. 70 ,ooo0 .

X'0.5 20 60 - Z.F.C,

X=0.85 50- \ V 10 •»

40- \

10 20 30 40 50 10 20 30 T (K) T (K) Fig.2 - Thermal variation of Fig.3 - "hermal variation of DC (H=400 Oe) for x = 0,5 ^A C (y"198Hz) for x = °-85

Measurements at lower temperature are necessary to identify the nature of the increase of the susceptibility curve below 13 K. In conclusion a preliminary study of the magnetic phase diagram of the system Zn Cd Cr S indicate that the competition between n.n. ferromagnetic and n.n.n. antiferromagnetic interactions stabilizes a spin-

REFERENCES (1) S. Viticoli, D. Fiorani, M. Nogues and J.L. Dormann, Phys. Rev. B 26_, 6085 (1982) (2) D. Fiorani, S. Viticoli, J.L. Dormann, J.L. Tholence, J.Hammann A. P. Murani and J.L. Soubeyroux, J. Phys.C, X6_, 3175 (1982) (3) D. Fiorani, J. Phys. C Y7_, 4838 (1984) (4) D. Fiorani, S. Viticoli, J.L. Dormann, J.L. Tholence and A.P. Murani, Phys. Rev.B 30, 2776 (1984)

Dino Fiorani ITSE-CNR, Area della Ricerca di Roma, CP 10 - 00016 Mnn1-i»r.n-f-nnrln Utarinno (Rrvna) - TTAT.V P 3 A I

OF THE HELICAL-FERROMAGNETIC TRANSITIONS IN MnAs, P -CRYSTALS K. Barner, Ch. Kuhrt k. Pny3. Institttt der Universitat Gottingen, Bunsenstr. 11-15 D-3UOO Gottingen A.F. Andresen Institute for Energy Technology, U-2007 Kjeller

MnAs can be retained in its MnP-type structure by application of external pressure. A reduction of unit cell volume can ^\lso be obtained by alloying with phosphorous. The magnetic properties of both systems are found to be equivalent /I/. In particular, with pressures 6 < P < 12 kbar or 0.05 < x < 0.13 one ob- tains a succession of magnetic order-order transitions (fig. 1, T. ,T. ). A helical to ferromagnetic transition has been found both at T., and T /2,3/. Judging from the magnetization curve and from the temperature dependence of the satellite reflections both transitions and the adjoining magnetic phases appear to be perfectly symmetrical. A look at the transport properties, however, reveals differences. Fig. 2 shows the resitivity p and the magnetoresistance Ap/p. While for the upper helical phase (i) Ap/p is small and positive, for the lower helical phase (II) it is larga and negative. p(T) itself is governed by spin disorder scattering and except for the high spin - low spin transition in the paramagnetic range nothing spectacular happens between T,< T < T^ A/; in contrast, for 0 < T < T, we observe a p(T )-anomaly. If one measures the carrier concentrations (fig. 3) one finds twc types of carriers (electrons, holes) and a different change in their concentrations n,p at T. and T. /5/. n,p appear to be constant in the existence regions of the magnetic phases suggesting a scattering process (in II) whose cross section increases on going to lower temperatures. Measurements of the thermoelectric power confirm this result. If one compares S(T) of x = 0.03 (which has no T^) and x = 0.10 (rig. 1*) one finds an extra contribution for T < T. while there is no significant change when T1< T < T-. A spin-flip type of scattering involving an excitonic unit has been proposed to be effective in II /6/, however, at present it is "till unclear how it can exist with conduction electrons and magnetic long range order present. Clearly, the average magnetic properties would not be affected too much by these scattering centers. This is confirmed by recent measurements of the magnetocaloric effect. P 3 A 2

ig. 1 Magnetic moment and Intensity of

if satellite reflection of MnAsQ ggP ^ versus temperature (T Heel-points, TJj. irder-order transition temperatures (h:helica. : ferromagnetic, p.: paramagnetic lov spin).

300K

Fig. 2. Resistivity p and magnetoresistance Ap/p of two spcimens (x=0.105 and x=0.1l)(p. paramagnetic high spin, T. second order MnP •«-»• NiAs phase transition spin disorder scattering) P 3 A 2

100 200K 400K Fig. 3. Reciprocal normal Hall constant 1/eR of two samples (x=0.11, x=0.10) versus temperature (T, ..: high spin - lov spin transition temperature). hi

Fig. h. Thermoelectric pover S of two whiskers (x=0.03, x=0.10) versus temperature. -.-.- diffusion TEP, TEP without T_.-anomaly.

References: 1. M. Menyuk, J.A. Kafalas, K. Dwight and J.B. Goodenough, Phys. Rev. 177, 9^2 (1969) 2. H. Fjellvig, A.7. Andresen and K. Barner, JMMM (198U) 3. S. Stolen, H. Fjellvag, A.F. Andresen and A. Kjekshus, private communication k. K. Barner, U. Neitzel, Phys. lett. 91A, 36l (1982) 5. H.J. Kxokoszinski, Thesis, Gottingen (1982) 6. H.K. Kroazinski, K. Barner, Phys. lett. 96A, 207 (1983) K. Barner, k. Physikalisches Institut der Universitat Gottingen Bunsenstr. 11-15, D-3^00 Gottingen p 3 A :

MAGNETIC PROPERTIES OF VITREOUS PHASES OF THE SYSTEM

S. Barnier, M. Guittard, M. Wintenberger and J. Flahaut Laboratoire de Chimie Min6rale Structurale Associ<§ au CNRS - LA 200, 4, Av de 1'Observatoire, 75270 Paris, France.

The preparation, thermal behaviour, optical and E.P.R. spectra of these glasses have been described in (1). Fig. 1 shows their composition range. The magnetic susceptibilities x have been measured with a V.S.M. from 78 to 300 K. In this temperature range they follow a Curie-Weiss law x = c/(T-9 ) (Fig. 2). The value of [0 I in- 2 + creases with the Mn concentration along the BK line of fig. 1 (table I) . The influence of the La^O^/Ga^S, ratio does not seem to be important. The |9 | values are quite large if compared to those of at MnGa_S4 (-50 K) , 3 MnGa2S4 (-28 K) or a wurtzite-type solid solution 0.15 MnS, 0.85 GaS1 5 (-35 K). In a MnGa2S4 there are mainly 90° Mn-S-Mn interactions. In the other two phases there are probably mainly super-superexchange interactions via two S and one Ga. In the glasses one can assume that there is a large number of # 180° Mn-S-Mn interactions via shared vertices of octahedra. For Mn-rich samples low field a.c. susceptibility measurements (10 Oe - 65 Hz) have been made from 1.2 to 4.2 K. Samples G and E show a spin glass-like behaviour with a susceptibility maximum at a well-defined temperature T (Fig. 3). For sample K one sees only a variation of the slope of the x (T) curve. In this case there are probably some isolated Mn + ions with magnetic moments which do not freeze even if the majority of Mn moments undergo a spin glass transition.

The Ts /|e | values are of the same order of magnitude (about 0.02) as those of the glasses of the system MnS-Ga-S^-GeS- (2) . P 3 A 3

Table I

Characteristic temperature 9 and T sg

Sample Composition e T P sg MnS GaS1.5 LaOK5 G 0.233 0.517 0.25 -125 K 2.4 K K 0.233 0.567 0.2 -130 2.2 K ? N 0.233 0.667 0.1 -95 no me as. E 0.2 0.6 0.2 -100 1.7 K I 0.166 0.534 0.3 -95 nothing for T>1.4K D 0.15 0.65 0.2 -68 no meas. C 0.1 0.7 0.2 -46 ii B 0.05 0.75 0.2 -18 rl

Wurtzite- 0.15 0.85 -35 II type solid solution

Fig. 1 Vitreous phases region (hatched) P 3 A 3

Sum E

HIM FIELD 0. C, WAS.

Fig. 2

versus T for a Jto rich sample. 233

0.166

Fig. 3 Low field a.c. susceptibilities

(1) S. Barnier, M. Guittard, M.P. Pardo, A.M. Loireau-Lozac'h, J. Flahaut, P. Porcher and J. Livage Mat. Res. Bull., 18, 1217 (1983) (2) S. Barnier, M. Wintenberger, J. Flahaut and J.P. Renard Mat. Res. Bull., 19, 1023 (1984)

Mailing adress of author : Mme M. WINTENBERGER Laboratoire de Chimie Mindrale Structurale 4, Av. de 1'Observatoire 75270 Paris Cedex 06 P 3 A 4

MAGNETIC PROPERTIES OF SOLID SOLUTION PHASES IN THE SYSTEM CrP — CrAs — MnP — MnAs H. Fjellvag, A. Kjekshus and S. Stolen Department of Chemistry, University of Oslo, Blindern, 0315 Oslo 3, Norvay A.F. Andreaen Institute for Energy Technology, 2007 Kjeller, Norvay

Among the MnP type phases, the most interesting structural and magnetic properties are found for the binary phases CrAs, iMnP and MnAs. The binary MnP type phases generally form extended solid solution ranges, and complete solid solution series are established e.g. for the systems CrP — CrAs , CrP — MnP and CrAs — MnAs vhile e.g. MnP —MnAs exhibits a broad two-phase region. The substitutional solid solution markedly alters the magnetic properties in some systems while others are only slightly perturbed. The present contribution is mainly concerned with the quar- ternary solid solution system Cr? — CrAs —MnP — MnAs [Mn, ,Cr,- kz P ; tji < 1.00]. The magnetic properties of the constituents of this system can be summarized as follows:

The binary phases CrP: P [=paramagnetic] U.2 < T < 1000 K. Curie-Weiss law not satisfied.

CrAs : P [T,T .=] 272 < T < [T =] Il8o K. Curie-Weiss law not satis- ——— a, l u fied. H [=helimagnetic along a] U.2 < T < 26l K t=T« ] with

135 at 2 K UstructuraH = 1.7 lU BMnP., TaP/2irc to * MnP,= 0.353H typ, e* x transitio2 " ° n is**' accompanie" d with hysteresis.

MnP: P [T =] 291 < T < 1000 K. Curie-Weiss law satisfied; 9 Ueff-2.8 uB, " 310 K. F ["ferromagnetic] [Ts»] 53 < T < 291 K, Up a 1. U. u at 60 K, moments along b. H U.2 < T < 53 K, u = l.k u_, T /2TTJ* * 0.116, *, o « 20° at 10 K. DC X »& MnAs: NiAs ,P [TD«] 390 < T < 1000 K; Curie-Weiss law satisfied; Ueff = U'5 WB' 9 * 31° K* MnP'P CTc d*] 31T < T < 39° K; Curie-Weiss law not satisfied due to gradual reduction of moment. NiAs.F k .2 < T < 306 K [»T_ . ] , U-, * 3.5 IU at 10 K, moments along a. The MnP,P to NiAs,F type transition is accompanied by hys- P 3 A 4

teresis. A metastable MnP,Ha [helimagnetic along a] state is obtained 'c? application of an external pressure at low temperature. The ternary phases

CrAs1_a.Pr: The Hfl state of CrAs extends to x % 0.07. The MnP.P to MnP,H type transition remains first order for 0.00 £ x <^0.07.

Mn Cr P: The H state of MnP extends only to t % 0.02. A magnetic two phase region exists below T for ^0.02 < t < 0.10.

The F state prevails up to t % 0.50. Tc and y_ varies linearly with the number of electrons [«,, _ - n^. „ _J . 1-t t MnAs P : The S and F states of MnP prevail for ^0.90 < x <_ 1.00. A broad two-phase region occurs for a.0.50 < x < ^0.90. Complex behaviour is found in the As rich concentration range. An MnP,F type state exists for ^0.08 < x < ^0.5. For ^0 .Ok < x < ^0.15 there exist two helimagnetic regimes , both having the spiral propagation vector along a [denoted H' and H for the low and high temperature states, respectively]. The H.' F and H states meet in a triple point. For 0.00 <_ x < -\X1.05 the MnP,K type state can be attained by application of external pressure.

Mnn .Cr.As: The NiAs ,F state of MnAs extends to t % 0.02. An H X — C £ CL state exists for ro0.09

Mn1 .Cr.As, P with fixed values of x • 0.05, 0.12 and O.lk. Magnetic properties of these series trill be discussed. P 3 A 5

HEAT CAPACITY AND MAGNETIC A.C. SUSCEPTIBILITY OF MnBhAs. J. Garcia, C. Rillo, J. Bartolomé, D. Gonzalez Departamcnto Termologìa, Facultad Ciencias, Ü. Zaragoza, 50009 Zaragoza, Spain. R. Navarrò Dcpartamento de Fìsica, E.T.S.I.I., ü. Zaragoza, 50009 Zaragoza, Spain. B. Chenevier, D. Fruchart Laboratoire de Cristallographie CNRS, 166X, 38042, Grenoble Cedex, France. P. Chaudouët E.R. 155, ENSISG," BP 46, 38402 Saint Martin d'Hères, France.

The ternary arsenide MnFhAs is an isotype of Fe2P (1), i.e. hexagonal P62m. It is formed by triangular channels of As atcms; the Rh atans occupy the tetrahedral sites on one channel, while the Mn atans sit at the pyramidal sites on the adjacent ones (fig 1) (2). The study of the thermal variation of the lattice constants evidenced the ccurrence of an abrupt change at T^ =158. K retaining the high temperature synmetry (3). Magnetization (M) measurements shew that the structural transition at Tj is coupled to the change frcm antiferrcmagnetic (AF) to ferromagnetic (F) order (upon heating). Associated to it, the electric resistivity (p) showed an ancmaly in dp/dT. At Tc-200 K the transition to paramagnetism together with a change over-fran metallic to

semi-metallic behavior at T>TC are reported (3). Moreover, for high magnetic

fields an intriguing induced magnetic manent appears above Tc in the magri etization data (3). In order to understand this puzzling feature, heat capacity (Cp) (fig. 2) and a. c. magnetic susceptibility, components X' and X"» (fig. 3) have been measured. The C curve shows an ananaly at Tj_=58 K, a first order sharp peak at Tt=156.9 K (with a thermal hysteresis of -£1=0.8 K), and a broad second order ananaly already starting at 100 K, with a maxiiman at Tp= 240.7 K. Naive interpolation of a smooth curve was applied as a base line to obtain the anomalous entropy, US, and enthalpy, £H, contents at T]_ and Tt. In the case

Fig.l. MnRhAs crystal structure Table I T(K) ÛS/R

• Mn x 58,+l. 1.4+.2 54.+.S O Rh «. 156.9+0.1 0.120+.002 28.7+.3 o As c 190.5rM3.5 * p 240.7+0.3 2.6+.3 320.+30. 238.6+0.2 * •Obtained fron %' experiments. P 3 A 5

of the anomaly at Tp the phonon contribution was substracted using a Debye function with0=300 K. The results are collected in Table I. In the X' curve a small change in slope is detected near Tj_, an abrupt increase

at Tt=157 K takes place in agreement with the AF-*-F transition, and a sharp 1 peak at Tc =190.5 K with a subsequent decrease, is observed. The 1/X curve included in the insert (fig. 3) exhibits a change in slope at T_=238.6 K in

a. O

Fig. 2. Heat capacity of MnRhAs. Curve a represents the estimated phonon contribution (0=300 K). Insert depicts warming (—») and cooling (—•) thermograms.

1 i o A 6 -25. 0.6 £ J X' \\ H MX' o £ 4.5 - 0.015 -12.5 - 0.45 SL r\ z 3. _ Q01 y , - 0.3 t J200 250 300 50 100 i i 1.5 i i - 0.15 i 1 \ 0 i t J 50 100 150 200 250 300 T{K)- Fig. 3. A.c. magnetic suceptibility X' ( ) and absortion X" ( —). Insert 1, expansion of low temperature measurements. Insert 2, inverse susceptibility. P 3 A 5

accord with the Cp anomaly. The phase transition at T is, consequently, identified as due to magnetic ordering. This interpretation is refrended by the AS content, only 7% lower than expected for a total effective moment ofyU=4.65 /^ (3), and explains the appearance of seme ordering in the range TC

ACKNOWLEDGEMENTS The Spanish C.A.I.C.Y.T. is acknowledged for funding this project and the help of E. Sorita in the C experiments is thanked.

(1) S. Rundquist, F. Jellinek, Acta Chem- Scand. 13, 425 (1959). (2) . A. Nylund, A. Roser, S. P. Senateur, D. Fruchart. J. Sol. St. Chem. £, 115 (1972). (3) B. Chenevier, D. Fruchart, M. Baanann, J. P. Senateur, P. Chaudeuet, L. Lundgren, Phys. Stat. Sol. (a) 84, 199 (1984). (4) B. Qienevier, M. Bacmann, D. Fruchart this conference.

J. Bartolome Departanento de Tertnolcgla P 3 A 6

MORIN-TRANSITION IN Ti-SUBSTITUTED HEMATITE: A MDSSBAUER STUDY T. Ericssona, A. Krishnamurthya>c and B. K. Srivastavab a) Institute of Physics, University of Uppsala, Box 530, S-75121 Uppsala, Sweden b) Department of Physics, University of Rajasthan, Jaipur 302004, India c) On leave from b.

Hematite (a-Fe-O,} is trigonal and weakly ferromagnetic (WF) at room temperature with the spins lying in the c-plane (hexagonal setting]. At~260 K there is a

spin-flip (Morin transition,TM) to the [111]-direction and the magnetic ordering is antiferromagnetic (AF) (1). In order to study the influence of Ti on this Fe Ti with 005 transition we made five samples of ( i_x x)2°3 °- - *- 0.025. We used

Fe, Fe203 and TiOo in appropriate proportions as starting materials. The mixtures were sealed under vacuum and heated at 105Q C for 10 hours. The characterization was done using X-ray diffraction ; no other phases except hematite could be detected.

Mdssfaauer transmiss-ion spectra (MS) were recorded between 1Q - 295 K. Centroid shift (CS) in irni/s was measured relative to a-Fe at 295 K and the quadrupole splitting Eg (mm/s) was defined as ((Vg-VgJ-^-Vi))/^ where v-j...Vg are the increasing peak velocities in a magnetically split sextet. Hyperfine fields were measured in tesla. At high temperatures MS could be fitted using only one sextet of Lorentzian lines. However, below 260 K two sextets were needed to get reasonable fits, reflecting the coexistence of the WF and AF phases (Fig 1). In the fitting procedure CS as well as line widths for the two patterns were constrained to be equal. The Mossbauer parameters for x=Q.Q05 are given in Table 1. The hyperfine parameters are the same, within experimental uncertainties, for the other compositions at the same temperature.

In pure hematite the Morin transition at ~260 K is sharp (2). As seen in Fig. 2 and Table 1, our samples show the coexistence of the two phases down to low temperatures. The AF fraction, obtained at low temperatures, decreases strongly with increasing Ti-substitution and is absent for Ti=0.025. This behaviour is Fe A1 x< 08 3 similar to what is earlier found in ( i_x x)2°3 ""*" °- ( «4)> However, Ti depresses the intensity of the AF-phase more strongly than Al does.

The differences in saturated hyperfine fields between the AF and WF phases were earlier found to be~0.8 tesla in hematite (1). In Al-substituted Fe20, de Grave et al (3) found a smaller differences0.4 tesla, which they attributed to spin P 3 A 6

canting. In our samples, the differences are~0.8 tesia and constant, thus indicating neither a canting nor a rotation in the (Fe-j_xTix)203 - system.

Fi gure 1. Mb'ssbauer spectra of (Fe^gg Ti0>Q1 )203 recorded at 295 K (above) and 80 K (below). Solid lines represent the fitted functions.

-10. -5. 0. 5. 10. Velocity (moi/s)

Table 1. Obtained WF-phase AF-phase T(K) E B B Int Mbssbauer parameters i CS o rnt EQ Fe Ti ) 10 0.51 -0.14 53.5 24 0.40 S4.2 76 ' for ( 0.995 0.005 2°3 80 0.50 -0.13 52.7 22 0.40 53.5 78 j at different temperat- 120 . 0.49 -0.13 53.2 24 0.40 53.9 76 ures. For definitions 150 0.47 -0.14 52.4 24 0.39 53.3 76 200 ! 0.45 -0.13 52.1 31 0.39 53.0 69 and units, see text. 220 0.44 -0.13 52.5 36 0.38 53.4 64 240 ' 0.42 -0.14 51.9 47 0.34 £2.6 53 250 ; 0.42 -0.14 51.8 63 0.30 52.5 37 260 I 0.42 -0.14 52.2 100 295 I 0.38 -0.19 51.0 100

The existence of an extended temperature range for the Morin transition as due to variations in composition or particle sizes in a sample has earlier been proposed (4). In Ti-doped single crystals of hematite Besser et ai (5) found, from magnetization and neutron diffraction measurements, that the Morin transition was depressed below 12 K for such a snail Ti-concentration as x=Q.OO2. Accordingly, The AF-phase in Table 1 should not have Ti in excess of 0.0O2, P 3 A 6

which will result in an unrealistic Ti distribution in the sample. Likely, induced strains and vacancies also influence 17. which makes it difficult to compare powder (as we used) and single crystal data. Makarov et al (6) found in a study of a single crystal of a natural hematite that TM was depressed below 90 K. However, when the same sample was crushed,TM showed up at~253 K, as expected for pure hematite.

80 - a) o o ~z 60 -. b) o c) 40 d) RA C LL LL 20

e) 0 - i i i i t t 50 100 150 200 250 300 TEMPERATURE (K)

Figure 2. AF fraction versus temperature for (^Q-i.yJ\)o^ x = a) 0.005, b) 0.010, c) 0.015, d) 0.020, e) 0.025

References: (1) F. van der Woude, Phys. Stat. Sol. U. , 417 (1966] (2) R. C. Nininger and 0. Schroeer, J. Phys. Chem. Sol. 33 , 137 (1968) (3) E. de Grave, L. H. Bowen and G.G. Robbrecht, Stud, in Inorg. Chem. 3, , 571 (1982) (4) E. de Grave, D. Chambaere and L. H. Bowen, J. Magn. Meter. 3Q , 349 (1983) (5) P. J. Besser, A. H. Morrish and C. V. Searle, Phys. Rev. H3 , 632 (1967) (6) E. F. Makarov, I. P. Suzdalev, G. Grol and I. A. Vinogrador, Sov. Phys. JETP 35 , 954 (1973) P 3 A 7

MAGNETIC STRUCTURES OF PeGe

J Bernharda), B Lebechb), 0 Beckman*) and T Freltoftb) a) Dep. of Solid State Physics, Institute of Technology, Dppsala University, Box 534, S-751 21 Uppsala, Sweden b) Physics Dep., Riso National Laboratory, DK-4000 Roskilde, Denmark

Iron monogermanide, FeGe, is known to exist in three polymorphs with monoclinic, hexagonal and cubic structures, respectively (1). Hexagonal FeGe has the B35 type structure (P6/mmm) and is antiferromagnetic below TJJ = 410 K with the spins parallel to the c-axis. Earlier susceptibility, pulsed field (2), magneto-resistance (3) and neutron diffraction (4) measurements indicate that the spins tilt away from the c-axis below ~ 30 K and form an antiferromagnetic cone structure. Several anomalies observed in the earlier pulsed field (2) and magneto-resistance (3) measurements indicate that changes in the spin structure may be induced by an applied magnetic field. We have made a detailed study of hexagonal FeGe with neutron diffraction (5). It shows that already below ~ 57 K the structure changes from a collinear antiferromagnetic structure to a c-axis double cone antiferromagnetic structure. The interlayer turn angle for the basal plane component is 194.4° independent of temperature. This corresponds to a periodicity of ~ 100 A. A re- interpretation of the above quoted data suggests that these findings are consistent with the macroscopic magnetic measurements and not inconsistent with earlier neutron diffraction data. With decreasing temperature the cone half angle increases to ~ 14° at 4.2 K, as can be seen in Fig. 1, and its temperature dependence shows a pronounced kink at 30 K, indicating a phase change at this temperature.

It has been suggested (2), (3) that the magnetic structure changed to a collinear structure at a critical field Bc of 1.4 T at 4.2 K and magnetic field perpendicular to the c-axis. We observe an anomalous decrease of the basal plane component m at Bc (Blc), but the basal plane modulated component does not vanish as would be expected if the structure changed to a collinear structure. With increasing temperature the critical field decreases and the anomaly

becomes less pronounced. Above Bc * 1.4 T (Blc, T * 4.2 K) we find

that ^i decreases with increasing field until B'c * 4.8 T is reach-

ed. At B'c the field dependence of nj_ is changed and above B'c \iL only decreases very slowly at 4.2 K. The cone structure is found to persist up to at least 9.4 T (Blc), which was the upper limit of the applied field. The interlayer turn angle is almost independent of applied field.

— Btckman at al (19721

20 40 60 TEMPERATURE IK) We have also done preliminary magnetisation measurements with field applied parallel and perpendicular to the c-axis. The results are

shown in Fig. 2. The transition at Bc = 1.4 T (Bic) is seen as

an anomaly in the magnetisation data. The anomaly at Bc = 6.9 T

HACHCT1C FIELD X C CT) MAGNETIC FIELD II C

Fig• 2. Preliminary magnetisation measurements on hexagonal FeGe. a) Shows magnetisation at 4.2 K with field perpendicular to c-axis and b) shows magnetisation at 9 K when the field was applied alonq the c-axis. P 3 A 7

(Blc) is the spin flip transition reported from pulsed field measurements (2).

As mentioned above there exist two other polymorphs of FeGe. The existence of modulated magnetic structures with relatively long repeat distances is characteristic of the magnetic ordering of all three polymorphs. Monoclinic FeGe (C2/m) has a complicated magnetic structure (6,7) which order antiferromagnetically at TN ~ 340 K and has a modulated component with repeat distance of

~ 395 A below Tc ~ 115 K. Cubic FeGe (P213, B20) is nearly ferromagnetic below TN (8,9). Recent small angle neutron scattering on a (~ 1.5 mm diameter) single crystal (10) indicates that below TN = 2 78.7 K the magnetic ordering is a long range modulated structure characterized by a modulation of the order Q = 0.010 A"1 which corresponds to a repeat distance of 630 A. The data

suggest that just below TN, 0 is along equivalent <100>- directions and turns toward equivalent <111>-directions at lower temperatures.

References 71 M Richardson Acta Chem. Scand. 2J_, 2305 (1967). 2. 0 Beckman, K Carrander, L Lundgren and M Richardsson Physica Scripta £, 151 ( 1972) 3. B Stenstrom and L J Sundstrom Physica Scripta 6_, 164 (1972). 4. J B Forsyth, C Wilkinson and P Gardner J. Phys. F: Met. Phys. a, 2195 (1978). 5. J Bernhard, B Lebech and 0 Beckman J. Phys. F: Met. Phys. 14, 2379 (1984). 6. D Fruchart, B Malaman, G Le Caer and B Roques Phys. Status Solidi a T%, 555 {1983). 7. G P Felcher, J D Jorgensen and R Wappling J. Phys C: Solid State Phys. Jj6, 6281 (1983). 8. Ii Lundgren, 0 Beckman, V Atha, S P Bhattacherjee and M Richar- son Physica Scripta _1_» 69 (1970). 9. C Wilkinson, F Sinclair and J B Forsyth 5th Int. Conf. on Solid Compounds of Transition Elements. Extended Abstracts p 158 (1976). 10. B Lebech, J Bernhard and T Freltoft. To be published.

Responsible Author: Jonte Bernhard, Institute of Technology, Box 534, S-75121 UPPSALA, SWEDEN Oral or Poster presentation P 3 A 8

MAGNETIC STRUCTURES OF FeSn2 G. Ventur-ini, B. Malaman and B. Roques Laboratoire de Chimie du Solide Minéral, associé au C.N.R.S. n9 158, Universi- té de Nancy I, B.P. n° 239, 54506 Vandoeuvre les Nancy Cedex (France) D. Fruchart Laboratoire de Cristallographie, C.N.R.S., associé à l'Université Scientifique et Médicale de Grenoble, B.P. 166 X, 38042 Grenoble Cedex (France) G. Le Caer Laboratoire de Métallurgie, associé au C.N.R.S. n° 159, Ecole des Mines, 54042 Nancy Cedex (France)

Introduction The stannide FeSn^ has a tetragonal - CuAljtype-structure. FeSn, is antiferromagnetic below T.. = 378 K with u = (1.6 ± 0.1) yn at OK (1). From their neutron diffraction study, Iyengar et al. (2) suggested that the (100) planes are ferromagnetic but the moment directions were not determined. We have therefore undertaken a complete study of FeSn- using magnetic measurements, neutron diffraction and Fe and Sn Mössbauer spectroscopy. Magnetic measurements Thermomagnetic analysis on powders and single crystals confirm that T., = 378 K but also reveal the existence of a second transition at T = (93 ± 1)K, as shown in figures 1 A and 1 B. Neutron diffraction study Neutron diffraction patterns on powder samples have been recorded at the I.L.L. (Grenoble) at various temperatures between 5 and 523 K with X = (2,526 ± 0,002)Â. The unit cell of the antiferromagnetic array is the same size as the chemiGal cell. Between T^ and T the magnetic structure is colinear characterized by ferromagnetic planes (100), antiferromagneticaly coupled along the [100] direction (figure 2) ; the spin direction lies in the (001) plane near the a axis (cf. Mössbauer study). Below T , FeSn, becomes non-colinear antiferromagnetic and the iron moments form a canted structure along the c axis (figure 3). The angle between the two spin directions is 18° and the moment per iron atom observed is

(1.64 ± 0.05)y at 5K. This magnetic behaviour of FeSn2 is similar to that observed for FeGe». A new description of the non-colinear structure of FeGe^ (3) is proposed and the moment per iron ato» is corrected to u = 1.21u_ in agreement with (4-6). P 3 A 8

Mflssbauer study 57, OnLy one Fe sextuplet is observed between 5K and T., which shows no indication of the existence of a transition at T . On the contrary, in the 119 magnetic state, two different Sn types of spectra are observed in agreement with the previous results. Only one sextuplet is observed between T\ and T,,. Below T a new site is observed while the site observed above 90K is strongly broadened. The Sn hyperfine fields are shown on figure 4. The hyperfine field of the new site strongly varies when the temperature decreases. As shown before (7), the Sn field is purely anisotropic above T , demonstrating that the Fe-Sn bond is covalent to a high degree. The spectra allow concluding unambiguously that the spin direction is close to the [100] direction, making an angle of about 15° with it at ^ 10Q K. This angle increases slowly from ^ 0°, at room temperature, when the temperature decreases. This has been confirmed by a Fe single crystal study between 100 K and 300 K. 119 Below T , the Sn spectra suggest that the spins are canted, according to mean magnetic structure determined by neutron diffraction, along [110] 119 direction. However, the details of the Sn spectra cannot be interpreted with this mean structure. Further experiments are planed to obtain a fine interpretation of this magnetic structure. REFERENCE (1) w. Nicholson and E. Friedman, Bull. Am. Phys. Soc, 8, 43 (1963) (2) P.K. lyengar, S.A. Oasannacharya, P.R. Vijayaraghavan and A.P. Roy, J. Phys. Soc. Japan, V7_, B. Ill, 41 (1962) (3) J.B. Forsyth, C.E. Johnson and P.J. Brown, Phil. Mag., 1£, 713 (1964) (4) J. Chenavas, Thesis of 3eme cycle. University of Grenoble - France (1964) (5) E. Kren and P. Szabo, Phys. Letters, U_, 215 (1964) (6) N.S.S. Murthy, R.J. Begum, C.S. Somanathan and H.R.L.N. Murthy, Solid State Com., 3, 113 (1965) (7) G. Le Caer, B. Malaman, G. Venturini and I.B. Kim, Phys. Rev., 26B, (9), 5085 (1982) MCemu/q) fig. 1 A : Temperature dependencies of the magnetization of "monocrystalline" FeSn2 at H = 2T (FONER) 1 T+ 2 T+ a) H t b) H // c

1OO TCK) P 3 A 8

fig. 1 B : Temperature dependencies of the magnetization of "monocrys- taLline" FeSn2 at H = 0,03 T (S.Q.U.I.D.) a) H t b) H // «T

50 150 CKl

z =1/4; 3^4 fig. 2 : Magnetic structure of ("*)z=3/4 FeSn2 at 300 K fig. 5 : Magnetic structure of FeSnj at 5 K

fig. 4 : Temperature dependence of the 119Sn hyperfine fields

TCRT P 3 A 9

PRESSURE DEPENDENCES OF THE CURIE TEMPERATURE FOR CoMnSi Ge, SYSTEM OF SOLID SOLUTIONS S. Niziofr Institute of Physics and Nuclear Techniques, University of Mining and Metallurgy, Mickievicza Av. 30, PL 30-059 Cracow, Poland R. Zach Institute of Physics, Technical University of Cracow, Podchorazych St. 1, PL 30-084 Cracow, Poland J. P. Senateur Ecole Nationale Superieure d'Ingenieurs Electriciens de Grenoble, 38042 Saint Martin d'Heres, France

The aim of chis work has been: - to investigate the influence of the magnetoelastic effect on the changes of the Curie temperature under pressure, - to examine the correlation interaction between crystallographic and magnetic phase transition under applied pressure. Macroscopic magnetic measurements and also neutron diffraction examinations performed for the CoMnSi Ge. system allowed to obtain a magnetic phase diagram (1,2). Depending on the composition the existence of two magnetic regions was ascertained: - the region in which the solid solutions, independently of their composition, show ferromagnetism within the whole range of magnetic order (0 fc x *-0,5); - the region for which the antiferromagnetic, noncollinear magnetic structure exists at low temperatures, and one observes a phase transition to the ferromagnetic phase at temperature T^^ (0.75 £ x £1.0). The changes of the temperatures of the magnetic phase transitions under pressure have been determined by use of the a.c. susceptibility method (3). The results of the investigations of dTj,/dp and dT^ /dp for the compositions x • 1.0 and x « 0.85 were reported elsewhere (3,4). It was observed, that T.^ „ decreases linearly with the pressure

(dTAF_F/dp 4. 0). The composition dependence of l/T_.dT_/dp for CoMnSi Ge. is presented in Fig.l. The dependence is not linear in the whole range of composition. One can observe that the dependence of 1/Tp.dTg/dp is linear both in the ferromagnetic (0 < j *-0.5) as well as in the metamagnetic • P 3 A 9

(0,75 £x £1,0) region. In both cases, l/T^.dT./dp decreases with the silicon content. ttle For CoMhSin a,Gen ic measurements of magnetization vs pressure have been carried out for the metamagnetic phase transition. At pressures p £0.8 GPa no changes of magnetization were found within the experimental error (deVdp — 0). The comparison of (BK)j (S- Bean-Rodbell parameter, K - compressibility) deduced from pressure measurements, with (BK)__ deduced from the exchange striction volume anomaly (5) was made. It was stated that the values of SK obtained in these two different ways are 2 2 1 2 2 1 comparable ((SK)r - -4. 10~ +_ !0~ GPa" , (BK)n =• -7.1O~ +_ 10~ GPa" ). On the other hand, the dependences of dT_/dp and d«"/dp do not satisfy the conditions set for the model of itinerant electron magnetism for this system. It is supposed that a model of narrow electron bands would be more appropriate for the interpretation of the magnetic properties of these compounds. In the light of these facts, the discussion of the CoMnSi Ge, system in papers (1,2) seems correct. The (p,T) phase diagram for CoMnSi_ .Ge- „ is presented in Fig.2. Here, similarily as in the case of CoMnGe, the collaborating magneto-structural phase transition is observed. The triple point TR also exists. For the compositions with higher silicon content the triple point can also be observed, but only at pressures greater than 1.5 GPa. This is connected with the increase of the temperature difference between the temperature of the transition from the hexagonal to orthorombic crystallographic phase and the Curie temperature in the orthorombic phase (6).

Acknowledgements The authors would like to thank Dr.A. Szytula from the Jagiellonian University and Dr.J. Beilly from C.N.R.S. Grenoble for their interest and valuable discussions. P 3 A 9

References (1) S. Niziol, Thesis, Zeszyty Naukowe AGH nr 870, z. 53 (1982) (2) S. Niziol, W. Bazela, Â. Szytula, À. Bombile, D. Fruchart J.Magn.Magn.Mat. 27, 281 (1982) (3) R. Zach, Doctor Thesis, Jagiellonian University (1984) (4) S. Niziol, R. Fruchart, J.P. Sénateur, J. Beilly, D. Fruchart, Proc. of V Int.Conf. on Solids of Transition Elements, Uppsala (1976) (5) H. Pascard, A. Globus, J.Appl.Phys. 53_, 2425 (1982) (6) S. Niziol, A. Zieba, R. Zach, M. Baj, L. Dmowski, J.Magn.Magn.Mat. 38, 205 (1983)

400

0 350 CoMnGe CoMnSi P [GPa]

Fig.l. The composition Fig.2. (p,T) phase diagram dependence of !/Tc.dT-/dp for CoHaSi0iGeQt} for CoMnSi Ge, x 1-x

S. Nizioi, Institute of Physics and Nuclear Techniques, university of Mining and Metallurgy, Al. Miclciewicza 30, 30-059 Cracow, Poland P 3 A lo

MAGNETIC AM5 STRDCTCRAL PHASE DIAGRAM CF Mn. Ni N 4—X X C. Rillo, J. Garcia, J. Bartolane Departamento de Tennologia, Facultad CLencias, U. Zaragoza, 50009 Zaragoza,Spain. J.A. Puertolas, R. Navarro Departamento de Plsica, E.T.S.I.I., U. Zaragoza, 50009 Zaragoza, Spain. D. Fruchart Laboratoire de Cr:stallographie CNRS, 166X, 38042, Grenoble Cedex, France. R. Madar E.R. 155, ENSIBS, HP 46, 38402 Saint Martin d'Heres, France.

Previous studies on Mn, Ni N solid solution phase diagram (1) showed three different phases (fig. 1): A paramagnetic phase (PAR) at high temperatures. A ferrimagnetic phase (F), analogous to the one for Mn.N (x=0) (2), with four magnetic sublattices; one formed by Mn(I), substituted by non magnetic Ni in the solution, which magnetic moments point in the [111] direction, and three formed by Mn(II), strongly bounded to the N atom. Their magnetic manents have components in the (111) plane antiferranagnetically coupled and in the fill] direction oposed to those of the Mn(I). The third region, (III), is the subject of the present contribution. Mn,NiN (x=l) orders below T" in a triangular antiferranagnetic arrangement in the (111) plane with fT symmetry, changing to (T at the lowest temperatures. Moreover, an uncompensated magnetic moment 1*20-002 iL/mol appears in the magnetization measurements (3,4). With the purpose to complete the phase diagram, a.c. magnetic susceptibility of both components,X andX"» at zero static field, has been measured between

400 i i r MnA.xNixN \ PAR 300 \ T(K) 200

100 - III

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Mn4N Mn3NiN Fig. 1. Phase diagram of the Mn, Ni N. the transition at (T.,T'.), HVV andTf are represented'by (S), (A), ($), respectively. Ta) andrdt data taken from reference (1). P 3 A lo

175 200 225 250 275

Fig. 2. Magnetic susceptibility, X'(T), of Mn4 Ni^N for x=0.4, 0.6 and 0.75 at if= 335 Hz. (•) heating and (o) cooling determinations.

77 K and 300 K and for 0.4<_x

For x=0.4 and 0.6 the transition tenperatures TA(F-*III) are clearly displayed and for x<0.75 the phase change PAR-*III (T'A)is also detected, being all tenperatures in agreement with previous data. Thermal hysteresis present at TV points out the first order character of the transition, at which a discontinuous change in the lattice parameter together with the magnetic ordering takes place.

•0 100

Fig.3. Magnetic susceptibility %' (T), of for x=0.825 to 1.0 upon heating. P 3 A lc

Additional anomalous shoulders appear for all concentrations. Sane of them, those

with the maximum at Tf, present frequency dependence (see fig. 4) and a remarkable absorption, while the other ones at T, and T_, (A in fig. 1) (T.

"X'(T), below Tf. For x=0.90 and 0.86, the frequency dependence of the %' maximum

at Tf (fig. 4) verifies the relation l/?gCln(2Tt>)). Furthermore, for x=0.84, magnetization vs. temperature measurements obtained either cooling with or without field, show differences (mainly at lower fields 800 Ce.) (1). Though all this phenomenology has been already observed in the m. Ga. N solution (6), where ferrimagnetic order coexists with a spin glass state, a more complicated behaviouor is present in the Ni solution. In order to elucidate unambiguously, the nature of all these anaralies, s'/stanatic magnetization measurements are actually in progress.

The Spanish C.A.I.C.Y.T. is acknowledged for funding this proyect.

(1) R. Madar, Ph. D. Thesis, Universite de Paris na 686, CNRS, (1970). (2) D. Fruchart, D. Givord, P. Convert, Ph. L'Heritier, J.P. Senateur, J. Magn. Magn. Mater. 17-19, 490, (1980). (3) D. Fruchart, E.F. Bertaut, R. Madar, G. Lorthioir, R. Fruchart, Solid State Camiun. £, 1793, (1971). (4) E.F.~Bertaut, D. Fruchart, Inter. J. Magn. 2, 259 (1972). (5) H. Ebrner, CM. Varnaa, Phys. Rev. Lett. 20, 845 (1968). (6) R. Navarro, J.A. Rojo, J. Garcia, D. Gonzalez, J. Bartolcme, to be published. P 3 A 11

MAGNETIC ORDERING IN THE SPINEL SOLID SOLUTION ZnCr2xi» l2- A. Wiedenmann, J. Rossat-Mignod DRF/DN, Centre d'Etudes Nucleaires, 85 X, 38041 Grenoble Cedex, France M. Hamedoun, J.L. Dormann, M. Nogues Laboratoire de Magnetisme, C.N.R.S., 92195 Meudon, France

s tne + In the solid solution ZnCr2XAl2-2x 4 magnetic Cr"* ions, located in the B-"'tes of a cubic spinel are diluted b> diamagnetic Al3+ ions. A neutron diffraction investigation was performed on a single crystal and powder srmples with different heat treatment for x = 1.0 and on powder samples with x = 0.9, 0.85, 0.80 and 0.5. Two distinct concentration ranges were found : for x >, 0.85 a long range magnetic ordering is built up whereas for 0.30 v< x 6 0.80 a spin glass behaviour is observed |l,2|. The compounds with x > 0.85 undergo a second order phase transition at the Neel-temperatures Tfl(x = i.0) = 15.5 K, TN(x = 0.9) = 14.5 K, T>j(x = 0.85) = 14 K to a helimagnetic order, which is characterized by the wave-vector !q = [0 0 0.79]. The magnetic moments are aligned ferromagnetically in the (001) layer and turn by an angle of 71° from plane to plane. At To = 12 ± 1 K the helical order transforms partly into a mixed phase by a first order transition. In the mixed phase the helical order coexists with two modulated structures defined by the wave vector $2 = [£ £ 0] and 1<3 = [l i 0J. In these collinear structures the moments are aligned along the [I 1 0] and [0 0 l] directions respectively, with a ++-- sequence along the wave vector (fig. 1).

The relative proportions of the three structures in the mixed phase vary from sample to sample, depending strongly upon the preparation conditions. The observed transition at

To and the coexistence of three structures in the mixed phase is surprising and unusual. This can be understood only if the free energy presents three local minima with nearly equal values corresponding to the observed structures. A small variation of the exchange integrals, resulting from the thermal variation of the lattice parameter, would change the relative energy and destabilize the helical ordering below To. The coexistence of the structures finally results from the high degree of frustration in the chromium spinel lattice. In fact, the exchange integrals for the six first nearest neighbours (n.n.), twelve second nearest neighbours (n.n.n.) and third nearest neighbours (t.n.n.) were estimated from a mean field treatment to be Ji = 2 K, J2 = -IK and J3 = + IK respectively. For x 4 0.80 no magnetic reflections appear in the neutron diffraction patterns down to T = 1.5 K. This indicates the absence ot any long range ordering and confirms the presumed spin glass state (SG). At low temperatures no indication for a reentrant behaviour, i.e. transition from the ordered to a SG state, was found. P 3 A 11

Strong spin correlations occur in the whole concentration range below 100 K giving rise to an important diffuse scattering contribution (fig. 2). The analysis of the diffuse intensity as a function of the scattering vector Q revealed that the short range ordering in the SG state is of the same nature as in concentrated samples above T^ : locally the magnetic moments are coupled ferromagnetically within (0 0 1) planes and turn by an angle of ^71° from plane to plane.

— [100] Helicd structure E, = [000.79] mr 1 [001]

o c Modubted structure J

k2 = [V2 V2 0] «£ mr// [110] •a K2

Modubted structure -500 = [1 V2 0] 0 0.5 1.0 1.5 2.0 Scattering vectorCA"1) Fig. 2 - Magnetic diffuse scattering contribution in ZnCrAlS4

[010] 2 0 1/2 1/4 3A 1 • o Fig. 1 - Projections of the magnetic i 3ravais 2 • a structures on the (001) plane in lattices 3 A the mixed phases of 4 P 3 A 11

The correlation function y { increases continuously with decreasing temperature- No anomaly or saturation was observed at the freezing temperature Tf, below which 2 the correlation length still increases. yx scales characteristically with x which agrees fairly well with a high temperature series-expansion above T = 20 K including

first and second nearest neighbour interactions (^ = 2K, J2 = - 2K). The inverse correlation length 1/ I decreases linearly with temperature leading to a finite value at T = 0 in the SG state (fig. 3).

Zn Cr2xAI2.2XS4

o x = 0.50 • xs0.80 A x = 0.85

10 20 30 40 50 60 Tempera ture(K)

Fig. 3 - Inverse correlation length as a function of temperature

• This work is part of a project n° 03-41 E-21 P of the BMFT (West Germany).

(1) M. Hamedoun, A. Wiedenmann, J.L. Dorman, M. Nogues, J. Rossat-Mignod, J. Phys. C (1984), to be published (2) A. Wiedenmann, M. Hamedoun, J. Rossat-Mignod, submitted to J. Phys. C (1984) P 3 A 12

HALF-METALLIC FERKCMAGNETS Magnetization and Electrical Transport Properties of some Heusler-type Alloys M.J. Otto, C. Haas and C.F. van Bruggen Laboratory of Inorganic Chemistry, Materials Science Centre of the University, Nijenborgh 16, NL-9747 AG Groningen

Recent selfconsistent spin-polarized ASW (augmented spherical waves) band calculations based on the local density approximation revealed interesting band structures of some of the Mn-based XMnZ Heusler alloys [i]. These materials were shown to have entirely different band structures for spin-up and spin- down electrons. The conduction and valence bands for spin-up electrons form a continuum of states like in metallic phases, while for the spin-down electrons a gap between the valence and conduction band appears like in semiconductors. Therefore these materials were called "half-metallic ferromagnets". In NiMnSb, for example, the Ni 3d band for both spin directions is full (10 electrons), the Mn 3d band is full for the spin-up direction (5 electrons), but empty for spin-down. The Sb 5p spin-down band is completely filled (3 electrons), leaving two electrons in the 5p spin-up band (Fig.1). As a result of this peculiar band structure the current carriers in the antimony 5p valence band are fully spin polarized at T=0K. As a consequence one expects interesting physical properties. The XMnZ Heusler alloys NiMnSb, PtMnSb, PtMnSn, PdMnSb crystallize in the C1b structure (space group F43m). This structure consists of three interpenetrating face-centered cubic lattices X (0,0,0 +f.c.c); Y ({ ,{ ,i + £ .c.c.); Z (2,2,2, +f.c.c.) (Fig.2). The structure lacks inversion symmetry on the Mn

MndJ

Nidi

N*(£) — •-X o-Mn *-Z Fig.1 Simplified model of the band Fig.2 Crystallographic structure of structure of NiMnSb. XMnZ Heusler alloys. P 3 A 12

sites. It is this broken symmetry which causes strong hybridization between Mn 3d and Sb 5p spin-up electrons, and this pushes the spin—up antimony 5p band up to energies above the Fermi level. Reason for interest in these materials is that the largest magneto-optical Kerr effect (2.5°) measured in any material at room temperature until now, was reported for PtMnSb [2]. PtMnSb, NiMnSb and PtMnSn are ferromagnetic alloys with localized magnetic moments. This localization of magnetic moments at the Mn sites in X-MnZ alloys was recently [3] attributed to the exclusion of the spin-down d-electrons from the Mn sites, leading to localized regions of magnetization in a system with itinerant electrons. This model can also be used for XMnZ Heusler alloys. The magnitude of the magnetic moments in the half-metallic alloys can be predicted theoretically (4U,, per formula unit) just by counting electrons in the spin-up and spin-down bands. This prediction is confirmed by the experimental values, which agree within experimental accuracy (Table 1). PdMnSb and PtMnSn are not half-metallic ferromagnets, according to the band structure calculations. PtMnSn has a lower saturation moment and smeared out behaviour of the specific heat around the Curie temperature. This effect is due to large atomic disorder, which causes also the large residual resistivity of this compound. Table 1. Compound cell parameter \i Curie temperature T

PtMnSb 6.197 1 3.96 pB 572 K from specific heat NiMnSb 5.927 A 3.98 V^ 740 K from resistivity

PtMnSn 6.261 A ~3.6 UB 330 K from specific heat Resistivity measurements on PtMnSb, NiMnSb and PtMnSn (Fig.3) showed a behaviour characteristic for spin disorder scattering of the charge carriers in a metallic ferromagnet. The charge carriers experience an exchange interaction with the localized spins, which become disordered with increasing temperature. The expected strong increase of the resistivity with increasing temperature below T is clearly observed in PtMnSb and NiMnSb. Above T the spin disorder scattering saturates and the resistivity increase linearly with temperature due to electron-phonon scattering (Bloch-Gruneisen law). The large residual resistivity of PtMnSn (about 60Z of the resistivity at T ) can be ascribed to scattering due to substitutional Mn-Sn disorder. The Hall effect of PtMnSb and PtMnSn exhibits an ordinary and a spontaneous contribution, as expected for ferromagnets [4], The total Hall resistivity p_ i3 given by p_ • p - E /J - R B_ + EuM, where R is the ordinary Hall coefficient (RQ - 1/(ne)), and Rg is the spontaneous Hall coefficient. B is P 3 A 12

300 600 900 T(K] .2 .4 1.2 Fig. 3 Specific resistivity Fig. 4 Spontaneous Hall coefficient. the magnetic induction, M^ is the z component of the magnetization, E is the Hall electric field, and J is the current.

The spontaneous Hall coefficient Rg is caused by scattering of the conduction electrons in the presence of spin-orbit interaction [4]. Due to the broken left-right symmetry of the spin-orbit interaction the electrons are asymmetric- ally scattered. According to the theory (for a review see ref. 4) two effects contribute to the anomalous Hall effect, i.e. the asymmetric scattering and the side jump which the electrons experience at the scattering event. That the

spontaneous Hall coefficient Rg is due to scattering of the charge carriers can

be concluded directly by comparing Rg (Fig. 4) and resistivity (Fig. 3) data for PtMnSb and PtMnSn (note the very large impurity resistivity and the large

value of Rg at low temperatures for PtMnSn). The ordinary and spontaneous Hall coefficients in PtMnSb are both positive. The spontaneous Hall coefficient of PtMnSb shows a remarkable behaviour

near Tc, but one should realize that the experimental errors in R are particularly large especially in that temperature region. We gratefully acknowledge that samples used in this investigation were kindly supplied by Dr. K.H.J. Buschow, Philips Research Laboratories, Eind- hoven Nl, and Dr. Z. Fisk, Los Alamos Natl. Lab., Los Alamos U.S.A.

References [1] R.A. de Groot, F.M. Mueller, P.G. van Engen, K.H.J. Buschow, Phys. Rev Letters 50, 2024 (1983). [2] P.G. van Kngen, K.H.J. Buschow, R. Jongebreur, R. Erman, Appl. Phys. Let- ters 42, 202 (1983). [3] J. Rubier, A.R. Williams, C.B. Scanners, Phys. Rev. B28, 1745 (1983). [4] C.L. Chien, C.R Westgate, The Hall effect and its application, 1980 Plenum Press, New York.

M..T. DttO- T.a>inrafr>l-v nf Tnnroanii' Iliiiiaitn WJ ^nknrrrh Ifi WT .07/; 7 Cmi', P 3 A 13

MAGNETIC PROPERTIES OF MIXED HKUK1JSR ALLOYS (T,T')2MnSn (T,T*..Co,Ni,Cu,Rh,Pd)

E. Ohl and H. Boiler

Institut fur PhysUcalische Chemie der Unlversitat Wien, WShringerstraBe 42 A 1090 Wien

Introduction In Heuslar alloy3 T.MnSn the main carrier of ferromagnetism is manganese, occupying the 4b) site in the L2. type of structure. The T atoms, situated in the 8c) position, appear not to contribute a magnetic moment except cobalt. The kind of T atom, however, strongly influences the magnetic interactions as reflected by the Curie temperatures 9 or 8 . In order to get more insight into the role of these elements, we studied a number of mixed crystal series (Oix-1) (1-5).

Experimental The alloys were prepared from high purity (^ 99.99%) metal powders by induction nelting and subsequent annealing at 1080 K followed by quenching. X-ray examination was done by the Debye-Scherrer technique. Magnetic measurements (magnetization as well as susceptibilities in the paramagnetic regions) were performed with a Faraday pendulum magnetometer (SUS10, A.Paar KG; Graz, Austria) in the range 80-1200 K and fields up to 13 koe.

Results The alloys used for magnetic measurements were all single phased. They appeared to be crystallographically well ordered. The lattice parameters vary linearly with x. The course of the Q and the 9 versus composition curves is almost parallel, the 9_'s being slightly higher (Fig.l). The magnetic moment of manganese has an almost constant value of u nj 4a in Heusler alloys without cobalt. In alloys containing cobalt Mn a u depends on x (u,, =• 4.2-3.6 u_). The moment of cobalt has usually a Mn Mn B constant value of u » 0.75 u . In alloys (Co. Cu )-MnSn and (Co Pd ) JtaSn, however, small amounts of Cu or Pd induce a sudden change of the moment (0.3 n_ to 0.75 \i_, and 0.75 \i to 1.00 a respectively) SB a a

Discussion The magnetic interactions in Heusler alloys are often discussed in terms of indirect exchange b«twa«n localized moments mediated by the conduction electrons (6). According to a model by Stearns (7) the P 3 A 13

NljMnS* [Ni,.,C«,)jMnSn

IRh,.1Co1)jMnSn

> Fig 1 : Curie Temperatures of (T,T )2HnSn Heualer Phases P 3 A 13 d, . -d. ,. _ interactions 3hould be dominant. On the other hand, spin itinerant localized polarized band calculations Indicate an itinerant character of all d-electrons, even those giving rise to the apparently localized magnetic moment of manganese (8). In thre« systems: (T,T')» (Ni,Co), (Nl,Pd) and (Co,Pd) the 9'a show a linear dependence on x, having a slope opposite to that of the lattice parameter curve. Thus geometrical and electronic factors act in the same sense. This general feature is also valid for the (Co,Rh)-MnSn system having a nonlinear 9 versus x curve. The minimum in (Cu Co ).MnSn may be inter- C 1— X X & preted as the result of the interplay of Co-Co, Co-Mn interactions on the one hand and the d.-d. interactions on the other hand. In (Ni. Cu ) MnSn the Z- •£. l ™X X it flat minimum may be explained as a superposition of increasing. d.-d» interactions and lattice enlargement. In (Rh Ni ).MnSn the Curie points also increase parallel to the lattice parameters at the rhodium side. It appears thus that Rh is enhancing ferromagnetism compared with Ni or Pd. This observation seems not to be explainable in terms of d.-d» interactions. Alloys containing cobalt have a special position among the Heusler phases. The variability of the magnetic moments of both manganese and cobalt indicate once more the complex character of ferromagnetism in Heusler alloys (9). (1) E. Uhl, J. Magn. Magn. Mater. ?S_, 221 (1981) (2) E. Uhl, J. Solid State Chem.42, 354 (1982) (3) E. Uhl, Monatsh. Chem. JU3_, 275 (1982) (4) E. Uhl, Solid State Comm. J53, 395 (1985) (5) E. Uhl and H. Boiler, J. Magn. Magn. Mater., in press (6) T. Kasuya, Solid State Comm. _15, 1119 (1974) (7) M.B. Stearns, J. Magn. Magn. Mater. J^S, 301 (1980) (8) A.R. Williams, V.L. Maruzzi, CD. Gelatt Jr., J. KObler and K. Schwarz J. Appl. Phys. _53, 2019 (1982) (9) S.M. Dubiel, J.V. Kunzlev, W. Schreiner and O.E. Brandao, Phys. Rev. B21, 1735 (1980)

Acknowledgement This work was supported by the Fonds zur Fordarung der wissenschaftlichen Forschung, Project Ho. P 4820. P 3 A 14

SOFT MAGNETIC AMORPHOUS RIBBONS G.Badurek Atom - Institut, Technical University of Vienna R.Grossinger, H. Sassix, A. Veider Institute for Experimental Physics, Technical University of Vienna Karlsplatz 13, A 1040 Vienna

Many electrotechnical devices (as e.g. transformers, inductivities, shielding devices etc.) are based on the existence of soft magnetic materials, which should exhibit a high value of the saturation magnetization M combined with a low value of the coercivity H in order to obtain high efficiency and to avoid losses. The standard material for this purpose is Fe-Si. Various technological processes exist in order to reduce H (1) . A low coercivity can be expected in materials where the intrinsic anisotropy as well as the magnetostriction is small. Additionally all local defects as e.g. grain boundaries or local stresses increase H . For a polycrystalline material like Fe-Si such local in- homogenities principally cannot be avoided. Therefore amorphous materials like e.g. Fea_B20 were a great hope because in the amorphous state an intrinsic magnetic anisotropy as well as grain boundaries should not occur (1) . A more careful study of these materials showed, however, that all Fe-containing ribbons exhibit a rather large positive magnetostriction (e.g. Fe8QB2 , X '••.3-0.10 ). A large magnetostriction is not desirable for a transformer core because the interplay between X and local stresses (caused by a local varying cooling rate by producing) causes an increase of the coercivity and additionally an accustic hum.

On the other hand for Co30B_0 a smaller but negative value for \ was found (X ^-4.10 ). It is evident that a composition of (Fe,

Co)30B_0 must exist where \ approaches zero. This was really found, however this material contains much of the expensive cobalt, and simultaneously possesses a low saturation magnetization (e.g.

Fe8OB2O : 4 ms = 15'8 kG; Co80B20 : 4 nMs= 11'50 kG? see for comparison the table). From the magnetic data Feo-B__ type materials would be the best,however the magnetostrictLve problem is not yet P 3 A 14

solved in a way that a large scale technical application seems to be possible. This discussion shows that magnitude and sign of the magnetostriction constant are the dominant factors for the hysteresis loop of soft magnetic amorphous ribbons. The ribbons, as obtained with a single roller melt spinning or planar flow cast- ing technique, allow due to their special geometry (2-25 mm broad, thickness 10-50 um, length: * m) stress dependent experiments

which clearly demonstrate the influence of Xs on the M(H) loop.

For a ribbon with positive (negative) xg the application of a stress in the ribbon axis causes an alignment of the moments parallel (perpendicular) to the ribbon axis. This fact is easily visible regarding the change of the M(H) loop which is demonstrated by figure 1 (4). This stress dependence of the hysteresis loop can be used in order to estimate the sign and the magnitude of

Xs (1»5). Domain observations of such ribbons under external stress by u- sing the Kerr-effect show this characteristic behaviour in a direct manner. Studying the

10 8 5 2 4 5 9 10 A change of the polarization tensor 01 electromotive force by polarized neutron transmission 02 C SI B °7S ,5 .O 6 through such ribbons (with and 03 X,= -3.5 lO" without external stress) supports

I 25 N/mm2 the results from optical measure- I ft ments concerning the domain structure and gives additional evidence about the spacial distribution of the magnetic moments in the sanple (5,6) . All these experiments Fe3gNi39(Mo, Si, had the aim to study magnetoelas- A.,r .8 • I0"6 tic properties of amorphous materials in order to develop new al- Fig. 1 Hysteresis loops of amorphous alloy strips with neg. and pos. ma- loys, to examine the influence of gnetostriction under tensile load a. the producing conditions upon the magnetic properties of such soft magnetic materials, when this magnetic behaviour is mainly determined by magnetoelastic effects. p 3 A 14

(1) Int. Conf. on soft magnetic materials SMM 6, Eger, Hungary, Sept, 1983 JMMM 41 (1984) (2) K. Handrich, S. Kobe, Amorphous Ferro- und Farrimagnetika, Physik Verlag 1980 (3) Glassy Metals: Magnetic, chemical and structural properties, CRC Press, Ed. R. Hasegawa, 1983 p 183r R. Boll, H. R. Hil- zinger, H. Warlimont (4) H. Warlimont, R. Boll, JMMM 26 (1982) 97 (5) R. GrQssinger, H. Sassik, Ch. Schotzko, A. Lovas, Int. Conf. on Rapidly Quenched Metals RQ5, Sept. 1984 Wurzburg (6) R. GrSssinger, A. Lovas, G. Wiesinger, G. Badurek, R. Krewenka, S. Hausberger, H. Sassik, H. Kirchmayr, Intermag 1984 Hamburg, IEEE, Transactions on Magn. Vol. MAG-20 No 5,Sept. 1984, 1337

Tab.: Magnetic properties of soft magnetic materials(2,3)

(5) (6) (6) (7) (8) alloy H P HV Bs c UL . PFe (kG) .10 (Oe) .10" .10" (yficiti) (W/kg) 1 7,8 460 0,025 0,9 20 350 60 120 5-10 2 6,7 430 -0,1 0,01 0,85 20 600 130 900 4-6 3 20,3 730 4 0,5 0,71 1 50 50 160 50-100 4 16,1 378 29 0,04 0,78 8 260 140 1000 10 alloy: Composition 1: Permalloy, Mu-metal: ,-Mo. (wt%); ID 4 2: amorphous VAC 6025, 6030: Fe5Co_QSi15B (at%);

3: Oriented T-S: FeQQ aSi, - (wt%);

4: amorphous metal: FegQB20 (at%); Remarksi B): measured with material thickness 50 urn (6): measured at 50 Hz (7): Mi crohardne s s ; (8): Iron losses, measured at 20 Hz with a field of 0,2 T

This work was supported by the Fonds zur FQrderung der wissenschaftlichen Forschung in Osterreich with the projects numbers 5020 and S 4208. P 3 A 15 RE-Fe-B A NEW FAMILY OF MATERIALS FOR PERMANENT MAGNETS R. Eibler, R. Grgssinaer. G. Hilscher. H. Kirchmayr, H. Sassik. G. Wiesinger Institut fur Experimentalphysik, Technische Universitat Wien, Karlsplatz 13, A 1040 Wien

A permanent magnet material well suited for technical applications B is first of all demanded to exhibit a high energy product ( H)max» the theoretical maximum being (M /2) (M ... saturation magnetization). The theoretical value to be gained is limited by the coercivity. This is the reason why the energy product of some materials, despite of a high saturation magnetization, is unreasonably small (e.g. pure Fe). A high quality permanent magnet can only be achieved if the coercivity is of the same order as the remanence M . A large coercivity is usually caused by a high intrinsic uniaxial magnetic anisotropy, which, however is due to an uniaxial crystallographic structure (e.g. hexagonal, tetragonal, ...). Up to now high quality permanent magnets were produced using 3d- metal rich RE-3d compounds (RE(3d)5,RE _(3d).7; RE ... light rare earth). Magnets based on SmCOc or Sm-Co..-, with energy products up to 30 MGOe are now available since several years. However the high costs of the raw material are restrictive for a large scale technical application. Very recently a new type of permanent magnets based on Nd-Fe-B has been developed (1,2) by drawing attention on the following points: i) The main element is Fe, causing a large saturation magnetization (37,1 Aig per formula unit) ( 1 , 6 ); ii) B stabilizes the formation of a compound with tetragonal

structure: Nd2Fe14B (P42/mnm); a=8,8oX, c=12,198 (3,5); iii) Nd raises the anisotropy field H. up to a value of about 8T at ambient temperature (fig. 1). Permanent magnets with the nominal composition Nd-eFe^-Bg have already been produced for which energy products exceeding 4OMGOe have been reported (4). The easy axis of magnetization is parallel to the c-axis at room temperature, whereas at about 135K it changes to an easy cone with a maximum angle of approximately 30° deviating from the c-axis obtained at T=4.2K (3,6). Commonly permanent magnets can be produced by applying two different methods; P 3 A 15

60 80 100 120 W 60 80 200 220 2W 260 280 300 Fig. 1: H of the Nd-Y-Fe-B series

too 200 300 T[*C) Fig. 2: Normalized temperature dependence of the anisotropy

field HA and the coercivity of a Neomax-Magnet P 3 A 15

i) The melting-grinding-sintering technique, established in the case of SmCOg (2) ; ii) A rapid quenching method as used for the production of amorphous ribbons (1), followed by grinding and plastic binding procedure•

For basic interest Hft of different master alloys was investigated. Fig. 1 shows that the main hard magnetic properties are based on

the Nd alloy, because Y-Fe-B has only a very small HA and no low temperature anomalies. For a technical application predominantely the magnetic properties at elevated temperatures are of importance. For this reason a tech- nically sintered Nd-Fe-B type magnet, kindly supplied by Sumitomo Special Metals Comp. was examined above room temperature. In fig.2 the temperature dependence of the anisotropy field H, , determined by applying the SPD (Singular Point Detection) technique (7), can be compared with the coercivity jH-, both having been normalized to the values at T = 273 K. Fig. 2 demonstrates that the intrinsic

coercivity as a function of temperature does not scale with Hft(T). Moreover the decrease of jR^, with raising the temperature ex-

ceeds markedly that of HA. The reason for this unique temperature behaviour is the subject of further investigations on this topic.

1) J. J. Croat, J. F. Herbst, R. W. Lee, F. E. Pinkerton, Research Pub.l. of General Motors GMR - 4492 (Okt. 1983) 2) M. Sagawa, S. Fujimura, M. Togawa, H. Yamamoto, Y. Matsuura, J. Appl. Phys. .55 2083 (1984) 3) R. Grossinger, P. Obitsch, X. K. Sun, R. Eibler, H. R. Kirch- mayr, R. Rothwarf, H. Sassik, Mat. Lett. 275 (1984) in print 4) K. S. V. L. Narasimhan, Proc. of Intermag Conf. (1984) Hamburg 5) J. F. Herbst, J. H. Croat, F. E. Pinkerton, W. P. Yelon, Phys. Rev. B 2± (1984) 4176 6) D. Givord, H. S. Li, J. M. Moreau, R. Perrier de la Bathie, E. du Tremolet de Lacheisserie, Proc. of REACIM (1984) St. PQlten, Austria, to be published in Physica 7) G. Asti, S. Rinaldi, J. Appl. Phys. 4£ (1974) 3600

This work was partially supported by the Treibacher Chemische Wer- ke, Austria. P 3 A 16

MAGNETIC STRUCTURES DETERMIHED BT NEUTRON DI7FRACTI0N - DESCRIPTION AND SYMMETRY ANALYSIS

A. Oles, W. Sikora, A. Bombikr M. Konopka Institute of Physics and Nuclear Techniques, Academy of Mining and Metallurgy, Al. Mickiewicza 3Ot JO-O59 Krakow, Poland

The information about a continuation (1) of the monography.'2.) Is- presented* The data compilation method, used is the same as in (2) . It concerns information regarding compounds investigated by neutron diffraction in last years. The new and essential part of information about each presented compound is based on symmetry analysis results. The symmetry analysis is the method developed by Izyumov and coworkers (3) • It uses representation description of magnetic structures proposed by Bsrtaut (4 ) . This method gives a possibility to construct all magnetic structure types allowed by the paramagnetic initial phase symmetry. Any type of magnetic structure in the frame of this method may be described by several coefficients. The symmetry analysis concerns the static situation and may be used for crystals with localized magnetic moments. It can be succesfully applied to the compounds with magnetic sublattices fully occupied by one sort of magnetic atoms. As the results of symmetry analysis there are given: - the wave vector star ; kj , and this from star arms k , which results from the translational symmetry change at phase transition ; - the representation f, of the initial phase symmetry space group to which the established magnetic structure belongs ; - the representation basic vectors vyl of the wave vector star - the order parameter p constructed for a given magnetic phase from coefficients c* by means of a linear combination of basic vectors: ^ . *>• v * "IP *• where $ is the vector describing r>, x,i *• x a considered magnetic structure. The results of the symmetry analysis give for some compounds the magnetic structures different than suggested by the authors of P 3 A 16

experimental papers* There are complicated structures for which the interpretation of neutron diffraction patterns is difficult. In that case, the symmetry analysis may be very useful. The description of topical neutron diffraction possibilities and of the algoritm, by which the symmetry analysis may be used in a simple way, are given in (1 ) «

M) A. Olss, W. Sikora, A. Bombik, M. Konopka, Scientific Papers of Academy of Mining and Metallurgy, Nr 1005, Phys. 1j (1984J (2; A. Oles, F. Kajzar, M. Kucab, ¥. Sikora, "Magnetic Structures Determined by Neutron Diffraction", PWN, Warszawa, 1976. (3) Tu.A. Izyumov, 0.7. Gurin, J.Magn.Magn.Mat. 36, 226, (1983) (A-) E.F. Bertaut, Acta Cryst. A24, 217, (1968) . P 3 A 17

STABILIZATION AND CHARACTERIZATION OF UNUSUAL OXIDATION STATES OR UNUSUAL ELECTRONIC CONFIGURATES OF TRANSITION ELEMENTS

G. Demazeau, M. Pouchard, B. Buffat and P. Hagenmuller

Laboratoire de Chimie du Solide du CNRS, 351, cours de la Liberation, 33405 Talence Cedex, France.

The stabilization of an oxidation state depends closely on two factors : - the stabilization of the corresponding electronic configuration, - the oxygen pressure used.

We set up a simple model emphasizing the interdependence between the electronic configuration of a transition element and the structural and chemical bonding factors characterizing its first environment.

In a first step the selected structural environment was a tetragonally distorted octahedral site.

Using in addition high oxygen pressures up to 70 kbar, unusual electronic configurations like : high spin Fe (t- d 2),medium spin Co (d d d d 2d 2 2), 4+221 g 3+ 6 1 yz zx xy z x -y low spin Co (add), low spin Ni (t_ d 2)| low spin Cu^+ g 2 p yz zx x^ § z (t d 2d 2 2) have been stabilized in a A'A Li _M _0 matrix deriving from the K.NiF. structure.

In return due to its isotropic electronic configuration, Fe (t_ e ) has been 2g g stabilized for the first time in the perovgfcite La.LiFeO.. Such unusual oxidation states or electronic configurationshave been characterized by magnetic measurements, ESR and Mossbauer resonance. P 3 A 18

CRYSTAL CHailSTRY AND MAGNETISM IN SILXCXDES LaFe. Rh Si (ThCr-Si.-TTfEE) P. Rogl and K. Hiebl Institut fur Physikalische Chemie der Oniversitat, W&hringerstraBe 42 A 1090 Wien und G. Wiesinger Institut fur Experlmentalphysik Technische Oniversitat, Karlsplatz 13 A 1040 Wien

Despite several research groups have focussed on the structural chemistry and the physical properties of the REFe.Si - as well as LaHh.Si -compounds, there is still a considerable number of confusing results. Bodak and Gladyshevskij (1972) were first to report the formation of LaFe Si crystallizing with the ThCr_Si -type of structure (X-ray powder data). Later Felner and coworkers (1975) confirmed the existence of LaFe Si with the ThCr.Si -type; they further reported magnetic susceptibility data of U . = 0.29 p. at 4.2 K and a weak ferromagnetic ordering at T = 668(5) K. eff B m From magnetization and Mossbauer effect studies most of the iron (94%) was concluded to be diamagnetic. More recent room temperature data however revealed only one type of Fe atoms in the structure without a magnetic moment (Umarij and coworkers, 1983). Agreement exists about the crystal symmetry in LaEhjSi with the ThCr.Si -type as well as about the superconducting transition at T =» 3.9 K (Ballestracci, 1976; Chevalier et al., 1982; Felner and Nowik, 1983). Measurements of magnetization, ac-susceptibility, electric resistivity and specific heat seemed to indicate a magnetic ordering at T = 7 K interpreted as iterant electron long range order of the Hh-4d electrons (?) (Felner and NowiJc, 1983). X-ray analysis of about 10 pseudobinary samples LaFe. Rh Si. in both the as-cast and the sintered condition revealed congruent: melting behavior as well as complete solid solubility throughout: the entire pseudobinary system for T >600°C. All powder patterns were indexed completely on the basis of a body centred tetragonal unit cell and the observed intensities and the extinctions in all cases proved structural analogy with the ThCr.Si -type (space group 14/rnmn). Using the atom parameters previously derived for CeOs.Si^ (Bwrvath, 1983),- the observed and calculated powder intensities are in excellent agreement for a statistical distribution of iron and rhodium on the 4d-sites. The dependencies of the lattice parameters and volumes of the pseudobinary section LaFe. RhxSi2 versus the concentration z are represented in Fig.l. P 3 A 18

LaFe2Si2- LaRh2Si2

• Ballestracci, 1976 a Umarji et ai.,1983 o this work a Feiner& Mayer.1975 7 Fetner8.Nowik.1983

20 40 so 80 LaRh2Si2 MOLE%LaRh2S52 "•

Fig. 1 : Lattice Parameters and volumes of the ternary alloys LaFe Rh Si_ with ThCr.Si.-type of structure

The magnetic behavior of LaT.Si. (T- Rh,Fe) is characterized by a temperature independent paramagnetism due to the metallic state. The Fe atom proves to be nonmagnetic (see also Omar ji et al., 1983), however at low temperatures (T< 100 K) small Curie tails appear, but we do not observe any si 3tai Dits kind of oagnatic ordering. La(Rh_Fe J2 2 * - strong field dependent susceptibilities below 650 X, due to a very weak magnetic moment VL •* 0*024 U_ at 2*5 K indicating that the spontaneous magnetization observed is not a bulk phase property. P 3 A 18

MSssbauer spactxa (4.2 K and roan temperature) unambiguously reveal a «m«Ti amount of iron on a second site without: a magnetic moment. In contradiction to earlier'data, LaRh-Si- does not undergo a superconducting transition above 1.5 K.

Acknowledgements This Investigations was supported by the Austrian Science Foundation (Fonds zur FSrderung der wissenschaftlichen Forschung in Ssterreich)through grant No. 5297. P.R. expresses his gratitude to the Hochschuljubilaumsstiftung der Stadt Wien for the KD-530 type microdensitometer and the MNT-306 micro- balanca. Thanks are also due to the Autrian Science Foundation for the use of the SUS-10 under grant No.4820.

References

(1) O.I. Bodak and E.E. Gladyshevskij, Vestnik. Lvov. Derzh. Univ. Sar. Khim. _1£, 29 (1972) (2) I. Falnar, I. Mayer, A. Grill and M. Schieber, Solid State Conunun. 16, 1005 (1973) (3) A.M. Umarji, D.R. Noakes, P.J. Viccaro, G.K. Shenoy, A.T. Aldred and D. Niarchos, J. Magn. Magn. Mater. _36_, 61 (1983) (4) R. Ballestracci, C.R. Acad. Sci. Paris, B282, 291 (1976) (5) B. Chevalier, P. Lejay, M. Vlasse, J. Etourneau and P. Hagenmuller Mater. Res. Bull. _17, 1211 and _18, 315 (1983) (6) I. Felner and I. Nowik, Solid State Commun. 47(10), 831 (1983) (7) C. Horvath and P. Rogl, Mater. Res. Bull. _18, 443 (1983) P 3 B 1

TRANSMISSION ELECTRON MICBOSCOPIC STUDY OF V.Si SINGLE CRYSTALS DEFORMED BY COMPRESSION AT 165O°C A Ben Lamine, F. Reynaud Laboratoire d'Optique Electronique du C.N.R.S., BP 4347, 31055 TOULOUSE Cedes, FRANCE and J.P. Senateur I.N.P.G., Genie Physique, ER 155, BP 46, 38402 ST MARTIN D'HERES, FRANCE

The superconducting compound V,Si (A15 structure) belongs to the large family of intennetallic compounds which are very brittle at room temperature and which can be deformed plastically at high temperature (1,2). After some early works (3,4)systematic studies of the macroscopic deformation of V Si at high temperature have been undertaken (5,6,7) during the last few years. However, to the best of our knowledge, there exists up to now only one published study using transmission electron microscopy, which shows creep subbondaries in V,Si deformed between 1280 and 1400°C (8). We present here a detailled study of the nature of the creep subboundaries and of the extrinsic dislocations inside the subgrains in samples of V,Si deformed by compression at 165O°C. This work forms part of a more general study of lattice defects in V-Si by transmission electron microscopy (9,10,11,12,13, 14). V.Si single, crystals have been compressed by two tungsten pistons in a -* 7 quartz vessel under vacuum. The stress was of the order of 1.4 x 10 Pa for a Q cylindrical sample of 6 mm diameter and of 10 Pa for a parallelepipedal sample of 2 mm x 2 mm section. The electropolishing of the samples for electron microscopy has been explained elsewhere (10). The thin films were observed either in a Philips EM 400 (120 kV) or in a high voltage electron microscope (1 MV) of the Laboratory. The determination of the Burgers vectors of the dislocations has been performed by computer simulation (15) with the two-beam dynamical theory of electron diffraction. Figure 1 shows an example of a creep subboundary observed in a V.Si sample deformed by 2,7 Z.Subboundaries ccmposed of one and two families of dislocations have been observed, in accordance with the results of (8). The Burgers vectors of the dislocations constituting the subboundaries are a <100>. The misorienta- tion, measured with the Kikuchi lines, is smaller than 1 degree. The determination of the plane of the subboundaries shows that both glide and climb have been involved in the formation of the subboundaries. Extrinsic dislocations have been also observed either inside the subgrains. P 3 B 1

or interacting with the dislocations of the subboundaries : their Burgers vectors have been shown to be of type a <100> , a <110> and even perhaps a <111> . From the results of this study, the .mechanical properties of V_Si may be explained as following : - at low temperature, the slip systems <100> jlOOJ observed experimentally provide only three independent slips systems, which is not enough , according to Von Mises criterion, to allow the samples to deform plastically, - at high temperature, the activation of the new slip systems <11O> j100j provides, together with the classical slip systems { 100}, enough independent slip systems to allow V.Si to deform plastically.

(1) G.E.R. Schulze und P. Paufler, Abh. Sach. Akad. Wissen. Leipzig, 51, 5 (1972). (2) P. Paufler und G.E.R. Schulze, Neure Entwicklung der Physik, Ed. P. Gorlich, A. Eckardt und P. Kunze, VEB Deutscher Verlag der Wissenschaften, Berlin (1974), 14. (3) E.S. Greiner and E. Buehler, Bull. Amer. Phys. Soc. ]_, 310 (1962). (4) H.J. Levinstein, E.S. Greiner and H. Jr. Mason, J. Appl. Phys. 37, 164 (1966). (5) A. San Martin, D.M. Nghiep, P. Paufler, K. Kleinstiick, U. Kramer and Quyen N.H., Scripta Met. U_» 1041 (1980). (6) M. Bertram, P. Paufler and K. Kleinsriick, Cryst. Res. and Techn. j_6, 89 (1981). (7) S. Mahajan, J.H. Wernick, G.Y. Chin, S. Nakahara and T.H. Geballe, Appl. Phys. Lett. 33, 972 (1978). (8) D.M. Nghiep, P. Paufler, U. Kramer, K. Kleinstiick and N.H. Quyen, J. Mater. Sci. _T5, 1140 (1980). (9) A. Ben Lamine, F. Reynaud, C. Mai and J.P. Senateur, Phil. Mag. 38A, 359 (1978). (10) A- Ben Lamine, J.P Senateur and F. Reynaud, J. Microsc. Spectrosc. Electr. 5_, 745 (1980). (11) A. Ben Lamine, M.J. Lahana, F. Reynaud and P. Stadelmann, J. Mater. Sci. Lett. _3. 431 (1984). (12) A. Ben Lamine, F. Reynaud, C. Colleix, M. AchSche, J. Sevely and K. Kihn, ICXOM 83, J. Phys. C2, 45, 709 (1984). U2) A. Ben Lamine, J.P. Senateur et F. Reynaud, submitted to J. Microsc. Spectrosc. Electr. . (14) A. Ben Lamine, R. Portier, J.P. SSnateur et F. Reynaud, submitted to Scripta Met.. (15) A.K. Head, P. Humble, L.M. Clarebrough, A.J. Morton and C.T. Forwood, Computed Electron Micrographs and Defect Identification, North-Holland Publ. Cy (1973). ? 3 B 1

Figure 1 : Electron micrograph showing a creep sub-boundary in a V-Si single crystal deformed by compression at 165O°C.

Franqois REYNAUD Laboratoire d'Optique Electronique BP 4347 P 3 B 2

TERNARY TRANSITION METAL SILICIDES (OR GERMANISES) BUILT UP OF INFINITE COLUMNS OF Si<<3«> - CENTERED SQUARE ANTIPRISMS AND TRANSITION METAL - CENTERED OCTAHEDRA. B. Chabot, E. Parthe and K. Cenzual Laboratoire de Cristallographie aux Rayon* X, Universite de Geneve, 24, quai Ernest Ansermet, CH-1211 Geneve 4 (Switzerland)

In two recent systematic studies of the structures of ternary rare earth - transition metal - silicides and homologues <1,2) it has been posssible to correlate geometrically a great number of different structure types. One of the aims of such an undertaking is to understand and interpret the sometimes complicated chemical formulas in relation with the structural features. In this work we will draw the attention to particular ternary structures of general composition R,T M where R s very late rare earth element, Sc or a transition element of the fourth group (occasionally also Mb or Ta); T = smaller transition element of Cr,Mn,Fe or Co group (occasionally also V) and M = Si or Ge. The structures of interest here are characterized by two kinds of infinite columns, both parallel to the shortest cell axis (of about 5 to €A>. One is formed of face-shared square aatiprisms centered by M atoms. The second kind of infinite column consists of face-shared octahedra formed by M atoms and centered by T ' atoms. It was of interest to find a classification scheme for structures presenting these two kinds of columns in order to predict compositions of new structures with similar features. The structures are conveniently classified according to : a>the composition of the antiprism columns b)the linkage of the octahedron columns Eight structure types, found with more than 40 different R x Ty M z compounds, are known with the structural features discussed above. Drawings of these structure types, in a projection along the column axes, are shown in the Table on the next page. Table caption : Eight structure types with infinite antiprisms columns and infinite octahedron columns, classified according to the composition of the M centered antiprism columns and the linkage of the T centered octahedron columns. Large circles : R atoms, medium circles : T atoms, small circles : M atoms. Fully drawn circles : height 1/2, dashed circles : height 0, double circles : height 1/4 and 3?4. Compositions of infinite M centered antiprism columns

R6T2M2

Edge conned ions with three other columns

§ Edge — connections 3i with two OWr columns

2fMnS<2 0t4S. Immn

No edge connectaons with other

•X) tnC73.CZ/t

CO ho P 3 B 2

In the three structures 2rF«Si2f ZrMnSi- and LuMnGe- there are extra M atoms inbetween th« antiprism column*, which assure that also the T atoms, which participate on the formation of the antiprisms, in octahedrally surrounded by M atoms.

The structures of Sa^Rm^Si^ and Hf2Ru3Si4 have recently been determined in our laboratory. The former had earlier been

mentioned in the literature with the composition Sc3Re_Si_ <3) but the correct stoichiometry was deduced from the classification scheme presented here and corroborated from single crystal X-ray diffraction data. The structure is centrosymmetric and the space group is C2/c instead of B2 as given in <3). It can be seen from the projections of the structures in the Table that Nb-Cr-Si- and Nb.Cr.Si- have the same atom arrangement but the composition of the antiprism columns is different. It is not known whether intermediate compositions are also possible. In Sc^Re.Si. and Hf.Ru^Si. the atoms are also arranged in a similar way but no system it known yet where both structure types exist.

A detailed account of this work will be published in Acta Cryst.B.

(1) E.I. Gladyshevskii & O.I. Bodak, "Crystal Chemistry of Intermetallic Compounds of the Rare Earth Metals" (in Russian). Lvov (USSR): Vishcha Shkola (1382) (2) E. Parthe & B. Chabot, in "Handbook on the Physics and Chemistry of Rare Earths" edited by K.A. Gschneidner Jr & L. Eyring, Vol.6, chapter 48, 113-334 (1384) (3) V.K. Pecharskii, O.I. Bodak & E.I. Gladyshevskii, Sov. Phys. Crystallogr. 24(4), 433-438 (1979)

Prof. E. Parthe Lab. Cristallographie, University de Geneve 24, quai E. Ansermet P 3 B 3

STRUCTURAL AND MAGNETIC PROPERTIES UPON HYDRIOATION OF SOME BORIDES OF REgFe^B TYPE. D. Fruchart, P. Wolfers Laboratoire de Cristallographie du C.N.R.S~, associe a l'U.S.M.G., 166 X, 38042 Grenoble Cedex, France.

New intermetallic phases stabilized by boron have recently given rise to a- considerable interest and are now classified among the highest performant permanent magnetic materials. Compared with the cobalt-rich compounds SmCo5, a significant lowering of the cost will be realized. As for most of the binary or ternary metal compounds containing both RE-metals and transition elements M, hydridation properties might be expected. The M/RE ratio and the presence of p-elements in the alloy are real factors governing the final H/metal ratio. Experimental evidence of a systematic formation of hydrides has recently been established for a large series of compounds based on iron and cobalt (1). In most of cases, the number of H atc~s absorbed is found 3-5 per unit formula K 0.4 % in weight). As for magnetic compounds alloying RE metals and d elements (ex. Laves phases, l^fREJg, ...) the magnetic properties of the alloys are significantly affected on hydriding. Here, an increase of both the saturation magnetization and the Curie point has been observed, between the virgin compound and its hydride. However a systematic reduction in anisotropy is observed for the H-filled compound. Neither the chemical, nor the magnetic properties could be well understood without the knowledge of the complete structure of the compounds. Recent and quasi simultaneous work have been carried out and more details have just been reported concerning these new materials (2, 3, 4). This structure, clearly related to the sigma phase also exhibits some local coordination encountered in the CaCUg-type. This knowledge allows us to apply Westlake's geometrical model theory (5) particularly successful in the case of tetrahedrally close packed alloys. A maximum of 5.5 H atoms per unit formula has been calculated which compares with the experimental value of 5.2 observed for

Nd2Fe143. Neutron diffraction experiments at various temperatures performed on hydrides and deuterides of Y, Nd, Ce compounds have allowed us to localize the H (and D) atoms. Simultaneously, the magnetic structures are analysed before and after hydridation. The results are discussed in terms of M-M distances, f-d polarization, p-d bonding and H screening.

(1) P. L'heritier et al., C.R.A.Sc. Paris, to appear (1984) P 3 B 3

(2) J.F. Herbst, J.J. Croat, F.E. Pinkerton, W.B. Yelon, Phys. Rev. B29, 4176 (1984) (3) D. Givord, H.S. Li. J.M. Moreau, Sol. State Conni. 50, 497 (1984) (4) C.8. Shoemaker, O.P. Shoemaker, R. Fruchart, Acta ïïryst., to appear. (5) O.G. Westlake, J. Less Com. Metals 90, 1 and ^1, 251 (1983)

D. Fruchart Laboratoire de Cristallographie, C.N.R.S., associé à TU.S.M.G., 166 X, 38042 Grenoble Cedex (France) P 3 B 4

THE USE OF THE "INHOMOGENOUS LINEAR STRUCTURE SERIES" ON THE STRUCTURAL DESCRIPTION OF SOME TRANSITION METALS COMPOUNDS Yu.Grin Institute of Physical Chemistry, University of Vienna, W3hringerstrafle 42, A 1090 Vienna, Austria

Structures, which are based on the various stacking of two dimensional fragments of more simple original structures differentiating on concentration and/or coordination characteristics, are defined as inhomogeneous linear structures (I). A possibility of combination exists only, if different fragments show the same atom distribution in the contact planes. Each series of structures can be described with the formulas of the original structures and/or with the unit cell formula. The composition of any number of a series is confirmed by certain coefficients of the unit cell formula. The symmetry of a possible structure of a series can be derived using modified Zhdanov-symbols, which differ from the regular Zhdanov-symbols by additional indices at every number. The lower index characterizes the structure from which the fragments originate. The upper one denotes the symmetry element included by the fragment. The symmetry of the symbols comply clearly with the symmetry of the structures. The number of the various symmetries of symbols corresponds with the number of possible space groups for the structures of the series under consideration. In the ternary system Ce-Ga-Ni three compounds have been found, where the structure could be explained and interpreted by the method described above. These alloys belong to the BaAl.-PtHg. series with the unit cell formula R X,' „ X" ^ i. m 4m+zn n (R= Ce, X'» Ga, X"= Ni):

CeGa,Ni._ m» 2 n" 2 space group P4/mmm Zhdanov-symbol (2)2)2

Ce2Ga.jgNi m= 4 n» 2 space group I4/nmim Zhdanov-symbol

Ce^Ga17Ni2 m=» 8 n- 2 space group I4/mmm Zhdanov-symbol

Ga in case of Ce4Ga]7Ni2 X'= |7/[g^ijy18« The remaining possible structures of this series reveal the symmetry of the space groups I4/mmm, I4mm, P4/nmm, P4/mmm and P4mm (2).

The structures of the Y2 +2 Ga2 Co2 +2 compounds belong to the series OC-ITl-UPt2:

YCo m « 2 n * 0 Cmcm (1.)2 m • 2 n » 8 Cmcm 0*4.).

Y4GaCo4 m - 1 n - 3 C2/m 1*3, m » 2 n « 4 Cmcm (1 *2,)

m ^ l ^x ^ I XXDDSQ i #% i«* P 3 B 4

The possible symmetry of hypothetical structures of this series corresponds with the space groups Cmcm, Cmc2j, C2/m, Amm2 and Cm (3). The so called chimney-ladder structures (4,5) represent a complicated example of inhomogeneous linear structure series. The fragments of three different compositions are necessary for their description. The formula is

L +_ +, M_ +2 +A » w^re for L- and M-components similar radii and coordination characteristics are required:

Z Ru2Sn3 p - 4 q-2 r =• 0 P4c2 C^2 \h 1 4 Ir3Ga$ p = 8 q = 2 r = 0 P4n2 ( 2 1)2 1 2 ] Ir4Ge5 p - 4 q=6 r » 0 P4c2 C^ 1 2 1^2 6 Rh]0Ga]7 p=28 q=6 r=0 P4c2 <>2V2 1 4V2 5 A ! 5 5 MnnSi)9 p = 28 q=2 r= 4 P4n2 <»2 I * 3 1 3 1 2 6 ] 6 J 6 5 Mo,3Ge23 p = 36 q=2 r= 4 P4n2 <>2 ] 3 1 3 I 2

Mn15Si26 p = 40 q=4 r=4 I42d (l^,^,^

RhJ7Ge22 p=!6 q = 20 r=4 I42d ('2 1 , I2 1 , ^>3>2 1 , 4 I ,)4 8 lo 8 V]7Ge3] p=52 q=2^ r=4 P4n2 ('2 ,'3 i'3 i>2 Mn2?Si47 p=68 q=2 r=12 P4n2 (126,124,135,134,135,134,136,),. The lowest symmetry of the hypothetical structures is space group P2. The other structures can be classified in three groups: monoclinic (possible space groups P2, 32, P2/c, B2/b), orthorhombic (possible space groups P222,, 12,2,2,, C222, F222, Pcc2, Pnn2, Pcca, Pnna, Fddd) and tetragonal (P4, 14, 14,, 14 22, 14 /a, P4c2, P4n2, I42d).

The structure model for Tc^Si7 (6,7) has been created by means of this method: space group Pcca, a- b= 5,737, c= 18,099 X. The calculated intensities of the reflexions are in satisfactory agreement with the experimental data.

(1) Yu.N. Grin, Ya.P. Yarmolyuk, E.I. Gladyshewskij, Sov.Phys.Cryst., 2_7, 413 (1982) (2) Yu.N. Grin, Ya.P. Yannolyuk, O.A. Usow, A..H. Kuzmin, W.A. Bruskow, Kristallografiya, 2_8_, 1207 (1983) (3) Yu.N. Grin, R.E. Gladyshewskij, A.N. Sobolev, A.P. Yarmolyuk, Kristallographia, 29, 1008 (1984) (4) H. Nowotny, The Chemistry of Extended Defects in Non-Metallic Solids, p. 223-237. Amsterdam: North-Holland Publ.Co. (1970) (5) H. Boiler, M.Chemie, J05, 934 (1974) (6) A. Wittmann, H. Nowotny, J.Less-Common Metals, £, 303 (1965) (7) J.B. Darby, J.W. Downey, L.J. Norton, J.Less-Common Metals, 8, 15 (1965) P 3 B 5

MBSSBAUER STUDY OF THE TERNARY SYSTEM CFe1_xVx)3 Ge

L. HgggstrBm, J. SjdatrSm, A. Narayanassamy and H.R. Verma Institute of Physics, Uppsala University, Box 530, S-7S1 21 Uppsala, Sweden

In the ternary system (Fe, V ). Ge one finds three types of crystal

fcc D0 and structures, hep (DO19)i < 3) sc (A15) stable for 0 - x £ 0.10, 0.17 - x ^ 0.28 and 0.55 - x - 1 respectively. Due to the large magnetic hyperfine fields at the iron nuclei, in the hep and fee phases, well resolved

Fe v Ge Mossbauer spectra are obtained. This implies that the ( i_x x^3 system offers a good possibility to study the influence on the iron hyperfine parameters of vanadium substitution at the iron sites for different geometrical and atomic arrangements. It is also possible to investigate if the V/Fe substitution is randomly distributed or if V prefers a particular iron site.

Four samples with x » 0, 0.03, 0.25, and 0.70 were prepared. Representative spectra are displayed in Figs. 1-2. Fittings of theoretical Mossbauer patterns to the measured spectra gave the resultx presented in Table 1. For the two samples having the hep phase, a spinflip transition occurs at 380 K. The spins are parallel with the c-axis for T - 380 K, and lie in the c-plane at T - 380 K. The linewidth and the electric quadrupole splitting change markedly at the spin flip temperature (150 1 and 0.09 mms* respectively). This can be explained by the fact that the electric quadrupole splitting, in this case given by .„ - e^ zz 3 cos 9 - 1 •*• n sin 9 cos 2$ AEQ , is a function of the angle 6 between V and the hyperfine field B, -. At the zz tiz transition,0 changes from 90 (T - 380 K) to three different values of approximately 15°, 45° and 75° corresponding to three different V -directions in the c-plane. The strength of the electric quadrupole interaction has been

eQ V Z2 -1 determined to ———c+ 0.33 mm s for Fe-Ge. The main influence of 3 Z V/Fe substitution in Fe.Ge is: 2) a decrease of B, . from 26.1(2) T for 6 Fe as nn to 22.5(2) T for 5 Fe and 1) V substitutes Fehi randomly a decrease of B, - fi hi 3) V1 Vha sa sa nsmaln (Tl - («10K« -0.0) 2 on s~ per V as nn) effect on the isomer shift for iron 4) the Curie temperature decreases by 7 Z from 650(5) K to 603(5) K. P 3 B 5

In Che cubic closed packed phase there 295 K exist two different crystallographic metal sites. It is found that V populates only one type of metal site. If V/Fe substitution is random within this metal site and the -40 -2.0 0 2.0 40 hyperfine parameters VELOCITY (mm/s) are mainly determined Fig. 1 Mossbauer spectra of (Fe V ) Ge. The large by the short range Q ? and small intensity pattern corresponds to iron interactions we expect nuclei having 0 and 1 V in the 1 nn shell 5 different Mossbauer respectively. patterns with the intensities given in Table 1. The spectra are in good agreement with this model. At four of these five different sites, iron experiences a magnetic field. This fact and the similarities with a-iron give a good 10K opportunity to study the hyperfine parameters for iron with many different well defined near surround- ings (Fig. 3). The change in isomer shift -8.0 -40 -2.0 0 2.0 40 8.0 is found to be +0.075 VELOCITY (mm/s) mm s and 0.02 mm s per Ge/Fe substitution Fig. 2 MBssbauer spectra of (FeQ _.V0 _5),Ge. The bars in the 1 nn and 2 nn indicate the outer line in the different six-line shells, respectively. patterns, except for the single line pattern with' 4 V in the 1 nn shell. P 3 B 5

The influence of V/Fe substitution on the isooer shift and the quadrupole splitting is however not possible to resolve. The spin polarization effect on the magnetic hyperfine field for iron has been estimated to

-8.6Z, -0.42 and -0.6Z OF* IF* 2ft 3Fe 4F* 1 *G» 4G* 4O IGt Ujt SO ?G» TO Mnn.-2nn ooys 6c per Ge/Fe (1 nn) 4V 3V ?v IV - - - - > 3nn always 12 F* Ge/Fe (2 nn) and Fig. 3 M«fn*tic hyperfint fielc ve.'ut: !IT iror jrres found V/Fe (3 nn) *s functior. of ntarest neighbour turrouniiinf i. substitution respectively. The Curie i B r r Compound (T) (!; temperature Fe3Ce 0.39 26.0 0.06 100 0.33 vas 0.39 0.37 26.1 22.i O.Oi 0.04 84 16 0.33 0.33 (F"O.9?VO.O3)3G* determined to be 295(5) K.

« V J Gt 75 0.2. > 3 V Ce '0. (1 30 (J.7C IC) 15. 8 * .0 17 .7 3i .5 27.0 49 42 9

Urns./1.) 0. 2b 0 .40 0 .40 0.40 0.40 0 32 0 37 0 .41

B(T) 30. 0 16 .6 12.2 7.0 0 0 0 0

itQ(»/t) 0 0 0.02 0 .06 :0.04 10 26 10. 35 :0 .41

0.41 0. It 0. 14 0 .14

T*b)t 1. Xrluli ei fitting of ch» Mctcbauti *p«ccr< Ai a tnpcrtture cf 10 K. £ it CIH- cencroid thiil. ik i» ihc tlectric qu4drupc.lt

The simple cubic compound {& -.V nonffia?netic down to at least 5 K. Here a decrease in isoner shift of -0.05 am s~ is found for a V/Fe (1 nn) substitution (Table 1). P 3 B 6

B CRYSTAL STRUCTURE OF Hd2CollfB. MAOETIC PROPERTIES OF HdjCo^B ABD *2°°1A D. Le Roux, H. Vincent and D. Fruchart Laboratoire de Cristallographie, associe" It l'U.S.M.G. C.N.R.S., 166X, 38042 Grenoble Cedex, France.

P. L'heritier and R. Fruchart ENSIEG, ER 155, B.P. 46, 38402 Saint Martin d'Heres, France.

Nd_Co1 B, isotype to Nd.Fe^B (1) has been prepared, for the first time to our knowledge. Constituents were meliwby means of induction heating in an evacuated so called cold crucible, at about 1200°C. Crushed ingots were then annealed for several hours at 1000°C. Single crystals suitable for structure determination and magnetostatic measurements could then be extracted from the bulk. A spherically ground fragment of .007 cm radius (uR = 1.'+) was used to measure 5706 reflexions with Z > 0 on an ENRAF-NONIUS CAD4 diffractometer at Ag radiation, up to 9 = 25°. Accurate cell parameters were found to be : a r 8.646(2) A c = 11.864(5) A and no hOZ, h+2= 2n + 1 reflexion was observed. Averaging of equivalent reflexions in —mm Laue class led to 654 independent observations. We then performed a least square refinement of atomic parameters in P—run space group, using all the reflexions, to R = .011 and R = .013. The weighting scheme was ui = -j , ' 0 (I) ( 01I with a(F) = ^[ Q . * .- > fi where aQ(I) is the standard deviation on observed intensities. Extinction was taken into account by means of SDP (2) correc-

Final positional and thermal parameters are listed in table I. We noticed sytem- atic differences between two sets of unaveraged equivalent reflexions, namely hk£, Rk2, h£e, HKi on one hand, and klu, fch£, k5i, EHi on the other. Moreover, hk.2 reflexions with |h| > |k| were systematically about 10 % stronger than those with |h| < |k|. This was observed on several samples. However, attempts to des- 42 cribe the structiire in orthorhombic subgroups of P—ran, with all or half the reflexions averaged in mmm class didn't give significantly better results. Fur- ther investigations by means of electron diffraction and imaging are in progress.

Preliminary results of ordering temperature determination on polycristalline samples of Hd.Co B and Y.Co^B (prepared in a similar way as the former) show a magnetization vs temperature dependence as in figure 1 (H applied 'v 100 Oe). M variation in the vicinity of T = 270°C for Md-Co.^B is probably due to a spin reorientation. Ordering temperatures are T = 720°C and T = 730°C respectively. P 3 B 6

Magnetostatic measurements on oriented single crystals of both compounds, from liquid helium to Curie temperature are being carried out. Nd.Co. B exhibits

axial anisotropy (parallel to c), while Y^Co^B (on the contrary to Y2Fe14B (3)) has a preferred plane of magnetization (perpendicular to c). Neutron powder diffraction patterns have been- recorded on both compounds in the same temperature range and investigation of magnetic structures are in progress.

Table 1 : Positional and thermal parameters . < .0005 A2 for Nd and Co, A^. < .004 A2 for B)

X z y Ull U22 U33 U12 U13 U23 4f .14414(4) X 0 .0062 .0059 -.0018 0 Ndl Ull Nd .72471(4) X 0 .0062 .0052 -.0014 0 0 2 tg Ull C°l 4C 0 H 0 .0077 .0088 .0066 -.0001 0 0

Co2 16k .72375(4) .06925(4) .37357(3) .0048 .0069 .0072 -.0011 -.0019 -.0001

16k .46300(4) .13985(4) .32192(3) .0054 .0051 .0059 .0007 -.0006 C03 .0010 .13165(8) X .25336(5) ,0064 .0041 .0009 -.0003 C\ 8j Ull U13 .40123(8) X .29495(5) .0059 .0063 .0011 .0010 C05 3j Ull U13 4e 0 0 .33420(6) .0057 .0064 .0026 0 0 C06 Ull B 4f .37665(96) X 0 .0103 .0074 .0018 0 0 Ull

Figure 1 : Magnetization vs temperature curves for and

Y2Co,4B

270 720 730

(1) C.B. Shoemaker, D.F. Shoemaker and R. Pruchart Acta Crystallographica, to appear (198«O (2) Structure Determination Package, Enraf-Nonius, Delft, Holland. (3) D. Givord, H.S. Li, R. Perrier de la B^thie Solid State Communications Sl_, 857-860 (198*)

D. LE ROUX Laboratoire de Cristallographie, associe" a l'U.S.M.G., P 3 B 7

Fe B ANALOGY CF INSERTION AND STACKING WAYS IN Na2 14 ' AND Fe5SiB2 PHASES.

Ph. l'Heritier, P. Chaudouet, R. Fruchart

E.R. 155 du CNRS, EUSIEG, Comaine Universitaire, B.P. 46, 38402 Saint Martin d'Heres, Prance.

C.B. Shoemaker, D.P. Shoemaker.

Department of Chemistry, Oregon State University, Corvallis, Oregon 97331, U.S.A.

The tetrahedrally close-packed structures of intermetallic compounds are characterized by exclusively tetrahedral voids and cannot acconodate "interstitial atoms" like boron or carbon. In many transition metals compounds, boron and carbon are in trigonal prisms. Consequently, the formation of ternary phases like MnsSiC {1) and Nd2 Fei4 B (2, 3, 4) needs a transformation of the close-packed structures from which these structures may be derived : laves-types structure (MgZn2) for MnsSiC (Fig. 1) and sigma-phase B Fi 2 for Nd2 Fei4 ( 9- )' A mirror plane occurs every second less densely populated plane, so that the pentagonal sites are systematically changed into pentagonal prisms (Mr^SiC) or hexagonal prisms (Nd2 Fe^4 B, Fig. 3). Similary, Fe5SiB2 (5) may be derived from an hypothetical structure "Fe^i" of Fe2B type (Fig. 4) in which the squared antiprisms surrounded by tetrahedral interstices are changed into squared prisms surrounded by trigonal prisms.

The structures of Fe5SiB2, Mn^SiC and Nd2Fei4B have the same stacking sequence : B' A B Aa B' Aa B A B' (6). Depending on the compound, the A plane consists of nets of squares and triangles (or pentagons and triangles) (or hexagons and triangles : "Kagame nets"), the B plane corresponds to the squared antiprisnatic site (or pentagonal) (or hexagonal), the A& plane is anti- symmetrical to the A plane, and B' is symmetry plane. P 3 B 7

Fig. 1.

Structure of MngSiC (orthorhontoic One 2^) projection on (100)

. Si atom X s 0.250 _X a 0.129

O Mn X s 0 X = 0.375 o C X =0

triangular prlim* with common raca : Fig. 2. P2typ«

Structure of

(tetragonal P42/nmm) "Kagoite nets", projection on (001) A" l

Z;0 Z = 1/2

Fig. 3.

Structure of atoms in mirror plane (B1) at z » 0 and z » 1/2 P 3 B 7

Fig. 4.

Structure of (tetragonal I4/mcm) sequence B plan*: a ABA* A* B"A* B*pl«M: projection along the c axis. fl boron* Fe5SiB2 i derived from structure (7)

B2- U3SI2trp«

Fe5SiB2 1 •F«2Si'VF»3B2"

(1) P. Spinat, R. Fruchart, M. Kabbani and P. Herpin, Bull. Soc. Fr. Min. Crist., 93_, 171 (1970) (2) J.F. Herbst, J.J. Croat, F.E. Pinkerton and W.B. Yelon, Phys. Rev. B, 29, 4176 (1984) (3) D. Givord, H.S. Li and J.M. Moreau, Solid State Com., 50/ 497 (1984) (4) C.B. Shoemaker, D.P. Shoemaker and R. Fruchart, Acta Cryst. C, 40, 1665 (1984) (5) B. Aronsson and I. Engstrom, Acta Chan. Scand., 14, 1403 (1960) (6) Ph. l'Heritier, P. ChaudouSt, R. Fruchart, C.B. Shoemaker and D.P. Shoemaker, submitted to J. Solid State Chem. (7) F. Bertaut and P. Blum, C.R. Acad. Sc. Paris, 236, 1055 (1953)

Ph. l'Heritier, E.R- 155 du CNRS, E.N.S'.I.E.G. B.P. 46, 38402 Saint Martin d'Herea. France. P 3 B 8

STRUCTURAL AND MAGNETIC PROPERTIES OF THE NEW REMnSi2 SILICIDES (RE - La-Sm) B. Mai"'", G. Venturini, M. Méot and B. Roques Laboratoire de Chimie du Solide Minéral, associé au C.N.R.S. ns158, Université de Nancy I, B.P. 239, 54506 Vandoeuvre les Nancy Cedex (France) D. Fruchart Laboratoire de Cristallographie,'C.N.R.S., associé à l'Université Scientifique et Médicale de Grenoble, B.P. 166X, 38042 Grenoble Cedex (France)

New ternary suicides with the formula REMnSi« (RE " La-Sm) have recently been isolated and characterized by X ray powder diffraction. Their magnetic properties are studied by means of susceptibility measurements between 4.2 and 750 K and neutron diffraction.

Structural properties These compounds crystallize in the TbFeSi. type structure (1) ; their lattice parameters are listed in Table 1. Their structure type and those of a-ThSi. and ThCr-Si- are closely related ; all of them are built up of similar atomic planes whose stacking sequen-

ces are respectively : (RE)XX in a-ThSi_, (RE)XX(RE)XT2X in TbFeSi» and (RE)XT_X in ThCr-Si- (RE - rare earths, T - transition metals and X « Si.Ge).

Magnetic study The REMnSi- compounds behave as normal ferromagnets below T_ varying from 385 to 440 K ; in addition, NdMnSi« displays a strong increase in suscepti-

bility at 40 K whereas PrMnSi2 becomes antiferromagnetic below 35 K (fig.I, 2 and Table 2). For all of these compounds, the reciprocal susceptibility obeys ex Curie-Weiss law at high temperatures according to the ferromagnetic behaviour. Neutron diffraction patterns have been recorded for PrMnSi- at 250 and 4.2 K. At 250 K, results confirm the ferromagnetic order of the Mn sublattice ; the moment lies in the (010) plane. At 4.2 K, data indicate that the structure is characterized by ferromagnetic layers of Pr atoms piled up along the b-axis with the sequence -M—, as shown on figure 3. The Mn layers are ferromagnetically coupled with the adjacent Pr layers, as it is usual with the light rare earths. The moments still are in the (010) plane.

Discussion Neutron diffraction studies ara in progress for the other RE compounds. However, results concerning PrMnSi- suggest that Mn subl2ttices always have P 3 B 8

Table I. Lattice parameters for REMnSi- compounds

Compounds a (A) b(A) c(A)

LaMnSi. 4.191 (3) 17.68 (1) 4.073 (3)

CeMnSi2 4.123 (2) 17.549 (9) 4.035 (3)

PrMnSi2 4.094 (3) 17.553 (6) 4.028 (4)

NdMnSi2 4.072 (0 17.487 (5) 4.017 (2) SniMnSi- 4.035 (I) 17.409 (6) 3.998 (1)

Table 2. Magnetic data for REMnSi. compounds

Moment at 4.2 K Compounds Transition temperatures (K) mole"') and 20 kG(uB) Tc

LaMnSi. 1.3 (1) 386 395 2.6

CeMnSi2 1.1 (1) 398 420 2.1

PrMnSi2 1.4(1) at 77 K 434 35 450 3.1

2.8(1)-(1.5 at 180K) NdMnSi2 40-441 460 4.2

SmMnSi2 1.5 (1) 464 470 3.3 P 3 B 8

the same anisotropic ferromagnetic order at high temperatures ; nevertheless, the interlayer interactions are rather strong in spite of great interatomic distances (y 9 A.) ; these interactions decrease when the distances increase between SmMnSi- and LaMnSi.-,* These conclusions are consistent with the magnetic measurements, At low temperature,, the increase of NdMnSi- susceptibility indicates a ferromagnetic ordering of Che Nd sublattice. However, the coupling scheme of Nd and Mn sublattices is not clear, since Mn atoms have a moment u - 1.3 11. a according to measurements on LaMnSi.- This study will be completed before the Conference and the magnetic interactions will be discussed.

REFERENCE (1) V.I. Yarovets and Yu.K. Gorelenko, Vestn. L'vovsk. Univ., Ser. Khirn., 23, 20 (1981). !W emu/9) ' • * Nd 60 60

500 TCK) fig. 1 : Magnetization vs temperature data for RE MnSi-.

field at 4.2 K for REMnSi2. PrMnSJ2 magnetization was also measured at 77 K.

3 : Magnetic structure of PrMnSi2 at 4.2 K. (Open circle : x » 0 ; dashed circle : x - 1/2). P 3 B '?

NEW TERNARY RARE EARTH - TRANSITION METAL GERMANIDES WITH TiMnSi, , Sc-Co.Si«,, 2 ' 5 4 10 OR U^COjSij TYPE STRUCTURES. SUPERCONDUCTIVITY IN THESE COMPOUNDS H. Meot, G. Venturi ni, E. Me Rae, J.F. Marêché, B. Malamsn and 8. Roques Laboratoire de Chimie du Solide Minéral., associé au C.N.R.S. n° 158, Université de Nancy I, B.P. n° 239, 54506 Vandoeuvre Les Nancy Cedex (France)

Mors than forty new germanides of formula RE^-T^Ge^,

With Lu5Rh4Ge1Q (Tc = 2.2 K) and Lu5Ir4Ge10 (Tc = 2K), there are now nine superconducting compounds in this class and some relationships appear between Tc and composition (Table 2). Germanides with U^Co^Si- type structures With T = Ru, we have characterized the first group of germanides which crystallize in the U_Co Si- type structure : REpSu-Ge- with rather large RE elements (RE = La-Sm, Gd-Er), as in the isostructural silicides. Recent synthesis of Tb,0s^Ge_ suggests the existence of a second group of isotypic

RE20s,Ge5 compounds with at least RE ~ Tb-La.

Y,Ru,Ge5 and La^Ru^Ge- do rv.t exhibit superconducting transition above

1.4 K, contrary to isotypic YjRhjSig (Tc = 2.7 IO and La-,Rh3Si5 (Tc = 4.45 K> (8). This difference confirms the strong correlation between the superconducting properties and the valence electron concentration. P 3 B 9

Table 1 the SHicides and germanides 2 Sc5Co4Si10(*), Sc2Fe3Si5 or CeNiSi2(A) types. These compounds are listed versus RE sire. New phases are indecripted by ; others are reviewed in (4)

N. T Ru OS Co Rh Ir

RE N. Si Ge Si Ge Si Ge Si Ge Si Ge

Sc ->-v *,* ->-7 X,* * Lu 9 e +v *,$,A -+ * * •+* * -rk Yb -HIT -Mir (III) (III,II) (III,II) (II,III:

-<•* * Tin •*%,•** A Er -K) +7,-*-* A 0 ->•* * Ho -K) •** A 0 •+* *

Dy •+O -»•* «,A 0 -He *

Y 9 •*C 9 •* * *

Tb -H) A 0 -+* 0 •*•*

Gd -K> A 0,A -Mir 0,A Eu(III) Sm ->0 A 0,A A Ce(IV) A Yb(II) Nd A 0,A A

Pr •*o A

Ce(IIl) •+0 0 A A La -K) 0,A A Eu(II)

Table 2 Supe-conducting transition temperatures (K) for the Sc5CO4Si-jg type faraily. The temperatures correspond to 90 and 10% points ; n.d. = not detected above 1.4 K ; (7) : reference

Co Rh Ir Os RE5T4(Si^Ge7^

Lu5T4Geio n.d. 2.2-1.6 2.01-1.94 l-U5T4Si to 3.76-3.72(7) n.d. 2.62-2.58(7) 8.6S-8.4K7) Y5T4Se-jo Y5T4S110 3.0-2.3(7) Sc5T4Si-)0 5.0-4.8(7) 8.54-8.45(7) 8.46-8.38(7) P 3 B 9

REFERENCES (1) J. Steinmetz, G. Venturi ni, B. Roques, N. Engel, B. Chabot, E. Parthé, Acta Cryst., B38, 2103 (1982) (2) H.F. Braun, K. Yvon, R.M. Braun, Acta Cryst. B36, 2397 (1980) (3) L.G. Akselrud, Ya. P. Yarmolyuk, E.I. Gladysehvskii, Sov. Phys. Crystal- Logr., 22, 492 (1977) (4) E. Parthé, B. Chabot, in Handbook on the Physics and Chemistry of Rare Earths, K.A. Gschneider, Jr. and L. Eyring, eds., Elsevier Science Pu- blishers, 113-334 (1984) (5) G. Venturini, J. Steinmetz, B. Roques, J. Less-Common Met., 87, 21 (1982) (6) M. Meyer, G. Venturini, B. Malaman, J. Steinmetz, B. Roques, Mat. Res. Bull., 18, 1529 (1983) (7) H.F. Braun and C.U. Segre, in "Ternary Superconductors", G.K. Shenoy, B. Dunlap, F.Y. Fradin, eds., Elsevier North Holland, 239-242 (1981) (8) B. Chevalier, P. Lejay, J. Etourneau, M. Vlasse and P. Hagenmuller, Mat. Res. Bull., 17, 1211 (1982) P 3 B lo

ORIGIN OF TOE STRUCTURAL DISTORTION BETWEEN PHASE I1 AND I IN COMPOUNDS S. M.iraglia, J.L., Hodeau, J. Chenavas, M. Marezio Laboratoire de Cristallographie du C.N.R.S., associe" a l'U.S.M.G., 166 X, 380^2 Grenoble Cedex (France). C. Laviron, M. Ghedira Faculte des Sciences et Techniques de Monastir, Monastir (Tunisie). J.P. Remeika Bell Laboratories, Murray Hill, N.J. 07974 (U.S.A.).

Compounds with formulae SnMJlh^Sn.. (M = La-Gd, Ca, Sr, Th) and (M = Tb-Lu, Y, Sc) have remarkable superconducting and magnetic properties. They become superconducting or undergo a magnetic transition at temperatures lower than ^ 10 K (1). The SnErl+RhfiSn1fl compound exhibits reentrant superconductivity (T MK, I ^ 0.6 K). Four different phases [I (2, 3) I1 (4), II (S) and II' (6)] exist for these compounds. The structures of all these phases contain a three dimensional network of corner-sharing RhSn5 trigonal prisms. The purpose of this paper is to determine the main difference between phase I1 and I. In the structure of these two phases, this network generates icc^ahedral and cuboctahe- dral sites which are occupied by the Sn(l) and M atoms (figure 1, a, b, c).

Figure 1 : Coordination of Rh (a), M (b) and Sn(l) (c) sites by Snf2) atoms in Sn(l)M3RhltSn(2)12.

These two atoms form a sublattice having the arrangement of an A15 structure.

Sn(l)M3Rh^Sn(2)12 compounds crystallise with the phase I or I' structure. The distortion of the phase I1 structure with respect to that of the phase I, is very small. Only very long exposure (200 hrs) precession photographs revealed the existence of superstructure spots which lead to double the three cubic phase I paroseters. It is, thus a good approximation to refine the structure of o phase I' compounds by using the phase I cell (az i, 9.7 A) and its space group (Pm3n). On this basis, we report here the results of structural refinements for Bio

the following compounds : Sn(l)M3Rh^Sn(2).*, M = La, Ce-, Pr, Gd, Th, Eu, Yb, Ca. There are not any appreciable shifts in the positional parameters between the compounds having the phase I structure and those having the phase I1 structure.

They all agree very well with the positional parameters of Sn(l)Yb3Rh^Sn(2)12 (phase I) (2). Consequently the interatomic distances and angles vary less than 2 % and 2°, respectively, on going from one compound to another. On the contrary, small differences exist in the thermal data of these compounds. Since they are significant, a detailed analysis allows the determination of the important features for the phase I1 distortion. Table 1 : Sn(2) thermal ellipsoid amplitudes and orientation of its largest axis, relatively to the cell axes, for different M compounds. La Ce Pr Gd Th Eu Yb Ca

0 P^A) 0.213 0.215 0.210 0.208 0 242 0.153 0.158 0.151

r2(A) 0.084 0.080 0.082 0.090 0.094 0.092 0.GJ5 0.093

r3(A) 0.075 0.070 0.071 0.079 0.083 0.080 0.082 0.081

.Angles of r1 with : X 90 90 90 90 90 90 90 90 y 18.2 19.3 18.3 18.9 25 .1 14.0 14.6 12.6 z 71.8 70.7 71.7 71.1 64.9 76.0 75.4 77.4

It can be seen from table 1 that a significant difference exists for the major axis of the thermal ellipsoid of the Sn(2) atom between the compounds with M = Eu, Yb, Ca, (phase I) and those with M = La, Ce, Pr, Gd (phase I1). It o o is on the average 0.154 A for the former compounds and 0.213 A for the latter. The compound with M = Th is unique, as it crystallizes with the structure of phase I, but the major axis of the Sn(2) thermal ellipsoid is the largest o (0.242 A) among the eight compounds reported here. An explanation will be given at the end. The other thermal data are practically the same for all eight compounds. Since the structure of the phase I1 compounds has been refined in a more symmetrical space group, the 40 % increase of one of the thermal-ellipsoid axes does not represent an actual increase in the thermal motion. It is rather due to a static distortion which, in the case of an average structure, reveals itself as an anomalous thermal vibration. In the phase I1 compound the major axis of the Sn(2) thermal ellipsoid makes an angle of *> 18° with the b axis and one of about 12° with the direction of the Sn(l)-Sn(2) bond. In phase I compounds, the corresponding angles are 14° and 16°, (figure 2, a, b). In all compounds, the major axis of the Sn(2) atoms is perpendicular to the Rh-Sn(2) and M-Sn(2) bonds. Therefore, the distortion of the P 3 B lo

phase I' structures consists mainly in the displacement of the Sn(2) atoms along the Sn(l)-Sn(2) bonds. The Sn(2) icosahedra around the Sn(l) atoms lose the m3 point symmetry and the twelve Sn(l)-Sn(2) distances do not have to be equal anymore. By comparing the results of the Th compound with those of the other compounds it can be deduced that the structure of SnTh-Rh^Sn.- has the phase I1 distortion, but there is no correlation between the distortion of any given Sn(l) site and the Sn(l) neighbours.

Figure 2 : Sn(2) thermal ellipsoid sizes and orientations, relatively to Sn(l)-Sn(2) bonds in phase I (a) and in phase I1 (b). The way the coordination polyhedra are arranged in the structure and the detailed analysis of distances led to the conclusion that these compounds have a strong covalent/ionic character rather than pure intermetallic one. The Sn(l), M and Rh atoms have a cation behaviour while the Sn(2) atoms have an anionic behaviour. It is reasonable to assume that the Eu, Yb, Ca and Sr atoms are in the divalent state, the La, Ce, Pr, Nd, Sm and Gd atoms are in the trivalent state, and the Th atom is in the tetravalent state. The structure of the M3+ and M-+ 2+ M compounds is distorted with respect to that of the M compounds. The structural analysis of the entire series corroborates the idea that an electron transfer takes place between the different sites.

(1) J.P. Remeika, G.P. Espinosa, A.S. Cooper, H. 3arz, J.M. Rowell, D.B. McWhan, J.M. Vandenberg, D.E. Moncton, Z. Fisk, L.D. Woolf, H.C. Hamaker, M.B. Maple, G. Shirane and W. Thomlinson, Solid State Comm., _34, 923 (1980) (2) J.L. Hodeau, J. Chenavas, M. Marezio and J.P. Remeika, Solid State Comm. J6_, 839 (1980) (3) J.M. Vandenberg, Mat. Res. Bull. _15, 835 (1980) (4) J.L. Hodeau, M. Marezio, J.P. Remeika and C.H. Chen, Solid State Comm. 42, 97 (1982) (5) J.L. Hodeau, M. Marezio and J.P. Remeika, Acta Cryst. B40_, 26 (1984) (6) J.L. Hodeau, These, U.S.M.Grenoble (1984)

J.L. Hodeau Laboratoire de Cristallographie C.N.R.S. 166 X 38042 Grenoble Cedex P 3 B 11

INVERSION BOUNDARIES IN M^ BOREDES J.P. Morniroli, M. Khachfi and M. Gantois Laboratoire de Génie Métallurgique (U.A. 159 CNRS) Ecole des Mines - Parc de Saurupt, 54042 NANCÏ-Cedex (France)

INTRODUCTION The structure of the three borides Ru-B., Rh^B. and Re-B- has been determined from X-ray experiments by Aronsson [1,2] and confirmed later by electron diffraction experiments [3]. It has hexagonal symmetry, belongs to the non-centro-symmetrical space ërouP P6~mc and is very close to the M_C_ carbide structure. Both boride and carbide have the same structural elements namely, metal atom octahedra, metal atom tetrahedra and prismatic interstices containing carbon or boron atoms, but these elements are arranged differently. As a result, the main difference between M_B, and ìi-C, concerns the local environment in prismatic interstices of the tetrahedra. Each of them has a typical local environment.

M7C. carbides usually rouCain many twins and antiphase boundaries which neither destroy nor modify the M^C, typical local environment [4]. On the contrary M_B_ borides exhibit very few defects [3].

The aim of this study is to characterize these defects by means of electron microscopy, in order to check whether their presence is linked to the local environment typical of the boride.

EXPERIMENTAL RESULTS Very rare planar defects are observed by electron microscopy in O thin foils. Some of them exhibit a bend (Figure 1).

Trace analysis indicates that the fault planes are the three equivalent {lToo} planes.

If contrast experiments are done with crystals having their c axis parallel to the electron beam, defects are out of contrast in bright field as well as in dark field, using one or several (hk.O) diffracted beams contained in the zero order Laue zone. This fact indicates that the defects do not affect the x and y coordinates of the atoms. Only the z coordinates are modified. With other diffracted beams, the defects are imaged as a fringes i.e. black and white fringes which are symmetrical in bright field and asymmetrical in dark field (Fig. 1 a,b).

From these properties, we postulate the model of a defect given in figure 2. It represents an inversion boundary between two crystals parts, where one structure is the inverse of the other. This defect is possible because the structure of M_B_ is non centro-symmetrical. P 3 B 11

(a) (b) (c)

U Figure 1 : Inversion boundaries in R 7B, (a - b) : Symmetrical and asymmetrical fringes in bright and dark field images, (c) : Contrast between domains on both sides of the boundaries obtained in multiple beam situation using (hk.1) diffracted beams.

o o

O O

xx X % X ./ Cb) a Local environment typical of MB. D Local environment O 0.932 typical of M C • 0.4.32 © 0.25 M *• Inversion point at 0 0.068 e 0.75 3-1/2 • 0.568 (a) Figure 2 : Model of inversion boundary in M^ (a) Projection on the (OOOl) plane, (h) Schematic description. P 3 B 11

Inversion boundaries have been observed in other materials (6 NiMo, x phase, GeTe,...) and were studied theoritically by Amelinckx et al [5] who indicated the way to identify them : a contrast between the domains situated on both sides of the boundary oust be observed in dark field, if a multiple beam situation prevails, anH if the reflections used do not belong to a zone giving rise to a centro-symmetrical projection, (hk.1) diffracted beams contained in the 1st order Laue zone with crystals having the c axis parallel to the electron beam do not give a centro- symnetrical projection and are suitable for inversion boundary identification. They were used to form dark field images which clearly show a contrast between the domains and validate the hypothesis of the presence of inversion boundaries in M3 (Figure lc).

CONCLUSIONS Very rare defects observed in ^B, borides are identified, by means of contrast experiments, as inversion boundaries separating domains in which 6- axes have opposite directions. One notices that at these boundaries, the local environments typical of the borides are destroyed and replaced by the ones typical of M^C., carbides. For this reason, such interfaces should have a high energy and should not occur frequently as observed experimentally.

[1] B. Aronsson, Acta Chem. Scand., _T3> I09> 0959) [2] B. Aronsson, E. Stenberg and J. Aselius, Acta Chem. Scand., _l£, 733, (1960) [3] M. Khachfi, E. Bauer-Grosse, J.P. Morniroli, T. Lundstrom and M. Gantois, Revue de Chimie Minerale, 2j_, 370, (1984) [4] J.P. Morniroli, E. Bauer-Grosse and M. Gantois, Philos. Mag., ^8, 311, (1983) [51 S. Amelinckx and J.Van Landuyt, Diffraction and Imaging Techniques in Materials Science, eds S. Amelinckx, R. Gevers and J. Van Landuyt, North-Holland Publishing Company, 1978, Amsterdam

J.P. Morniroli. Laboratoire de Genie Metallurgique (U.A. 159 CNRS) Ecole des Mines - Pare de Saurupt, 54042 NANCT-Cedex (France) P 3 B 12

CRYSTAL GRCWIH AND TRANSPORT PROPERTIES OF THE REFRACTORY METAL

DISXLICXDES MoSi2 AND WSi2-

J.P. SEMATEOR, 0. TB3MAS, R. MADAR

ER. 155. CNRS. ENSZEG. Donaine Uhiversitaire BP 46, 38402 Saint Martin d'Heres, France.

O. LABORDE

C.R.T.B.T., CNRS, B.P. 166 X 38042 Grenoble Cedex, France.

E. ROSENCHER

C.N.E.T. B.P. 98 38243 Meylan Cedex, France

During the last few years metal silicides have been subjected to intensive analysis both from a theoretical standpoint and for their potentialities in the silicon based very large scale integration technology (MoSi2»

HDwever, despite their technological importance, little is known about the transport properties of these materials. Bulk data are scarce (1) (2) and moreover have been questioned recently (3) since they have been measured on sintered materials. low temperature data are absent. Only recently have been reported some measurements on thin films obtained by CVD, sputtering or evaporation on silicon substrates (3) (4), but accurate studies on well characterised single crystals are generally unavailable. This situation results from the difficulties encountered in the synthesis and crystal growth of these materials : they have very high melting temperatures (.a 2000" C) and react with alnost all kind of crucibles. In order to determine accurately their transport properties, we have grown large single crystals of the disilicides which are of interest for P 3 B 12

electronic devices. In this paper we will focus on MoSi2 and WSi2<

Experimental Single crystals of MbSi2 and WSi2 have been grown by Czochralsky pulling from the melt in a modified Hukin type cold crucible with a 40 Hit/ RF hea- ter. The main problem for Czochralsky growth in Hukin type crucible is related with the volume variation of the melt during the pulling operation ; this leads to a variation of the tenperature and generally the part Which can be pulled weights less than 50 % of the original melt. In our modified crucible a solid bar of the material is pushed into the melt from the bottom of the crucible at a speed which compensates exactly the volume of the crystal pulled. The synthesis is obtained by direct melting, under 1.5 bar of pure argon, of silicon lumps (99.999 %) and metal rods 99.99 %). The synthesis and the crystallisation are carried out in the same crucible. In a first step polycrystalline rods (z 3 x 3 x 30 mm) of tetragonal MoSi2 and VJSi2 are obtained by pulling from the melt with a tungsten needle. Despite the fact that numerous cracks exist in these rods, in relation with the anisotropy of the thermal expansion coefficient, some nonocrystalline seeds can be extracted from these samples. In»second step, starting from (001) oriented seeds, we have been able to obtain large single crystals of MoSi2 ( 8 ma, 1 = 30 iim) and WSi2 with out any crack.

Transport properties Measurements of the electrical resistivity have been made on rectangular monocrystalline bars between 4.2 and 293 K for both [100] and [001] orientations. Figure 1 shows a typical measurement of the resistivity versus tempera- 2 ture for a MoSi2 bar along the [100] direction. The straight line give the T coefficient of the variation of the resistivity with tenperature. These results will be discussed in conparison with those obtained previously for thin films (4).

J.P. SENATEUR ER 155 CHBS. ENSIBG. BP 46, 38402 Saint Martin D'heres, France. P 3 B 12

References (1) D.A. ROBINS Philos. Mag, 4, 322, (1958) (2) V.S. NESHPOR and G.V. SAMSCNÖV Sov. Phys. Solid. State, 2, 1966 (1960) (3) S.P. MUPARKA, M.K. REED, C.J. DOHEETY and D.B. FICASER J. Electrochem. Soc, 129, 293 (1982) (4) P.H. WOERLEE, P.M. Th. M. VAN ATTEKÜM, A.A.M. HOEBEN G.A.M. HUPKX and R.A.M. WXTERS. Appi. Phys. Lett., 44, (9), 876 (1984)

1 10 1000 T(K) Figure 1 : Logarithmic plot of the tenperature dépendance of resistivity for

MoSi2 single crystal along the [100] direction.

J.P. SENATEUR ER 155. CNRS. ENSIEG. BP 46, 38402 Saint Martin d'Hères, Franc«.! P 3 B 13

NEW TERNARY RARE EARTH - TRANSITION METAL GERMANIDES WITH YbjRh^Sn^,

BaNiSiw U4Re?Si6, ThCr2Si2 OR CaBe2Ge2 - TYPE STRUCTURES. SUPERCONDUCTIVITY IN THESE COMPOUNDS G. Venturini, M. Méot, H. François, B. Nalaman, J-F. Marêché, E. Me Rae and B. Roques Laboratoire de Chimie du Solide Minéral, associé au C.N.R.S. n" 158, Université de Nancy I, B.P. n" 239, 54506 Vandoeuvre les Nancy Cedex (France)

The new germanides reported in this work belong to the systems RE - T - Ge (RE = rare earths, T = Co, Rh, Ir, Ru, Os). They are divided into two groups according to their compositions : - phases of the Yb,Rh,Sn.., CD or BaNiSo, (2) types,

- phases of the U,Re7Si, (3), ThCrjSi' (4) or CaBeJîe2 (5) types. These compounds have been characterized by X-ray powder diffraction. Structural changes in each group are discussed. The occurence of superconductivity is investigated. Germanides with the Yb,Rh,Sn.., or BaNiSa, type structures The new phases are presented in Table 1, with already known germanides (6) and stannides (7) of the Yb,Rh,Sn., type ; for each T eLement, these compounds are listed as a function of the RE atom sizes. The stabilities of the two structural classes appear to be determined by the relative sizes of RE atoms on one hand and T and X = Si or Ge atoms on the other. This relation can be explained by the architecture of the Yb,Rh,Sn.., and BaNiSn, types in which RE atoms are embedded in three-dimensional networks of T and X atoms, with strong covalent interactions. These networks are very rigid and, in each of them, the size of the RE sites is determined by T-X or X-X distances. In the germanides of the YbjRh^Sn.., type, the RE atoms occupy rather small germanium cuboctahedra which become stretched to their limit when the RE size increases ; the structural change to the BaNiSn, type, with larger RE sites, reduces the T-X distances to normal values (fig.). Germanides with the l^Re-ySi^, ThCr,Si, or CaBe^Ge, type structures The new phases are presented in Table 2, with other already known members of their structural classes.

The structural changes from the U^Re-Si, to the ThCr2Si2 or CaBejGe^ types are chiefly governed by the same size factor as in the first group but the electronic structure of T elements also appears to intervene in this

case. The changes between the ThCr,Sip and CaBe2Ge2 types are less clear. P 3 B 13

Superconductivity The occurence of superconductivity was investigated above 1.4 K in the new gemanides of diamagnetic rare earths. No transition occurs in the

compounds crystallizing with the Yb^Rh.Sn., or ThCr2Si2 type structures or in Lu.Rh-Ge, ; on the contrary, superconductivity is observed in Lalr^Ge, (T =1.5 K) which crystallizes in the same CaBepGe, type structure as

LaIr2Sip *T =1.6 K) (12). Measurements are in progress on the other U,Re-Si, - type germanides and on the BaNiSn- type group. Complete results concerning the conductivity over the 1.4 to 300 K range will be discussed and compared with data on isotypic compounds.

Table 1 Table 2 Germanides and stannides of the Germanides of the Yb-Rh.Sn.-C*) or BaNiSn, (0) types or types

! T Co Rh Ir Ru Os Co Rh Ir Ru Os (8) Ge Sn Ge Sn Ge Sn Ge Sn Ge RE (7) (7) (7) (7)

Sc 7(9) 7(9) 7(9) 7(9) Lu * * * *(6) *(6) A 7 ' 7 7 7 Yb(III) * *(6) *(6) 7,A 7 A 7 Tm * * *(6) *(6) A 7,A 7 A 7 Er * * * *(6) *(6) A 7,A 7 A 7 Ho * * * *(6) *(6) A 7,A 7 A Dy * * *(6) *(6) A 7,A 7 A Yb(III,II) * * Y * * * *(6) *(6) A A 7 A Tb * * * * *(6) *(6) A A 7 A Gd * * * * *(6) *(6) A ACQ) A Eu(III) *(6) (III,II) Sra * * * * * *(6) * *(6) A A A Ce(IV,in; *(6) *(6) Yb(II) * * * A Nd 0 * * * *(6) * *(6) A A A Pr 0 * 0 * * * *(6) * *(6) A A A A Ce(III) 0 * 0 * * * A A k A La 0 * 0 * 0 * 0 * 0 A A<11) A A Eu(II) 0 0 0 A A A P 3 B 13

REFERENCES (1) J.L. Hodeau, J. Chenavas, M. Hare-io, J.P. Remeika, Solid State Commun., 36, 839 (1980) (2) W. Dorrscheidt, H. Schäfer, J. Less. Common Met., 58, 209 (1978) (3) L.G. Akselrud, Ya. P. YarmoLyuk, E.I. Gladyshevskii, Oopov. Akad. Nauk Ukr. RSR, Ser. A, 359 (1978) (4) Z, Ban, M. Sikirica, Acta Cryst., 18_, 594 (1965) (5) B. Eisenmann, N. «ay, W. Müller, H. Schäfer, Z. Naturforsch., 27b, 1155 (1972) (6) C.U. Segre, H.F. Braun, K. Yvon, in Ternary Suparconductors, G.K» Shenoy, B. Dunlap, F.Y. Fradin, eds., Elsevier North Holland, 239 (1981) (7) G.P. Espinosa, A.S. Cooper, H. Barz, Mat. Res. Bull., 1_7, 963 (1982) (8) W.M. Me Call, K.S.V.L. Narasimhan, R.A. Butera, J. Appi. Cryst., 6, 301 (1973) (9) N. Engel, B. Chabot, E. Parthé, J. Less - Common Met., 96, 291 (1983) (10) 0. Rossi, R. Marazza, R. Ferro, J. Less-Common Met., 66, P17 (1979) (11) E.. Hovestreydt, Unpublished results (12) H.p. Braun, N. Engel, E. Parthé, Phys. Rev. B, 28, 1389 (1983)

fig. : Distances d(Co-Ge), d(RE-Ge) and d(Co-RE) versus RE radius for RECoGe, and RE-Co.Ge..» P 3 B 14

THE HIGH-TEMPERATURE MICROHARDNESS OF SOME REFRACTORY MATERIALS AS MEASURED ON SINGLE CRYSTALS

I. Westman, T. Lundstram Institute of Chemistry, Box 531, S-T51 21 Uppsala, Sweden M.M. Korsukova, V.N Gurin, S.P. Nikanorov A.F. Ioffe Physico-Technic&i Institute, Academy of Sciences, Polytechnics! Street 26, 19U021 Leningrad, USSR

Determination of strength properties of "brittle materials c. i be quite difficult in particular at elevated temperatures. Tensile or bending tests require complex apparatus as well as relatively large and specially shaped specimens, which are difficult to prepare. It is much simpler, however, to measure the microhardness, which is fundamentally the resistance of the material to plastic deformation but which is also closely related to other strength properties.

The measurements were carried out on single crystals rather than on polycrystalline materials to avoid influence .?ror. pores and grain boundaries on the inden- tations made. The crystals were prepared using high-temperature solution growth

from an aluminium bath. The following materials have been studied: TaC,'Ta3p, and

The crystals were characterized by chemical and electron microprobe analyses, single crystal and powder X-ray diffraction techniques and scanning electron microscopy. The microhardness measurements were carried out in a Union high- -temperature microscope, equipped with a microhardness attachment. A Vickers squcre-pyramidal diamond or sapphire indenter was used.

Results from the temperature interval room temp. -800°C will be presented and discussed. P 3tf 1 5

TEMPERATURE DEPENDENCE OP THE LATTICE CONSTANTS OP TRANSITION METAL DISILICIDES I. Engstrgm, B.LSnnberg Institute of Chemistry.University of Uppsala, Box 531, S-751 21 Uppsala,Sweden

The metallic nature of the transition metal silicides is the main reason for their use in the field of microelectronics, but other physical properties such as thermal expansion,nucleation and crystal growth are of utmost importance for their applicability from a technological point of view. In the present investigation the thermal expansion of the group IV to VII transition metal disilicides was measured by X-ray methods. The result reveals the directional character of the interatomic forces in the disilicides in relation to the position in the oerio- dic chart of the transition metal and the structure tyne of the di- silicide. The rate of volume expansion of the disilicides reflects broadly the order of their melting points. Thus the expansion is largest for CrSi~,about 5-5 %,and smallest for ReSi~ ,about 2.6 %, in the temperature interval 300 to 1500 K. Comparing disilicides belonging to the TiSi~ family, their thermal properties vary from an almost isotropic behaviour in' VSi- and f to a very anisotropic in CrSip- The thermal expansion of the two disilicides representing the Zp structure type ZrSi2 and HfSi- is very similar. The thermal exoan- sion in one of the crystallographic directions is markedly small for both. The main part of the thermal expansion thus occurres in the directions of the a- and c-axes. The thermal expansion coefficients calculated from the observations made are 8-9 (X10 -K~ ) at room temperature and 11-15 (x10 -K~') at 1300 K. The most extreme expansion coefficient 22 (x1Q -K ) was obtained for the a-axis of CrSi_. P 4 A 1

STRUCTURAL INVESTIGATIONS OF BISMUTHIDES OF TITANIUM, ZIRKOKIUM, AND HAFNIUM. Belga Block and Wolfgang Jeitschko Anorganisch-Chemisches Institut, Universitat Mtinster, D-4400 MOnster, West Germany

The system titanium-bismuth has been investigated before (1-3). We confirm the existence of a titanium-rich compound with a titanium:bismuth ratio near 3:1. The tetragonal compound Ti_Bi is also confirmed. In addition to these two compounds we find a new phase with the approximate composition ~TiBi. Its powder pattern can be indexed on the basis of a tetragonal cell with the lattice constants a = 1O.316(1) S, c = 7.369(1) %, V - 784.2(1) ?3 for the titanium-rich side of the phase.

Previous investigators of the systems zirconium-bismuth (4) and hafnium- -bismuth (5) reported pyrophoric compounds with the approximate compositions — Zr,Bi and ^ZrBi. In addition the compounds ZrBi^ and HfBi_ were found to crystallize with a TiAs. type (6) structure. We confirm the existence of ~Zr3i, ZrBi_, and HfBi_. The new compound Zr_Bi, was found to be isotypic with the hexagonal Mn.Si, structure. Its lattice constants are: a = 8.73O(2) A, c = 5.988(3) 8, V = 395.2(2) £3. We have refined the structure of HfBi. from single-crystal X-ray data to a conventional residual of R = 0.055 for 840 F values and 37 variable parameters. The lattice constants were determined from Guinier powder data: a - 15.63(1) 8, b = 10.144(4) X, c - 3.968(2) 8. Two new compounds with hafnium contents near 50 at % were also observed. We continue the investigations of these systems.

(1) H. Nowotny and J. Pesl, Monatsh. Chem. 82_, 336 (1951). (2) H. Auer-Welsbach, H. Nowotny, and A. Kohl, Monatsh. Chem. 8£, 154 (1958). (3) J. Obinata, Y. Takeuchi, and S. Saikawa, Trans. ASM S2_, 1059 (1960). (4) D. Eberle and K. Schubert, Z. Metallk. 59_, 306 (1968) . (5) F. Hulliger, Nature 204, 991 (1964). (6) S. Wenglowski, G.B. Bokij, and E.A. Pobedimskaja, Z. Strukt. Khim. 5_, 64 (1964). P 4 A 2 MnRuAs, MnRhP : PHYSICAL PROPERTIES AMD MAGNETIC STRUCTURES B. Cheneyier Institut Laue Langevin, 156 X, 38042 Grenoble Cedex, France. M. Anne, M. Bacmann, D. Fruchart Laboratoire de Cristallographie, associé à l'U.S.M.G., C.N.R.S. 166 X, 38042 Grenoble Cedex, France. P. Chaudouët E.R. 155, ENSIEG, B.P. 46, 38402 Saint Martin d'Hères, France.

Crystallographic structures of MnRuAs and MnRhP belong to the Fe^P-type, space group P52m (0 with the following cell parameters at 300 K : MnRuAs MnRhP a = 6.5377(3) A a = 6.2446(4) A C = 3.6267(2) A c = 3.5943(3) A In these structures, metalloid atoms form tetraedrons and square-based pyramids stacked in triangular cross section channels, along the c axis. The 4d atoms are located on tetrahedron sites of a channel and the manganese atoms on the pyramidal sites of the adjacent channel. Magnetic properties

MnRuAs and MnRhP are ferromagnetic with Curie temperatures Tc = 521 and 401 K respectively (2). In high field range (- 20T) at 4.2 K, saturation is reached with fields of about 2-3 Tesla and the magnetic moments per formula unit are 3.94 and 3.06 ug. At room temperature, in both of the compounds, the satura- is not reached even with a field of 20 T. Thermal evolution of the paramagnetic susceptibility of MnRuAs clearly diverges from a Curie-Weiss law even far above T . In the phosphide this discrepancy is less important. The paramagnetic Curie temperatures extrapolated to -= 0 from the highest temperatures (» 1200 K) give the ratios : ~ = 0.76 (MnRuAs) and ß- = 0.86 (MnRhP). p 9P Electrical resistivity The electrical resistivity of the compounds has been studied on sintered samples with a four-probe method. The resistivity variations versus temperature exhibit a metallic behavmur and are very similar. Nevertheless in the MnRhP case the slope of the curve p(T) decreases when getting near Tc = 401 K. This evolution has to be compared with the MnRhAs case at the magnetic order-disorder transition (3). Magnetic structures Magnetic structures have been refined using neutron diffraction patterns obtained at room temperature with a position sensitive multidetector (DN5) in P 4 A 2

the reactor SILOE of C.E.N.G. Magnetic contributions are located on nuclear peaks. Since the number of crystal parameters [2] and magnetic moment carriers [3] is small and the magnetic cell is the same as the chemical one, a refinement profile method (4) was used. The models are uniaxial and most of the magnetism is localized on manganese. Refinements results are reported in tables I and II. Table I : MhRuAs Table II : MnRhP

B(A2) B(A2) w w MX(UB) Mz(uB) Mn(3g) 0.5965(9) 3.24(10) 1.10(24) 1.09(12) Mn(3g) 0.5967(9) 1.34(25) 2.77(10) 0.40(11)

Ru(3f) 0.2573(5J 0.69(7) Rh(3f) 0.2617(8) -0.32(15) 0.52(10)

Asdb) 0 - - 0.88(14) Pdb) 0 - 0.19(15)

As(2c) 1/3 0.33(9) P(2c) 1/3 - 0.73(14)

R(Wp) : 9.30 % R(Wp) : 10.33 %

R(I)NUC : 4.32 % R(DNUC: 4.68% M) : 3.72% m R(DMAG: 4J2% 46 reflections 44 reflections - MnRuAs The magnetic moments are ferromagnetically ordered along a direction located at about 71° from the c axis. Their value at 300 K, 3.42(10)uB is in fair agreement with the magnetization measurements. No significant magnetic moment has been detected on the ruthenium sites. - MnRhP As for MnRuAs, the magnetic structure of MnRhP is mainly based on the order of the moments of the manganese atoms. The structure consists in a stack of ferromagnetic (001) plane.',, with a direction of anisotropy located at 26° from the c axis, the momen* value being 3 08(10)^. In the hypothesis of a possible rhodium polarization (5,6,7), the refinement shows that these ato'.s effectively carry a magnetic moment of 0.32(15)uB parallel to c pointing in the opposite direction to the corresponding manganese "torn component. The resulting value confirms the saturation moment measured under magnetic field (2.6nB at 300 K),

(1) J. Roy-Montreuil, Doctorat d'Etat, Orsay, 1982. (2) P. Chaudouet, These, Institut National Polytechnique de Grenoble, Nov. 1983. (3) B. Chenevier, D. Fruchart, M. Bacmann, J.P. Se"nateur, P. Chaudouet, L. Lundgren, Phys. Stat. Sol. (a), 84, 199 (1984). (4) H.M. Rietveld, J. Appl. Cryst. 2, 35" (1969). (5) J.S. Kouvel, J. Appl. Phys. 37, 7, 1257 (1966). (6) G.G. Low, Proc. Int. Conf. orTMagnetism, Nottingham, 133 (1964). (7) I. relner, I. Nowik, Phys. Rev. Letters 45, 2128 (1980). B. CHENEVIER Institut Laue Langevin, 156 X, 38042 Grenoble Cedex (France). P 4 A 3

MAGNETIC BEHAVIOUR OF toRuP B. Cheneyier Institut Laue Langevin, 156 X, 38042 Grenoble Cedex, France, M. Bacmann, D. Fruchart Laboratoire de Cristallographie, associé â l1U.S.M.G., C.N.R.S., 166 X, 38042 Grenoble Cedex, France. P. Chaudouët E.R. 155, ENSIEG, B.P. 46, 38402 Saint Martin d'Hères, France.

MM'X compounds (where M, M1 = 3d, 4d metal and X = As or P) became in the recent years a subject of intensive investigations (1). Several compounds belonging to the series MnM'X (M1 = Ru, Rh) are being presently studied (2, 3, 4). We therefore report now some physical properties of MnRuP. The structure of MnRuP belongs to the Fe.->P-type, space group P6~2m, with the o £ cell parameters a = 6.257 and c = 3.52, A. Magnetic measurements performed between 4.2 K and room temperature show an unusual magnetic behaviour. The magnetization versus temperature measured under magnetic fields up to 15 T, reported in fig. 1, indicates complex antiferromagnetic orderings below T», = 270 K. Furthermore, the paramagnetic susceptibility measured on a powder sample diverges from a Curie-Weiss law at a temperature T = 3 T„.

130 KO»

120K0»

KO» 20KOB 10 KO« ?KSÎ_ 100 180 200 290 300 TK

Figure 1 : Magnetization versus temperature P 4 A 3

400

TEMPERATURE

10C0 121X1

Figure 2 : Reciprocal susceptibility versus temperature

In order to investigate the magretic structures existing between 4 and 27G K, neutron diffraction experiments were performed using the position sensitive detector DN5 in the reactor SILOE of the C.E.N.G. In the range 170-27C K, we observe a set of satellite reflections related to a temperature dependent propagation vector q^ = [0 q 0] with q = 0.369 at T = 190 K. Between 170 K and 6 K, the diffraction patterns exhibit an additional set of satellite reflections superimposed on the previous one. All the new reflections are indexed in the orthohexagonal cell th q2 = [qx 0 0], qx = 0.401 at T = 160 K. q1 and q2 are orthogonal. In this temperature range q1 remains fixed (q = 0.389) but q2 reaches the value qx = 0.480 at T = 6 K. The magnetic structures determination is in progress.

The electrical resistivity of MnRuP has been studied on sintered samples with a four-probe method. The resistivity variations versus temperature indicate a metallic behaviour and exhibit several anomalies.A strong anomaly is observed near T = 173 K in agreement with the change occuring in the neutron diffraction patterns. Measurements performed around this temperature show the existence of a rather large hysteresis (a 12 K) which suggests a first order transition. At lower temperature p(T) shows another anomaly near 100 K. There is still to explain this electrical behaviour relatively to'th*- thermal variation of the P 4 A 3 modulated magnetic structure.

(1) R. Fruchart, Ann. Chim. fr. 7, 563 (1982) (2) B. Chenevier, D. Fruchart, M7 Bacmann, J.P. Sfinateur, P. Chaudoufit, L. Lundgren. Phys. Stat. Sol. (a) 84, 199 (1984) (3) B. Chenevier, M. Bacmann, D. Fruchart, J.P. SSnateur, R. Fruchart, Phys. Stat. Sol. (a), to be published. (4) B. Chenevier, M. Anne, M. Bacmann, D. Fruchart, P. Chaudoufit, this conference.

M. Bacmann Laboratoire de Cristallographie, associ6 a I'U.S.M.G., C.N.R.S., 166 X, 38042 Grenoble Cedex (France). P 4 A 4

SOME UNUSUAL FEATURES ABOUT THE HAOETIC BEHAVIOUR OF MnRhAs 8. Chenevier Institut Laue Langevin, 156 X, 38042 Grenoble Cedex, France. M. Bacmann, D. Fruchart Laboratoire de Cristallographie, associS a l'U.S.M.G., C.N.R.S., 166 X, 38042 Grenoble Cedex, France.

Magnetic measurements performed on MnRhAs (P62m) indicate a ferromagnetic

behaviour between Tc = 200 and Tt = 160 K. At Tt, MnRhAs undergoes a first order transition and is antiferromagnetic at lower temperature (1, 2). Neutron diffraction experiments show that the moments of the manganese atoms order with magnetic structures consisting of a stack of ferromagnetic (001)

planes. When T < 158 X, Svu and S_ components are both coupled in a ++— sequence. xy z In the high temperature range (165 < T < 200 K), S components are ordered in a Ay +-+- sequence and S components are terromagnetically coupled. The magnetic symmetry changes from P1 to Pm (3). The thermal variation of the magnetic moments of the Mn-atoms, determined from the Bragg reflections of the neutron diffraction patterns, exhibits a discontinuity when the magnetic reordering takes place (fig. 1). Simultaneously a small rotation of the moments relatively to the c axis (AS < 10°) occurs and a weak induced moment (= 0,2 uB/at) appears on the rhodium atoms when T > 158 K. Furthermore, on the neutron diffraction pattern performed at T a 160 K, we obser/e the superimposition of the L.T. and H.T. patterns with two additional satellite reflections related to a propagation vector q * [0 0 1/25]. These satellites reveal the existence of an intermediate long period phase resulting from th* occurence of periodicity defects at the transition. A simple mechanism, involving antiphase-domains, is proposed. When 200 < T < 240 K, we don't observe any magnetic contributions as Bragg reflections on the neutron diffraction patterns but a diffuse scattering can be identified. It implies that the neutrons are not scattered by a perfect three- dimensional arrangement of magnetic moments and suggests the existence of rather large magnetic correlations in this temperature range. This hypothesis is in agreement with the strong induced ferromagnetism observed above Tc (2). For example, at T = 220 K, the extrapolated magnetization per formula (from HJ is about 1 ug ! Moreover recent heat capacity measurements show a broad peak around 240 K which corresponds to the rise of strong magnetic correlations when T decreases (4). Between 160 and 200 K, i.e. when MnRhAs exhibits a ferromagnetic behaviour, a diffuse scattering signal, weaker than the previous one, is still observed on P 4 A 4

the neutron diffraction patterns. In this temperature range, domains of 3-D and short range orderings (low-D magnetic order) could coexist, corresponding to a decrease of some interactions. This would provide an explanation of tin* magnetic moment reduction observed in the H.T. phase as calculated from the Bragg peaks (fig. 1). 1 5 Thermal evolution of the magnetic moment on the Mn-atoms, as deduced from the magnetic part of the Bragg scattering.

ptan(a£),/'c

TOO

(1) H. Bacmann, D. Fruchart, J.P. Senateur, Proc. VII Int. Conf. Solid Compounds of Transition Elements, Grenoble, June 1982. (2) B. Chenevier, 0. Fruchart, M. Bacmann, J.P. SSnateur, P. ChaudouSt, L. Lundgren, Phys. Strt. Sol. (a) 84, 199 (1984). (3) B. Chenevier, M. Bacmann, D. fruchart, J.P. Secateur, R. Fruchart, Phys. Stat. Sol. (a) to be published. (4) J. Garcia, C. Rillo, J. Bartolome, D. Gonzalez, R. Navarro, B. Chenevier, P. ChaudouSt, D. Fruchart, this conference.

M. BACMANN Laboratoire de Cristallographie, associe 4 l'U.S.M.G., C.N.R.S., 166 X, 38042 Grenoble Jedex (France) P4 A5

NEW CHANNEL AND CLUSTER STRUCTURES FORMED IN TERNARY ALKALINE EARTH AND LANTHAN COMPOUNDS G. Cordier, P. Woll and H. Schafer Abt. Anorganische Chemie II der Technischen Hochschule Darmstadt, Eduard-Zintl-Institut, Hochschulstr. 4, D-6100 Darmstadt

Formerly it has been shown, that in many Zintlphases, i.e. compounds of strong electropositive metals with semimetals, clustering of the semimetals occurs depending on the charge of these elements. It is of special interest, that in analogous ternary phases with transition and semimetals these elements together form clusters, too. New examples are:

BagAg4Si42: cubic, S.G. Pm3n (No. 223) a = 1047.8 pm, Z = 1, § = 3.92 g/cm3, NagSi.g structure.

I.a,Ag.Sn4: orthorhombic, S.G. Immm (No. 71) a = 1567.6(5) pm, b = 740.9(3) pm, c = 472.5(2) pm, Z = 2, § =8.01 g/cm , new structure type.

La12Zn_Sb3o: orthorhombic, S.G. Imm2 (No. 44) a = 1946.1(6) pm, b = 1543.0(5) pm, c = 435.1(2) pirt Z = 1, 3 =6.93 g/cm", new structure type. ta12Mn2Sb3Q: orthorhombic, S.G. Imm2 (No. 44) a = 1961.1(6) pm, b = 1537.6(5) pm, c = 431.4(2) = pm, Z = 1, sx 6.93 g/cm , La.,2Zn2Sb30 structure.

The structures are presented. 6' 2 STELLIGE INTELLIGENZ - 8V2 STELLIGE PRÄZISION

Solartron Instruments

DikiiuilmuUicnetcr Moueli 7151 r Modell 70« 1 • b'isteiliges DMM • 8'i Stellen • U. I. R i4-Leiieri. Temp. • Skalenlänge 140.Ü00.000 • Analogausgang für Schreiber oder • Stabilität 1.2 ppmiTag Oszilloskop • 11 ppm/lahr • 500 Speicherplätze für Daten • 10 nV Empfindlichkeit • Echtzeiiuhr • bis 1400 MO • Re-Start nacn Netzausfall • „TRl'E-Ohm • IEEE und RS232 eingebaut • 4-Leiter Messung • RechenDrogramme für Grenzwerte: Ratio; • Datenspeicher für 1500 Meßwerte Statistik. Max..Min.; Skalierung; Offset; • GPIBund RS232C ScMumfauiyew|

IOO MHz OszillcMliop Modell 3224 100 MHz Oszilloskop Modell 5277. • 4 s 100 MHz von 5 mV - 5 V/cm •Je Shot bis 100 MHz • 4-kanaliger alternierender Trigger • 2000 Div/us • 4 Kanäle 'ivertierbar • 5 x 100 MHz von 5 mV - 5 V • 4-kanai;^er xiy Betrieb • 3-kanaliger alternierender Trigger • Cursoren für At Messungen • j Kanals invertierbar • Innenraster mit Beleuchtung • Cursoren für At Messung • Incerne Abfragerouiine • DVM für externe DC bis 1500 V gegen Fehlbedienung • Babysitting Betriebsart • Nichtflüchtiger Tastenfcldspeicher • Nichiflüchiiger Tastenfeldspeicher • 5 lahre Gewihrieismngsfrisi SOLARTRON Scrtlurnb erger

Sdrioabaratr MaBmräta G«Ji.b.H., A-1120 Wian, Mmdiinqar Hauptitrofl« 4«, Telafan 0222/83 56 28, Telax 132598 P 4 A 6

P AND Tc As THE CRYSTAL STRUCTURE OF ^°2 3 2 3 L.H. Dietrich and W. Jeitschko Anorganisch-Chemisches Institut, UniversitSt Munster, D-4400 MQnster, West: Germany.

The previously prepared (1) title compounds crystallize with a new structure type which can be derived from the monoclinic (C2/m) structure of Mo.As, by distortions due t~> differences in metal-metal bonding. "Tiese lower the symmetry to the triclinic soace group PI with the following lattice constants. Tc.P,: a = 6.266(1) 8, b = 6.325(1) 8, c = 7.683(2) 2, a = 95.79(1)°, S = 101.76(1)°, Y = 1O4.34(1)°; Tc ASj: a = 6.574(1) 8, b = 6.632(1) S, c = 8.023(2) 8, a = 95.69(1)°, 8 = 102.03(1)°, y = 104.31(2)° with Z = 4 formula units per cell. The isotypic structures were determined and refined with isotropic thermal parameters from single-crystal X-ray data to conventional residuals of R = 0.049 (1790 F values and 42 variable parameters) for TcoP3 and R = 0.053 (3104 F's and 42 variables) for Tc As .

Fig. 1. Relation of the monoclinic centered Mo-As -type subcell to the triclinic

PI cell of Tc2P and Tc.As,. In the upper left-hand corner the standard C2/m setting for the cell of Mo.As3 is outlined. This is the cell given by Taylor, Calvert and Hunt (2). The structure has been determined and described by Jensen, Kjekshus, and Skansen (3) in the C2/m cell shown in the upper right-hand corner. These authors have also given lattice constants for the alternate 12/m setting. In the lower right-hand corner the relation of the standard C2/m cell of Mo_As, to the ?I ceil of Tc.P and Tc.As, is shown in a view almost perpendicular to the pseudo-2/m axis. P 4 A 6

to the fifth T-T interaction. Thus the metal atoms can relax in their distorted TASg-octahedra to form shorter bonds to the four remaining T neighbors. A further analysis of the bonding situation, shows that two-electron bonds cannot fully account for both structures, but nevertheless most near neighbor interactions will come close to such bonds.

(1) R. Ruhl, W. Jeitschko, and K. Schwochau, J. Solid State Chem. 44, 134 (1982). (2) J.B. Taylor, L.D. Calvert, and M.R. Hunt, Can. J. Chem. 43_, 3045 (1965). (3) P. Jensen, A. Kjekshus, and T. Skansen, Acta Chem. Scand. 2£, 1003 (1966) .

Fig. 3. Metal-metal bonding in Mo_As and Tc-As,. For both structures corresponding sections of the two-dimensionally infinite sheets with metal-metal bonds (in 8 units) are shown. For clarity the considerably distorted octahedral As coordinations of the metal atoms are not shown. Almost all metal-metal bonds occur across common edges of the TAs,-octahedra. The only exceptions are the o horizontal bonds of 3.257 A in the Ho_As~ structure and the corresponding bonds in Tc These bonds are formed across common faces of the octahedra. GLASS-FCBMIHG ABILITY IH PALLADIUM-CONTAINIHG SYSTEMS WITH SOME METALLOIDS M. ^i "er and M. EL-Boragy (x) Max-Planck-Institut fur Metallforschung, Institut fur Werkstoffwissenscharten, Seestr. 75, D-TOOO Stuttgart 1, FRG

In order to sttidy the energetic factors having an influence on glass-forming ability, a systematic investigation of palladium-rich glassy alloys with the metalloids of B groups vat made. The aim of this s'rudy was to search for some relationship between the 'behaviour of the crystalline and non-crystalline phases in the system investigated. For this purpose the usually crystal chemical criteria were used, e.g. average atomic volume, valence electron concentration or average valence electron density. In Figure 1 the glass-forming region of the ternary system Pd-Cu-P is shown (1). To compare the prepared glassy alloys with the phase equilibrium and the types of structures of the stable phases, an isothermal section at 775 K in the system Pd-Cu-P was investigated (2). The section is shown in Figure 2. The metallic glass is formed in two and three phase equilibria region, assuming there are no other phases present at temperatures higher than 775 K (Fig. 2). This observation is in a good agreement with the confusion principle.

• glassy alloys 775 K a gtauy aMoyB •**» enrstaMint ohuny \

A i /%w. \\

\ \ \ f: \ A A. \ • \ , \ \ \. \/ \ > mot* fraction o* I? «p ' Cu3P

Figure 1. Results of the x-ray Figure 2. Isothermal section of the analysis on the splat-cooled phase diagram Pd-Cu-P at 775 K specimens in the system Pd-Cu-P

(x) on leave from the Faculty of Engineering, Suez Canal University, Port Said, Egypt P 4 A 7

A survey of some palladium-containing ternary non-crystalline alloys with 1 2 the elements of the B and B groups as the second and with phosphorus or arsenic as the third component shows that in the interval of valence electron concentration 1.35...1.50 is a glass-forming gap (no single glass phase). Similar results were found in the system Cu-Ag-Mg (3). An analogous behaviour is observed in the palladium-containing systems with two different metalloids. It was found that the glassy alloys Pd. Si, Pd,Ge, Pd, P and Pd^As axe miscible. In all these pseudo ternary glasses, the value of valence electron concentration 1.35 is not crossed. On the other hand, there is no miscibility of the metallic glasses PdiGe and Fd?Se. In this case both glassy compounds have a different characteristic valence electron concentration: For PdiGe, IC < 1.3 and for PdpSe, IC > 1.5. This observation of behaviour of the metallic glasses is analogous to the energetic stability criteria for Cu and CsCl type structure. For the representative glassy alloys the macroscopic density p and the e value of the first x-ray diffraction peak were used to verify the Nagel-Tauc criterion (25C, = K ) (!+). The value of the diameter of the Fermi sphere was calculated from the approximation (T = 0) given by en,

where L = Avogadro number, x. = mole fraction and A. = atomic weight of the i component. The wave number K was calculated from Bragg's law,

where X = wave lenght used. The correlation between 2K_, and K for some palladium-containing metallic glasses is given in Fig. 3.

2KF in For many of the glassy alloys the equation 2K_ = K fits the data very F p well. Some deviation (less 10?) are Figure 3. Correlation between the diameter of the Fermi sphere and the wave observed if the B metal has a large number £ for some palladium-containing atomic volume, or if the concentra- glassy aSloys. P 4 A 7

ion of the metalloid is higher than Xp . = 0.25.

(1) M. Ellner, M. EL-Boragy and B. Predel, phys.stat.sol. (a), 80, 519 (1983) (2) M. El-Boragy, M. Ellner and K. Schubert, Z.MetalUtde. 7£, 302 (198U) (3) U. Mizutani and K. Yoshino, J.Non-Cryst.Solids, 61-62, 1313 (198U) (1;) S.R. Nagel and J. Tauc, Phys. Rev.Letters, 35_, 380 (1975)

M. Ellner, Max-Planck-Institut fur Metallforachung, Seeatr. 75, 7ooo Stuttgart 1 P 4 A 8

Ni,SnP - A NEW COMPOUND MADE IN TIN FLUX H. Fjellvag Max-Planck Institut fflr Festktirperforachuag, D-70O0 Stuttgart 80, BRD S. Furuseth Department of Chemistry, University of Oslo, Blindern, Oslo 3, Norway

The crystal structure of Ni2SnP has been determined from single crystal X-ray data. The structure is discussed in relation to the crystal structures of NiX (NiY_) phases (X_,]f being IV or V main group elements respectively).

Introduction No representative with the MnP type crystal structure is found among the NiY_ pnictides, thus opposing what is generally found for other T_P and TAs phases (T = V,Cr,Mn,Fe Co). However, the only representatives among TX phases (T_ • 3d-elements) crystallizing in the MnP type structure are NiSi and NiGe, see e.g.(l). More structural information on ternary phases (including solid solution series) may help in understanding the mechanisms responsible for the choice of structure type. From a study of various Ni_XY phases, it is here reported on structural data of Ni^SnP. i.— 2

Nil Ni2 Sn p ylrt • • • • y«3/4 A V 0 D

Fig.l. The crystal structure of Ni2SnP (short Ni-Ni distances are outlined) P 4 A 8

Results and discussion Single crystals of Ni-SnP were prepared by crystallization from tin melts containing stoichiometric amounts of nickel and phosphorus. The compound can equally well be synthesized by direct reaction between stoichiometric amounts of the elements in evacuated silica ampoules, at 700 °C.

Ni-SnP crystallizes in the orthorhombic space group Pnma, having the unit cell dimensions a - 12.8260(7) A V - 234.64 A3 b = 3.5943(2) A c - 5 .0896(2) A z = 4

The structure was solved from single crystal X-ray data, and refined co a conventional R-value of 0.030. A projection of the structure along [010] is given in Fig.l; Ni-Ni contacts with interatomic distances 2.80 A and less are outlined.

Fig.2 shows the coordination polyhedra for the four atoms in the asymmetric unit, distorted trigonal prisms for Sn and P, and distorted octahedra for the two crystallographic different Ni atoms (Nil and Ni2). All the indicated interatomic distances fall within the expected range, with the exception of one rather long Ni2-Sn distance (3,009 A) , which in turn implies a heavily distorted octahedral coordination for Ni2. The Ni atoms are additionally coordinated to other Ni-atoms, c_f. the zig-zag Nil-Ni2 chains along [001] in Fig.l. Similarly the Ni2-Ni2 separation (2.801 A) also suggests bonding interactions.

The two different coordination octahedra of Nil and Ni2 are sharing a common face, thereby forming chains along [001], which, in turn, are connected by sharing edges. The preference of the Ni-SnP crystals to form needles with axis parallel to [001] probably reflects strong bonding interactions in this direction. P 4 A 8

2.509

Nil • Ni 2 • Sn

Fig.2. Coordination polyhedra in Ni.SnP (Interatomic distances are given in A )

3y comparing the crystal structure of Ni.SnP with those adopted by tha binary HiX_ and MiY_ phases, clear relationships to £..£. the NiP and MnP type structures (2,3) are found. This is especially evident when concentrating on the metal sublattice, c_f_. Fig.l. In the MnP type structure the two-dimensional pattern produced by Ni-Ni bonds in Ni.SnP are joined into a three dimensional net. In JJi.SnP the large Sn atoms effectively separate the two-dimensional blocks .

(1) Kjekshus, A. and Pearson, W.B. Progress in Solid State Chemistry 1_, 83 (1964) (2) Larsson, E. Arkiv Kemi 23^ nr.32, 335 (1964) (3) Selte, K. and Kjekshus, A. Acta Chem. Scand 1T_, 3195 (1973)

Sigrid Furuseth, Department of Chemistry, University of Oslo, Blindern, Oslo 3, Norway P 4 A 9

NEUTRON DIFFRACTION STUDY ON Ag6Ge1QP12 Helmer Fjellvag , Wolfgang Honle, and Hans Georg von Schnering Max-Planck-Institut f. FestkbVperforschung, Heisenbergstr.1 , D-7000 Stuttgart

As a part of the studies of binary and ternary phosphides, AggGe^P., was synthesized arid its crystal structure characterized by means of single crystal x-ray diffraction (1). The prominent building units of the diamond structure related compound are the Agg octahedra, tetrahedraily capped by four Ge(II) atoms thereby forming units of Ag^Ge with two electron - four center bonds (Fig. 1). The P atoms complete the coordination around Ag (two P atoms) and Ge(II) ( three P atoms), and the cluster units lAggGe^P^] thus formed are interconnected in a bcc arrangement by the remaining six Ge(IV) atoms. 8y assuming all bonds from P to Ag and Ge as two electron bonds, 10 electrons remain for the AggGe* unit, thus leading to four Ge(II) and an Agg + cluster unit. The cluster units [AggGe^-l are isosteric with [RhgfCOhfCO)^] and

[(SCug)Sb-S.-] > however in the latter the center of the Cug octahedron being occupied by an additional sulphur atom. A striking common feature of the crystal structure analyses for [AggGe.P.-lGeg and the mineral tetrahed-

rite [(SCug)Sb^S12]Cu5{2) are large elongated thermal ellipsoids for the octahedra forming atoms. For a better understanding of these phenomena, a single crystal neutron diffraction study on AggGe.QP^ ( Instrument D 9 , ILL, Grenoble, \ = 0.8438 8 ; crystal habit: cube with edges 0.38 cm ) at different temperatures ( 220, 150, and 72 K) was performed..

Fig.1 Ge(ll) P 4 A 9

The preliminary results of the isotropic structure refinement clearly prove an emoty Agg octahedron (as opposed to tetrahedrite) and a large thermal parameter for the Ag atoms (being three times larger than those of Ge and P atoms). Turning to anisotropic temperature factor coefficients did not alter this picture. However, AF syntheses ( isotropic refinement) indicated split positions (xyz) for the silver atoms. Introducing this finding into the isotropic structure refinements, a temperature factor of the same magnitude was obtained for all the atoms (Table 1 ). Further details will be presented at the conference.

Table 1: Ag6Ge1QP12 (Space group I ?3m) Positional and thermal parameters

Method X N N N N Temp. [K] 295 220 150 72 72 Lattice .a. constant1 J 10.322(13) 10.303(5) 10.293(5) 10.284(5) 10.284(5) 0.1946(2) Ag 12 e 0.1954(1) 0.1947(4) 0.1946(3) 0.1946(2) 0.0083(T2) -0.0036(11) B(Ag) 1.90(6) 1.03(6) 0.73(5) 0.38(3) 0.15(5) Gel 8 c 0.2885(2) 0.2889(2) 0.2890(2) 0.2893(1) 0.2893(1) B(Gei) 0.71(5) 0.38(4) 0.26(4) 0.14(2) 0.15(2) Ge2 12 d 1/4; 1/2 ; 0 B(Ge2) 0.51(8) 0.26(3) 0.17(3) 0.11(2) 0.11(2) P 24 g 0.1283(2) 0.1288(2) 0.1289(2) 0.1289(1) 0.1289(1) 0.3605(3) 0.3604(3) 0.3601(2) 0.3600(2) 0.3600(2) S(P) 0.56(12) 0.24(3) 0.15(3) 0.11(2) 0.12(2) N(hkl) 94 444 447 444 444 R(unweight.) 0.040 0.061 0.055 0.042 0.035 B(Ag)/¥(Ge,P) 3.2 3.5 3.8 3.2 1.2

(1) H.G. von Schnering and K.G. HSusier, Rev. Chim. Miner. J13,71(1976) (2) E.Hackovicky and B.Skinner, N.Jb.Miner. Mh. 3,141(1976) P 4 A lo

STRUCTURAL AND MAGNETIC INVESTIGATIONS ON THE ORTHORHCWIC Co D. Fruchart Laboratoire de Cristallographie du C.N.R.S., associ§ a l'U.S.M.G., 166 X, 38042 Grenoble Cedex, France. S. Niziol Academy of Mining and Metallurgy, Al. Mickiewickza 30, Krakow, Pologne. B. Chenevier Institut Laue Langevin, 156 X, 38042 Grenoble Cedex, France. i£ Roudaut Laboratoire de Diffraction Neutronique, D.R.F., C.E.N.G., 85 X, 38041 Grenoble Cedex, France.

1. Introduction

The solid solution between Co2P and Mn2P exhibits a wide range of composition with the CooP structure type (Pnma). In this domain it was observed strong correlations between the degree of metal atom ordering onto the pyramidal and tetrahedral sites, and the cell parameters behaviour. The magnetic phase diagram versus Co/Mn ratio makes evident that the magnetic couplings are also strongly dependent on the degree of substitution :

- first on the P (pyramidal) site Co2P — CoMnP - secondly on the T (tetrahedral) site CoMnP — Mn,P. CoMnP has the highest Curie point T = 589 K of the series of isotype phosphides and arsenides. In a previous study, structural and magnetic properties have been analyzed (1). In spite of the non-observed lowering in crystal symmetry (2), the possibility of a rather complicated non-collinear structure setting on both P and T sites, has been demonstrated. The ferromagnetic behaviour of CoMnP is mainly due to the Mn atoms, but the Co atoms exhibit a non negligible magnetic moment (0.65 uB) not reported in another analysis (3). In the Co-rich part of the phase diagram, the magnetic ordering temperature may be fairly connected to the mean magnetic moment located on the P sites. The magnetic properties appear to be more complicated in the range of composition CoMnP-CoQ ^Mn^ gP. At low temperature the ferromagnetic ordering turns into a metamagnetic behaviour. 2. Structural and Magnetic investigations

Magnetic and neutron diffraction results concerning CoQ gMn1 2P and Co0.5Mn1.5p are here rePorted- At first, the experiments permitted to check up the metal ordering process, governed by phenomenological criteria systematically, reviewed in (4). The main features are : P 4 A lo

for y = 0.2 It exists three magnetic domains, T > 455 K paramagnetic state, 105

TZ = 1/9 and turns into TZ = 3/8 at lowest temperatures. for y = Q.5 At room temperature the compound is paramagnetic, it turns into the conical long wave model when temperature decreases. The wave-vector is temperature independent T = 3/8. At T = 220 K, the material remains ordered, contradicting the 'interpretation of all the previous bulk magnetic measurements (5, 6). The ordering point would be situated slightly above 220 K. Table 1 shows the crystal and magnetic parameters measured on the two samples, at different temperatures. 3. Discussion In the paper on CoMnP (1) we have pointed out that the variations of the Curie

point Tc and the magnetization could be simply expressed in terms of the substitution of Co by Mn on the P sites. Versus the ordered substitution in

(Co1_xMnx)CoP one has :

m = x . [uCo(T) + uMn(P)]

with MMn = 2.55 uB, uCo = p • 0.65 u8, p = 3 if x < 0.3 (polarization phase) and p = 1/x if 0.3 < x < 1.

In Co-| Mn1 P, the range 0 < y < 0.6 ..cannot be analyzed in a so clear cut schema, in order to fit the ordering temperature. However, the change of the easy direction of magnetization (6) is fairly supported by the neutron diffraction results. The easy anisotropy axis corresponding in CoMnP to the major Fmode (along y) turns in the A mode (along z, axis of the conical structure). One has to remark the large similarities of the long wave structure (conical = helix + orthogonal component) with the "double helix" structure-type encountered in binary phosphides and arsenides (7). More recently this type of "double helix" has been discovered in the isotype ternary gentianides and silicides (8, 9). All these series have in common the crystallographic space group Pnma, and the occupancy by the more magnetically active atom (Mh is our case) of the same site 4c (octahedral, or octahedral displaced = pyramidal). But the considerations on the P k A lo

nature and strenght of the exchange integrals do not exactly apply here (7), sinca we have to consider the non negligible P-T interactions with y increasing. If they are negative, these interactions could explain the fast decrease of the ordering temperature with y.

(1) D. Fruchart, C. Martin-Farrugia, A. Rouault, J.P. Se"nateur, Phys. Staic. Sol. (a) 57, 675 (1980) (2) D. Fruchart, M. Bacmann, P. Chaudouet, Acta Cryst. B36, 2759 (1980) (3) H. Fujii, S. Komura, T. Takeda, T. Hokabe1, T. Okamoto, J. Magn. Magn. Mat. 14, 181 (1979) (4) TC Fruchart, Proc. VII CICSET Grenoble, France (1982), published in Ann. Chimie Paris, Masson Ed., 7, 563 (1982) (5) A. Roger, These, Universite" de Paris VI (1970) (6) T. Okamoto, H. Fujii, T. Hihara, K. Eguchi, T. Hokabe, J. Phys. Soc. Jap. 50, 12, 3882 (1981) (7) laee a review paper : A. Kallel, H. Boiler, E.F. Bertaut, J. Phys. Chem. Sol. 35, 1139 (1974) (8) STNiziol, H. Binczyka, A. Szytula, J. Todorovic, R. Fruchart, J.P. Senateur, D. Fruchart, Phys. Stat. Sol. (a) 45, 591 (1978) (9) S. Niziol, A. Bombik, W. Bazela, A. Szytula, D. Fruchart, J. Magn. Magn. Mat. 27, 281 (1982) P 4 A lo

Table 1 : Crystal and Magnetic Structure of Co1 Mn1 P

T(K) y a(A) b(A) c(A) TZ M(P) (Hn) w(T) (Hn/Co) 610 0.2 5.968(2) 3.559(2) 6.768(2) 0 ParaiMgnetic state

300 0u2 5.918(3) 3.522(2) 6.718(4) 0 2.25(5) 0.88(5) CoMnP type 1.01 2.01 0.17 0.41 0.73 0.27 186 0.2 5.9229(7) 3.5201(5) 6.7195(8) 0 2.22(5) 1.04(5) CoMnP type 0.85 1.92 0.68 0.63 0.60 0.57 55.4 0.2 5.967(2) 3-5439(7) 6.7662(9) 0.106(5) 2.30(5) 1.03 2.25 0.48 0.45 0.93 29 0.2 5.969(4) 3-544(3) 6.768(4) 0.105(5) 2-36(5) 0.94(5) 2.31 0.48 0.15 0.93 4.2 0.2 5.969 3.544 6.768 0.372(5) not calculated, not enough data

300 0.5 5.977(3) 3.559(2) 6.799(4) 0 Paramagnetic state 226 0.5 5-969(6) 3.554(4) 6.776(6) 0.37(1) 1-2(2) 0.35(20) 0.85 0.85 -0 0.85 140 0.5 5.969 3.554 6.776 0.345(25) not calculated, not enough data 64.5 0.5 5.9C

In the CoMnP type of structure (T = 0) the moments are distributed on three

components GYFJ\, ; in the conical long period structure, moments are expressed A y L in tii and v.,,.

D. Fruchart Laboratoire de Cristallographie, C.N.R.S., associe h. 1'U.S.M.G., 166 X, 38042 Grenoble Cedex (France) P 4 A 11

PREPARATION PND Ct&SXKL STRDCIDRE OF HoCo^P, ^° ISOKPIC LSHUfflNOID COBALT PHOSPHIDES Ursula Jakubowski and Wolfgang Jeitschko Anorganisch-Chemisches Institut, Univer3itat Munster, D-44OO Munster, West Germany.

The new compound HoCOjP, was prepared by reaction of the elemental components in a tin flux with the atomic ratio Ho:Co:P:Sn * 1:4:1:30. After annealing for 7 days at 850 °C the tin-rich matrix was dissolved in cold slightly diluted (1:1) hydrochloric acid. This treatment does not attack the shiny black needles of the compound.

The crystal structure of HoCo.P- was determined from single-crystal X-ray data by direct methods and subsequent difference Fourier syntheses. It is orthorhombic with the space group Pmmn-D-. • The lattice constants- are listed in Table 1 together with other isotypic compounds. There are Z = 6 formula units in the cell. Full-matrix least-squares refinements resulted in a residual of R = O.O55 for 2441 unique structure factors and 50 variable parameters.

The HoCo^P. type belongs to a large family of structures of which the Fe_P and TiNiSi type structures may be considered as the most simple representatives. The phosphorus atoms are all in a trigonal prismatic environment of metal atoms which is augmented by three additional metal atoms outside the rectangular faces of the prism. There are six different sites for the cobalt atoms. Four of these cobalt atoms have the coordination number 12 with 8 metal and 4 phosphorus atoms. The latter form (somewhat distorted) tetrahedra. One cobalt atom has 15-fold coordination (10 metal and 5 phosphorus atoms). The five phosphorus atoms form a square pyramid. These coordination polyhedra also occur in the Fe_P and TiNiSi type structures for the P and Si atoms. The coordination of one cobalt atom is somewhat unusual, but occurs also in the structure of Sc-Co.gP.. (Reinbold and Jeitschko, this conference). It consists of six Ho atoms forming a trigonal prism and three phosphorus atoms outside file rectangular faces of the prism. The two different Ho atoms have 18-fold coordination including six P atoms which form a trigonal prism.

Table 1. Lattice constants <2> of compounds with the orthorhombic HoCo,p, type structure. Standard deviations in the last significant digits are given in parentheses. Compound a b c V(X3) TbCo,P, 10.586(4) 3.706(2) 12.267(4) 481.3 OyCo,P, 10.570(3) 3.698(2) 12.250(3) 478.9 HoCo,P, 10.560(2) 3.683(1) 12.233(2) 476.4 ErCo,?, 10.546(2) 3.678(1) 12.201(2) 473.2 TaCo,P, 10.531(3) 3.666(1) 12.174(5) 470.0 XbCo,P. 10.514(2) 3.657(1) 12.155(3) 467.3 LuCo3P. 10.524(2) 3.659(1) 12.164(3) 463.4 P 4 A 12

GROWTH AND CHARACTERIZATION OF LOW RESISTIVE CDSIASa SINGLE CRYSTALS M^g^W... Kiaiiel, M.Ch. Lux-Steiner, M. Obergfell and E. Bucher Fekultaet fuer Physik, Universitaet Ronstanz, Postfach 5560, 7750 Konstanz, Fed. Rep. of Germany

Cadmiumsilicondiarsenide is a p-type semiconductor with several interesting properties suggesting it as a potential candidate for solar cell and photo cathode applications: the direct band gap (1.47eV) is almost ideally suited for photovoltaic energy conversion (1,2); bandstructure and p-type conductivity as well as the crystal field splitting and spin orbit coupling suggest the possibility of achieving 100* spin polarized electrons from a (001) surface, when suitably irradiated by photons (3); and the components are low cost materials. For both applications it is necessary to use low resistive crystals, that is why this work deals with a) the reproducible growth of large and low resistive CdSiAs2 single crystals with mirroi—like faces and b) the influence of the transport agent on the electrical properties. a) For the reproducible growth of large (up to 1.5 cm length) single crystals a modified vapour phase transport technique was

fig.l Single crystals of CdSiAaz grown by Br2 (scale 1 an) P 4 A 12

used within a closed-tube system (4). Lumps of polycrystalline CdSiA32 material which was first synthesized fron the pure elements was exposed stationarily together with a transport agent to a snail temperature gradient below the point of peritectic decomposition (around 350 °C ). During several experimental runs it was noticed that the growing process was very sensitive to the growth parameters: these were (i) the temperature profile along the growth ampoule, (ii) the kind and amount of the transport agent, (iii) the homogenity of the prereacted polycrystalline source material, (iv) the size of the lumps and the arrangement within the growth ampoule and (v) the size of the ampoule itself. Some typical crystals grown by Br2 are represented in fig.l. Crystals grown by CdCl2xHaO are shown in fig.2. The habit of these crystals is usually prismatic elongated, the growth direction is [111], the reflecting faces belonged to the forms {101},{112} and (112}. b) Ohmic contacts could be obtained by careful surface cleaning, etching and evaporating gold spots. For electrical characterization we determined the resistivity and the hall mobility between room temperature and 160 K by the van der Pauw method. Crystals grown by Brz showed resistivities in the range of 100-300 ohmcm, whereas those grown by CdClaxHzO showed values from i-10 ohmcm. From the slope of the resistivity in fig.3 the activation energy and net free hole density are calculated. Assuming a non-compensated semiconductor, it yields activation energies of 490 meV and 260 meV for sample 3A and 3B

fig.2 Single crystals of CdSiAs2 as grown by within the ampoule (scale 1 mm) P 4 A 12

, respectively. These values are in good agreement with the data of other authors, assuming a saturated Cd/As growth atmosphere in our ampoules (5). Free hole densities of 6-10l* cm~3 and 6-1015 cr3 are determined for sample 3A and 3B at room temperature , respectively. The hall mobility varies from 150-250 cm2/Vs at room temperature for differently grown samples. Between 160 and 300 E the hall mobility of sample 3B increases as T~1-2.

ia = 3B

r • a2 =

10 is

1000 10 /K"'/ 2 I • S S 7 ' fig.3 Specific resistivity of sample 3A (grown by Brz) and 3B (grown by CdCl2xH20) from 160 K to 300 K (1) Ch. Dovletmuradov, K. Ovezov, V.D. Prochukhan, Yu.V. Rud and M. Serginov, Sov. Phys. Semicond. 10 (1976) 986. (2) A.F. Caroll, C.F.Smith, L.C. Burton and L.H. Slack Rev. Electrochem. Soc. 128 (1981) 8 . (3) F. Meier, ETH Zuerich, private communication. (4) M. Avirovic, M.Ch. Lux-Steiner, U. Elrod, J. Hoenigschmid and E. Bucher, J. of Crystal Growth 67 (1984) 185. (5) G.K. Averkieva, N.A. Goryunova, V.D. Prochukhan, Yu.V. Rud and M. Serginov, phys. stat. sol. (a) 5 (1971) 571.

M.H.W. Kimmel Universitat Konstanz Fakuitat fur Physik Postfach 5560 7750 Konstanz P 4 A 13

STRUCTURAL CHEMISTRY AND PHYSICAL PROPERTIES OF TERNARY PNICTIDES

RE Rb2 Zz (RE - RARE EARTH, X - P, As)

R. MADAR, P. CHAUDOUET, D. BOURSIER, J.P. SENATEUR

ER. 155. CNRS. ENSIEG. Domalne Unlversltalre BP 46, 38402 Saint Martin d'HSres, France,

B. LAMBERT

Laboratoire da Cristallographie du CNRS 25, a/enue des Martyrs, 166 X, 38042 Grenoble Cedex, France.

Ternary phosphides and arsenides RE Rh2 P2 and RE Rti2 AS2 have been synthetized from the elements. We have been able to prepare these compounds for RE : La, Ce, Pr, Nd. Our attempts to synthetlze these alloys with the other rare earth were unsuccessful. All these compounds were found to be isotypic and to cristallize with

the primitive tetragonal structure of CaBe2Ge2 which in an ordered derivative of the BaAl^ type (1). This is a suprising result since most of Che 400 known

ternary compounds with the general composition MT2 X2, where M Is ar alkaline earth or rare earth metal, T is a transition or post transition element, and X * B, Ga, Si, Ge, P, As or Sb cristallize with the body centered tetragonal ThCrjSij type structure (2) (another derivative of the BaAl^ type). Moreover all other ternary phosphides and arsenides with this general composition, reported up to date (3), belong to the ThCrjS^ type of structure.

We have listed in the table 1 the lattice constants of the RE Rh2 X2 phases and given in the table 2 an example of the evaluation of the powder pattern of La Rh2 As2« Since we are Interested in the relation between structural features and the occurence of superconductivity, the superconducting behavior of these alloys was studied at low temperature. No superconductivity was found for temperature as low as 1.5 K. P 4 A 13

The magnetic behavior of these alloys was studied In the temperature raage 300 < 1 < 1250 K (figure 1). For RE - Pr and Nd the magnetic susceptibilities closely follow a typical Van Vleclc paramagnetism of free RE3+ ion. In the case of Ce Rh£ ^ the interpretation is more complex. No ferromagnetic ordering was encountered in these alloys at 4.2 K.

References : 1. H.F. BRAUN, N. ENGEL and E. PARTHE Phys. Rev. B, _28_, 1389, (1983) 2. E. PARTHE and B. CHABOT In K.A. GSCHNEIDNER Jr., L. EYRING (Eds), Handbook on the Physics and Chemistry of Rare Earths, Vol. 6, North Holland, Amsterdam (1983) 3. R. MARCHAND and W. JEI'XSCHKO J. Solid. Stat. Chem.,_24_, 351, (1978) 4. A. MEWIS 2. Naturforsch., 35 b, 141, (1980)

S f J

EN XE.VIN 3CO 4C9 acs TED 11=0 1220 :xa

Figure 1

R. MADAR. ER 155, CNRS, ENSIEG, Doaaine Universicaire, BP 46, 36402 Saint- Martin-d'Hdres, France. P 4 A 13

Table 1 : Lattice constants of the RE IUi2x2

R.E HE Rh2 P2 HE Rh2 As2 a(A) c(A) a(A) c(A)

La 4.189(6) 9.577(4) 4.315(0) 9.862(0)

Ce 4.157(i) 9.501(9) 4.233(2) 9.850(2)

Pr 4.156(1) 9.513(2) 4.266(3) 9.856(6)

Nd 4.142(6) 9.488(9) 4.240(6) 9.859(1)

Table 2 : Comparison of calculated and observed both d-spacings and Intensities for La Rh

hkl dobs(A) dcalc(A) rcalc ^bs 002 4.931 4.931 0.2 w 101 3.952 3,953 0.02 vw 110 3.051 3.051 14.4 m 111 2.910 2.914 1.0 w 103 2.612 2.614 70.0 vs 112 2.594 2.594 89.0 ws 004 2.465 2.465 3.0 w 200 2.157 2.157 100.0 ws 202 1.976 1.976 0.4 vw 114 1.916 1.917 33.0 s 211 - 1.893 0.2 - 105 1.792 1.793 52.0 s 213 1.664 1.664 92.0 vs 006 - 1.643 0.2 - 204 1.622 1.623 10.0 w 220 1.524 1.525 70.0 s 222 - 1.457 0.3 — 116 1.446 1.447 51.0 s 301 - 1.423 0.2 - 215 1.378 1.379 74.0 vs P 4 A 14

TERNARY BARE EARTH COPPER AND SILVER PNICTIDES WITH

ThCr2Si2-RELATED STRUCTURES

M.H. waller and W. Jeitschko

Anorganisch-Chemisches Institut, Corrensstrafie 36, Universitat MOnster, D-4400 Munster, West Germany

Several new ternary compounds RTxPny with R = rare earth metals, T = Cu or Ag, and Pn » P, As or Sb were prepared by reaction of the elemental components under vacuum in alumina crucibles sealed in quartz ampoules at temperatures of about 9OO C. The resulting products are black or grey, sometimes with a metallic appearence; they are stable in dry air. Single crystals suitable for structure determinations were isolated from the solidified melts or from sintered lumps.

The X-ray powder patterns of all compounds can be indexed with tetragonal subcells corresponding to the ThCr2Si2~type structure. The true symmetry of some compounds which were investigated by single crystal X-ray methods was found to be orthorhombic with lattice constants a "WO 8, b •x* c <\» 5.5 8. Single crystal structure determinations and refinements were carried out for. LaCui+xP^2

{R = 0.057), CeCu1+xP^2 (R = 0.024), and LaAg^P,^ (R = 0.055) . They are isotypic, the space group is Cmmo with Z =» 8. The structure parameters of the cerium compound which was investigated most thoroughly are given in Table 2.

The structure is very similar to the structures of SrZnBi2 (1) and CaBe_Ge_ (2). The main differences to these structures arise by the partial occupation of the positions of the Cu(2) and the P(3) atoms which are located in layers at x = 0 and 1/2. In these layers the P(3) atoms form 4-membered planar rings which are connected to each other by Cu(2) atoms. The Cu(2) atoms are thus coordinated tetragonal-pyramidal by P atoms, while the Cu(l) atoms have a tetrahedral coordination formed by P(l) and P(2) atoms. The P(3) and Cu(2) atoms give rise to the superstructure: the P(3) atoms are considerably displaced from their positions in the tetragonal subcell, and the Cu(2) atoms are no longer on 4-fold positions. In addition the positions of the Cud) atoms are also not fully occupied. The refinement of the occupation factors of all copper and phosphorus atoms resulted in the composition CeCUj i2Pi 97'

All compounds were identified by their X-ray powder patterns. Some of them were found to be isotypic with CeC^ 12P1 g?, others could be identified as listed in Table 1. However, most of them could not be ascribed to one of the so far known structure typas of the ThCr2Si2-PbFCl family. Especially interesting is the occurrence of high temperature modifications of the compounds CeCu. P^ P 4 A 14

and LaAg P . Their powder patterns could be indexed on the basis of a L ' X ""* tetragonal cell.

Table 1. New ternary compounds RT Pn with ThCr,Si_-related structures

! cu p Ag P Cu AS Ag As Ag Sb x y ^*x y x y ^*x y ^x y

Y B E F

La A A,D E A F

Ce ArD A E A F

- Pr A A E C F

Nd A E c F

Sm B E c F

Gd S E c F

Tb B E c F Dy B E F

Ho B E F

Er 3 E F

Tm 3 E F

Yb E

Lu E

p A . .. orthorhombic CeCu. i2 i 97 structure with a-2oS, b »c « 5,5 S 3 . . orthorhombic structure with a - 20 S, b - c - 5.5 8, but not isotypic wlthCeCVl2P1.97 C . .. orthorhombic SrZnSb. type structure (3) with a — 20 A, b — c - 4 S D . ,. tetragonal structure with a — 4 X, c - 20 S E . ,. tetragonal structure with a - 4 8, c - 20 2, but not isotypic with D F . . tetragonal structure with a - 4 S, c - 1O or 20 A P 4 A 14

Table 2. Atomic parameters of CeCu. .12P1.97 2 Atom Cmmm Occ. X y z B[8 ] Ce(l) 4h 1 0 .11690(2) o 1/2 O.736{5) Ce(2) 4g 1 0 .37867(2) 0 a 0.641(5) Cud) 8m 0.98(1) 1/4 - 1/4 0.2522(2) 0.777(7) Cu(2) 4h 0.31(1) o.4543(3) 0 1/2 2.33(9) P(l) 4h 0.98(2) o.3240(1) 0 1/2 0.62(2) P(2) 4g O.99(2) 0 .1771(1) 0 0 0.57(2) P(3) 8n O.98(3) 0 0.2262(6) 0.2044(3) 2.43(4)

R = 0.024, 38 variable parameters, 821 F -values. B: isotropic equivalent thermal parameter from anisotropically refined atoms. Lattice constants: a = 19.649(5) 2, b = 5.550(1) S, c = 5.522(1)'8

(1) G. Cordier, B. Eisenmann, H. Schafer, Z. Anorg. Allg. Chem. 426, 2O5 (1976). (2) B. Eisenmann, N. May, W. Muller, H. Schafer, Z. Naturforsch. 27b, 1155 (1972). (3) E. Brechtel, G. Cordier, H. SchSfer, Z. Naturforsch. 34b, 251 (1979). f A

NEW R2RhuX7 COMPOUNDS STRUCTURALLY RELATED TO Zr Fe P? TYPE J.Y. Pivan, R. Gue"rin and M. Sergent University de Renn.es I, Laboratoire de Chimie Mine rale B, L.A. 254 du C.N.R.S., Avenue du General Leclerc, 35042 Rennes Cedex (France)

We report the results of X-ray investigations of a new family of ternary compounds R Rh X? (R = RE, Zr, Y ; X = P, As). Our previous work on binary system Rh-As reported on new arsenides Rh,As_ and Rh As. (0 sx< 1.5) (1, 2), both structurally related to 1 b ' 7 iX * Cr12P7-type (3). The filling-up of trigonal prismatic arsenic sites, unoccupied in Rh .As., by elements as rare-earth, yttrium, zirconium, led to the new ternary arsenides mentionned above. Isotypic ternary phosphides were also synthesized. Lattice constants are given in Table I. Crystal structures of

Rh As,, Rh As. (x = 1.5), Ho Rh As, and Zr Rh P7 were solved. Table I

1 3 (A) c (A) c/a

219 0 Rh!0.5A37 9 (2) 3 532 (1) , 383 RK!2AS? 9, 29? (1) 3 657 (1) 0 .393 3, 0 Y2Rh!2As7 9, 883 (3) 869 (4) 391 Ce2Rh.2As7 9, 897 (3) 3, 916 (2) 0 396

Dy2Rh12A,7 9, 894 (3) 3, 866 (2) 0 390 Ho Rh As^ 9, 892 (3) 3, 859 (2) 0 390-

Er2RhuA.? 9, 890 (3) 3, 358 (3) 0 390

2r,Rh12P7 9. 516 (I) 3, 773 (1; 0 396

Y2Rh]2P7 9, 619 (3) 3, 792 (2) 0 394

.Vd2RhuP? 9, 600 (3) 3, 820 (2) 0, 398

Cd2Rh12P7 9. 626 (4) 3. 790 (2) 0, 393

Tb2RhuP7 9, 624 (4) 3, 790 (2) 0. 394 D ah P 9, 595 (3) 3, 790 (2) 0, 395 y2 12 7 Ho2Rh12P? 9. 587 (3) 3. 789 (2) 0, 395

Er2Rh12P? 9. 568 (3) 3, 780 (2) 0. 395

Yb2Rh]2P7 9. 582 (2) 3, 792 (2) 0, 395 P 4 A 15

The structures of new ternary compounds (Fig. 1) are closely related to that of Zr Fe ,P. which is a filled-up Cr _P_-type structure (4, 5, 6, 7).

Fig. 1 Projections onto the (001) plane of Ho L

Fig. Z Location of the metalloi'd atom along the c axis.

O Rh.Fe.Cr o As.P However, they differ from Zr Fe ,P_-type by the location of non-metal atom along the c axia ; this atom, which appears at the same height as near-neighbour metal atom3 in Zr.Fe. .P.,-structure, occupies a different height in

Ho_Rh As, or Zr Rh ,P7 structures (Fig. 2). This result modifies the coordination of this atom which is now in a distorted or regular metal octahedron, instead of a tricapped trigonal prism. Surrounding metal atoms are also disturbed by this phenomenon : iron atoms occupy alternately |FeP | tetrahedra and | FePJ pyramids in Zr Fe P_, contrarywise to that observed in Ho Rh As. or Zr Rh P. where rhodium atoms are tetrahedrally coordinated. The location of metalloid atom along the c axis was also previously discussed for binary compounds Cr P_ and Rh As. (2). First physical measurements performed on ternary compound

Ho_Rh _P7 exhibit a metallic conductivity and an antiferromagnetic order (T^. = 2.7 K) at low temperature. Upon 80 K, a Curie-Weiss paramagnetic behavior appears. The experimental value of magnetic moment is in agreement with that expected for Ho 3 free ion. Magnetic and electrical measurements at low temperature for all compounds of this family are in progress.

(1) J.Y. Pivan, R. Guerin, M. Sergent, C.R. Acad. Sci. ^er:.e n, 9_, 533 (1984) (2) J.Y. Pivan, R. Guerin, M. Sergent, J. Less Common Met. (to be published) (3) H.K. Chun, G.B. Carpenter, Acta Cryst. B35, 30 (1979) (4) E. Ganglberger, Monatsh. Chem. 99_, 557 (1968) (5) "W. Jeitschko, B. Jaberg, Z. Anorg. Allgem. Chem. 467_, 95 (1980) (6) R. Guerin, M. Potel, M. Sergent, J. Less Common Met. 78_, 177 (1981) (7) S. Maaref, R. Madar, P. Chaudouet, J.P. Senateur, R. Fruchart, J. Solid State Chem. 40, 131 (1981).

R. Gu&rin, Laboratoire de Chimie Mine rale B, Campus de BeauJieu, ,, Avenue du General Lederc, 35042 Rennes Cedex (France) 4 A 16

STRUCTURAL CLASSIFICATION IN TERNARY PNICTIDES CHEMISTRY j.Y. Pivan, R. Gue"rin and M. Sergent University de Rennes I, Laboratoire de Chimie Mine"rale B, L.A. n' 254 du C.N.R.S., Avenue du General Lederc, 35042 Rennes Ce"dex (France)

Our study on ternary system Ho-Ni-P revealed three new compounds Ho Ni, P,_, Ho.Ni.-P,, and Ho, Ni, ,P ., in addition to the compounds 3= 19n U 6Z013 ZOn 00 43

HoNi,P2 (ThCr2Si2-type) and Ho2Ni12P7 (Z^Fe^P^type) previously reported (1, 2). These new phosphides with the same metal/metalloi'd ratio of 2:1 as

Ho_Ni1,P7 crystallize in an hexagonal unit-cell with one formula unit per cell. The structural studies as well as refinement of Ho Ni P_ were performed on r eedle-shaped single-crystals and led to the following results given in Table I. Table I

H0 Ni P Ho Ni P Ho^Ni-.F „ 2 12 7 5 19 12 H°6Ni20P13 20 66 43

a (A) 9.063 (1) 12.298 (2) 12.676 (4) 23.095 (7) c(A) 3.673 (1) 3.762 (2) 3.730 (2) 3.742 (4) V (A3) 261.2 492.0 519.1 1728.6

space group P6 P62m P63/m P63/m

- Ho,Ni,nPn, : The structure derives from Hf,Co .P,-type (3) by substitution S 17 1^ 2 4 3 of hafnium atoms in pyramidal phosphorus sites by nickel atoms.

- Ho£Ni20P13 : Jt is of Zr6Ni20P13*type' WC have Previously reported (4) and can be viewed as a filled-up Rh_ Sit -type structure (5)

2Q6643 ^3 compound is a new structure-type, phosphorus polyhedra are the same as in Ho.Ni P and Ho,Ni P , but the number is more 5 17 12 0 7.0 13 important. All these structures, included Ho.Ni.-F.., exhibit a trigonal prismatic metal coordination of the phosphorus atoms as observed in numerous structures with same metal/rnetallofd ratio (3 - 10). P 4 A 16

Moreover, in agreement with unit-cell variations (Fig. 1), the three H Ni P can b structures Ho Ni P?» Ho^Ni P13 and °2Q 66 43 * considered as belonging to the same family of structures of which Ni.P (Fe F-type) ; in fact, Ni.P may be regarded as the most simple representative. Indeed, a general o 3 formula of the members of this series can be written : A , , ,B, ,,, ,,C , ,> , which must be modified as follows : n(n-l) (n+l)(n+2)' n(n+l)+l (A,B) . ,. , , ,.C , .» . when n increases. For example, the second v 'n(n-l)+(n+l)(n+2w ) n(n+l)+l formula has to be applied when n = 6. Other terms of the series are actually in study. Such a classification was tried by Engstrom for binary compounds Fe P, Th_S._ (anti Cr P_) and Rh-.Si , (5), but the formula he gave: B , ,C , - was not correct. The most simple description of the members of this series is to consider that they are built up, in projection on the (001) plane, by triangular metal composite units (Fig. 2). Each composite unit is in fact subdivided into n2 simple triangles occupied either by C atoms (n = 1), B and C atoms (n = 2, 3) or A, B, C atoms (n = 6). From our work on Ho-Ni-P system, structural relationship in pnictide chemistry between Fe P, Z.r_Fe P_, Zr.Ni P and Ho Ni, ,P types is established. In the same way, this relationship can be extended to binary structures Cr P.. and Rh Si as shown below.

JVttf 7 ' C6Rh20Si13 FezP

(1) W. Jeitschko, B. Jaberg, J. Solid State Chem. 3,£, 312 (1980) (2) W. Jeitschko, B. Jaberg, Z. Anorg. AUg. Chem. 46J_, 95 (1980) (3) E. Ganglberger, Monatsh. Chem. 99., 566 (1968) (4) R. Guerin, E.H. El Ghadraoui, J.Y. Pivan, J. Padiou, M. Sergent, Mat. Res. Bull. 22, 1257 (1984) (5) I. Engstrom, Acta Chem. Scand. 1%, 1924 (1965) (6) S. Rundqvist, F. Jellinek, Acta Chem. Scand. H., 425 (195 9) (7) E. Ganglberger, Monatsh. Chem. 99., 557 (1968) (8) V.N. Davydov, Yu. B. Kuz'ma, Dopov. Akad. Nauk. Ukr. R.S.R., ser. A, i, 82 (1981) (9) L.G. Aksel'rud, V.I. Yarovets, O.I. Bodak, Ya. P. Yarmolyuk, E.I. Gladyshevskii, Kristallografiya, 21., 383 (1976) A 16

(10) Ya. P. Yarmolyuk. L.G. Aksel'rud, E.I. Gladyshevskii, Kristallografiya 23, 942 (1978)

0 (A)

Fig. 1

20- Unit cell parameters versus n for C(A>

Ni2P (n = 1)

H°2Ni12P7 (n = 2) H°6Ni20P13 (n = 3> H°20Ni66P43 (n * 6) / 15

t 2 3 4 5 S 7 3

Fig. 2 Projections on the (001) plane of

(a) Ni&P3 (n = 1)

(b)Ho 2Ni12P? (n = 2) (c)Ho 6Ni20Pl 3 (n = 3) (d) Ho L = ( 20Ni66P 43 (B (d) large circles ::Ho smf.i:. circles : P P 4 A 17

STRUCTURAL CHEMISTRY OP TERNARY SCANDIUM COBALT PHOSPHIDES Bike J. Reinbold and Wolfgang Jeitschko Anorganisch-Chemisches Institut, Oniversitit Munster, 0-4400 Munster, West Germany.

Several new scandium cobalt phosphides were prepared by reaction of the elemental components in a tin flux. ScCoP has a TiNiSi type structure (1) with the orthorhombic lattice constants a » 6.268(2} 8, b = 3.75O(l) 8, 3 c = 7.050(3) 8, V = 165.7 8 . ScCo^ is isotypic with YCb5?3 (2). Its orthorhombic lattice constants are: a = 11.691(5) 8, b = 3.533(1) 8, c = 10.206(3) 8,

V = 427.6 8 . The compound Sc.Co P7 has the hexagonal Zr_Fe P? structure (3) with a = 8.973(3) S, c = 3.531(1) 8, c/a = O.3935, V -- 246.2 S3. Sc.Co .P.. has a new structural type which is closely related to the Hf-Co.P, type (4). It has hexagonal symmetry with the space group P62m-D, and the lattice constants a = 12.123(7) 8, c = 3.633(2) 8, c/a = 0.2997, V = 462.4 8 , with one formula unit per cell. The structure was determined and refined with isotropic thermal parameters from single-crystal counter data to a residual of R = 0.026 for 848 unique structure factors and 25 variable parameters .

The structure of SccCo,-P,. can be derived from the HfoCo,,P, (= Hf0Co,rP,_) structure by substitution of the three hafnium atoms of site 3f with cobalt atoms. This site has the coordination number 15 and apparently is too small for the large scandium atoms which occupy two sites with the coordination number 18. The other cobalt atoms have even smaller coordination numbers. Three of them occupy sites with 12-fold coordination and one a site with only 9-fold coordination (6 Sc + 3 P). Sc_Co -P.. belongs to a large family of structures where the phosphorus atoms have trigonal prismatic coordination which is augmented by thr=s additional neighbors outside the rectangular faces of the prism.

(1) C.B. Shoemaker and D.P. Shoemaker, Acta Crystallogr. _18_, 900 (1965) . 1'2) U. Meisen and W. Jeitschko, J. Less-Common Met. 1O2, 127 (1984). (3) E. Ganglberger, Monatsh. Chea. 99, 557 (1968). (4) E. Ganglberger, Monatsh. Chem. 99_, 566 (1968). P 4 B 1

PREPARATION AND STRUCTURE REFINEMENTS OF TERNARY AND QUATERNARY

RHENIUM PNICTIDES WITH RegP13 STRUCTURE

R. Rûhl Berufsgenossenschaft der Keramischen und Glas-Industrie,. D-87OO Würzburg W. Jeitschko Universität Münster, Anorganisch-Chemisches Institut, D-44OO Münster

The rhombohedral polyphosphide RögP.^ (1) with a cell volume for the hexagonal cell of V » 1768 A forms a continuous solid solution Re.As P.-, up to V = _ O X 1J~X 1981 % (2,2). We have refined the structure of one of these compositions (V = 1958 A ), and of a quaternary compound with a small amount of antimony (V = 1981 A ). The structure refinements led to the compositions Re^As- ,P, . o y.b j . 4 (R = 0.052 for 1833 F-values and 40 variable parameters) and Re^Sb 7Asg IPT O (R = O.O67 for 1156 F-values and 32 variable parameters). The structure refinements of the ternary and quaternary compounds are of interest because the Re,P,, structure has five different sites for the P atoms o l J The first P atoms which are substituted by As or Sb atoms are those with the greatest distances to their neighbors. On the other hand, the position of the P atoms which form 6-membered rings with the very small P-P distances of 2.142 S in Re.P,,, is the position with the highest P content in the newly refined structures, b 1J The average Re-Re distances in the rhombohedral Re. clusters do not change very much with the substitution of the P atoms by As or Sb atoms. They are all slightly smaller than the average Re-Re distance of 2.899 A in the rhombohedral Re groups of Re2P5 (4): Re6P13 - 2.853 A\ Re6Asg ßP3 4 - 2.859 X, Re6SbQ 7Asg ^ 2 - 2.890 X. The ternary rhenium arsenophosphides and the quaternary rhenium antimonyarseno- phosphide have been prepared with iodine as a mineralizer. The samples were all annealed at temperatures between 9OO and 950 C in evacuated silica tubes, usually with a great excess of arsenic and antimony (for example: the starting composition was "ReAs.g ,PQ 7" for the compound with V » 1981 A and the extrapolated composition RecAs. _P, ,). No binary rhenium polyarsenide was found besides

Re3As7 (5,6) . The structural chemistry of these compounds will be discussed together with those of closely related other compounds (3). (1) R. Rûhl and W. Jeitschko, Z. Anorg. Allg. Chem. 466_, 171 (198O). (2) R. Rühl, 0. Flörke and W. Jeitschko, J. Solid State Chem. 53_, 55 (1984) . (3) R. Rühl, Dissertation, Giessen (1982). (4) R. Rûhl and W. Jeitschko, Inorg. Chem. 2_1_, 1886 (1982) . (5) S. Furuseth and A. Kjekshus, Acta Chem. Scand. 20, 245 (1966). (6) M. Klein and H.G. v. Schnering, J. Less-Common Met. U_, 298 (1966). PREPARATION AND CRYSTAL STRUCTURES OF THE POLYPHOSPHIDES TiMn,Pl2 AND NbFe,Pl2 Udo D. Scholz and Wolfgang Jeitachko Anorganisch-Chemisches Institut, Univer3itat MUnster, D-4400 MUnster, West Germany

Single crystal3 of the new compounds TiMn,P12 and NbFe,Pn were prepared by reaction of the elemental components in presence of iodine and in a tin flux respectively.. The crystal structure of TiMhIP1j was determined by Patterson and difference Fourier methods from single-crystal diffractometer data. It crystallizes in the monoclinic space group C2/c with four formula units per cell and the lattice constants: a = 16.059(3) A, b = 5.794(1) A, c = 10.635(2) A. g = 115.22(1)°, V = 895.3(2) A' . The full-matrix least-squares refinement resulted in a final

R-value of 0.021 (74 variables, 2954 F-values). NbFe2P12 crystallizes with a superstructure of the TiMn2P12 type structure. The refinement of the TiMn2P,2- like subcell resulted in a R-value of 0.036 (70 variables, 2844 F-values). How- ever, weak superstructure reflections which violate the C centering extinction rule indicate a primitive monoclinic cell. The structure was finally . efined in the space group P2/c to a residual of R = 0.040 (146 variables, 4863 F-values). The lattice constants for the conventional P2/c cell are: a = 10.649(2) A, b = 5.706(1) A, c = 14.929(2) A, B = 102.71(1)°, V= 885.0(1) A3. This cell corresponds to the I2/a setting of the TiMn2P12 structure.

The atructure of TiMn2Pi2 (Fig. 1) can fully be rationalized by classical two- electron bonds. The P atoms are all tetrahedrally coordinated by either two metal and two P atoms (P.'~) or by one metal and three P atoms (P°). The Ti atoms have square-antiprismatic P coordination ("d*sp3 hybrid") and obtain a d°-system (Ti*+), while the Mn atoms in octahedral P coordination ("d*sp3 hybrid") obtain a d'-system (Mn2+). The Mn atoms form Mn-Mn bonds via common edges of the MnP,-octahedra (Fig. 2). In this way all d states of the Mn atoms are optimally utilized as was observed before in the three modifications of MnP, [l-3].

In NbFe2P,2 a total of three additional valence electrons is available. In 2 analogy to the Mo atoms in the structure of MoFe2Pu [4] the Nb atoms obtain a d - system. The Fe atoms are in equal part3 in the oxidation states Fe*+ (d*- system) and Fe3+ (d'-system). The Fe3+ atoms form Fe-Fe bonded pairs (Fe-Fe distance: 2.966 A) while the Fe1+ atons with a d*-system repel each other (Fe-Fe distance: 3.362 A). Thus the superstructure of NbFeiPu is fully rationalized by our rationalisation of the bonding situation. Actually we had expected the superstructure to arise through the two different (bonding and antibonding) Fe-Fe interactions before we solved the structure. P 4 B 2

Figure 1: Crystal structure of TiMn2Pi2. The heights of the atoms are indicated in hundredths of the y coordinate. The TiP,-square-antiprisms are linked to the MnPs- octahedra only via common corners, while the MnP6-octahedra form pairs by a common edge. In the upper right-hand part of the drawing the 18-membered ring of the Pi,l'" -poly-"anion" is shown.

ptei R3I Pill

P(2) PHI PIII v^Pi3i PIS)

P131 PI6I

Figure 2: Near-neighbor environments of the Ti and Mn atoms in TiMn2Pi2. Interatomic distances are given in k units.

References: [1] W. Jeitschko, R. RUhl, U. Krieger and C. Heiden, Mat. Res. Bull. 15,1755 (1980). [2] R. RUhl and W. Jeitschko, Acta Crystallogr. 837, 39 (1981). [3] tf. Jeitschko and P. C. Donohue, Acta Crystallogr. B31, 574 (1975). [4] U. Florke and V. Jeitschko, Inorg. Chem- 22, 1736 (1983). P 4 B 3

EFFECT OF EXTERNAL PRESSURE AND CHEMICAL SUBSTITUTION ON THE PHASE TRANSITIONS IN MnAs A. Zigba Institute of Physics and Nuclear Techniques, University of Mining and Metallurgy, al. Mickiewicza 30, 30-059 Krakow, Poland R. Zach Institute of Physics, Technical University of Cracow, ul. Podchora±ych 1, 30-084 Krakow, Poland H. Fjellvag and A. Kjekshus Department of Chemistry, University of Oslo, Blindern, Oslo, Norway

The concept of "chemical pressure" is used when the effect of external pressure is simulated by substitution. MnAs offers an exceptional possibility of quantitative examination of chemical pressure idea by comparing the effects of external pressure and substitution on phase transitions of different type. Both cation Mn^^As/T: Ti,V,Cr,Fe,Co,Ni/ and anion MnAs^^/X: P,Sb/ substitutions are possible, among which Mn.. .Ti.As and MnAs,, Sb I —X X I —X X exhibit the effects of negative pressure. Figure 1 shows schematically the pressure phase diagram of MnAs (i) . It contains several phases with different crystal structures /NiAs-type or MnP-type/ and with different kinds of magnetic order /P-para, F-ferro and H-helimagnetism/; phase transitions of the first- and second-order are denoted by solid and dashed lines. The transition NiAs,F*±MnP,H was chosen as a calibrating one, since in low temperature it is essentially pressure-driven. In substituted compounds the special pressures

Ps,i and Pg,d /defined as the maximum pressures values for NiAs-*MnP and MnP-»NiAs phase boundaries, see Fig.1/ are shifted in comparison with MnAs, and this shift is a measure of chemical pressure P . Figure 2 shows that for Mn^^Ti^As system the measured Pg ./t/ and Pg d/t/ dependences are linear up to t » 0.08, hence chemical pressure is proportional to Ti concetration with coeffinient £ • P* /t - - 45 kbar. For t • 0.10 the special pressures are bigger and the difference Pg ^ - Pg ^ is decreased. This may indicate the breakdown of the P 4

chemical pressure concept for higher Ti concentration. The values of £ were determined in the similar way for other substituting elements. The effect of substitution on the phase transitions, magneto-

structural Tg *nd structural TD, may be described by dimensionless 1 coefficients T^ dTc/dt and Tp dTD/dt. These coefficients estimated from experimental T-t phase diagrams (2) are compared in Fig.3

with calculated from the known pressure derivatives dTc/dP, dTD/dP for MhAs and the coefficients of chemical pressure £.. The chemical pressure appears to be the major factor determining the substitutional shift of transition temperatures. The systematic difference A A

between T^ dTc/dt andETg dtc/dP may be semiquantitatively explained by effect of dilution of magnetic Mh sublattice by nonmagnetic atoms of Ti,V,Co and Ni (3). For the structural transition the large systematic difference between Tl dT^/dP and ST^dT^/dP may be linked to the effect of disorder, difference of electron number and/or the change of atomic mass. The more detailed report including the analysis of lattice constants and the results for anion substitution will be given elsewhere (4).

References (1) N. Menyuk, J. A. Kafalas, K. Dwight, and J. B. Goodenousrh. Phys. Rev. 221, 9^2 /1969/. (2) K. Selte, A. Kjekshus, A. F. Andresen, and A. Zie;ba, J. Phys. Chem. Solids J58, 719 /1977/ and references therein. (3) A. Zifba, H. FjellvSg, and A. Kjekshus, J. Phys. Chem. Solids, in print . (4) A. Zie;ba, R. Zach, H. FjellvSg, and A. Kjekshus, to be published. P 4 B 3

NiAs,\_ 3^v- — 8

0.04 0.08 0.12 Fig.1. Fig.2. t

P-T phase diagram The special pressures Ps .?„ . of MnAs after (i). vs. titanium content for system.

10- Mnw Tt As A 8- - \ • \ 6- • / \ dTD U- dP

2- dT Ti \ / V * Fe\ Co dt i i i vf \ Cr -2-

— /. — if - ~^J -Ml ' Us Tc dP -6- \ A" 1 dTc dt -8- V Fig.3. The comparison of experimental /•,•/ and calculated /O,D/ changes of transition temperatures upon cation substitution.

A. Zif^bs, Institute of Physics and Nuclear Techniques, University of Mining and Metallurgy, Al. Mickiewicza 30, 30-059 R/oLko'w Poland P 4 B 4

NONSTOICHIOMETRY IN B8-TYPE NiSb R. Leubolt, H. Ipser, P. Terzieff, and K.L. Komarek Institut fur Anorganische Chemie der Universitat, wahringerstrafie 42, A-1090 Wien, Austria

The appearance of phases with a NiAs-derivative structure is common to many transition metal - B-metal systems. With a few exceptions, most of theso B8-type phases deviate from the ideal composition ratio of 1:1 due to either the subtraction of transition metal atoms from their regular sites or the addition of interstitial transition metal atoms. In NiSb both the additive and the subtractive type of structure are combined into one phase field ranging from about 43 to 51.5 at% Sb at 1300 K (Fig.1).

1100 AV Liquidus O SoliOus t OTA • invariant arrest J

• smots thirmal analysis 1000 ' + nnase boundary from rhermodyn

897 i 3 900 865:5

800 40 45 50 at% Sb

Fig.1 Phase boundaries of the B8-type phase NiSb P 4 B 4

In principal accordance with earlier observations (1) the congruent melting point was determined to be at 47 at% Sb and 1432 ± 5 K. Isopiestic measurements (2) yielded an inflection point of the partial thermodynamic properties of Sb at about 50.6 at% Sb indicating that the interstitial (additive) and the transition metal deficient (subtract!ve) type of structure are not strictly separated. This is also reflected by the composition dependence of the lattice parameters which show a change of slope located at a composition higher than 50 at% Sb (Fig.2). The densities, determined by a pycnometric method using kerosene, were found to vary linearly with composition in both the interstitial and the transition metal deficient regions with a discontinuity at about 50.6 at% Sb (Fig.3). Magnetic measurements at room temperature yielded a flat minimum between 50 and 51 at% Sb characterized by a nearly non-magnetic behaviour. Similar results have been reported by other authors (3).

1 r I 1 I 1 ' 1 '11 • (A) -*3CC < 5.20

5.15 -

5.10 i • 1 , 1 . i.l. ..4 46 48 50 af%Sb 54 42 44 46 48 50 at% Sb 54

Fig.2 Lattice parameters of B8-type NiSb as function of composition (O,v quenched from 1300 K and 1173 K) 9.0

X X 8.8 " X X X X X •*> "TV X X S 8.6

K 0 X -

8.4 -

- - 1173 K

8.2 1 1 i 1 | i 50 52 54 at%Sb

Fig.3 Densities of B8-type NiSb as function of composition (x, calculated densities)

(1) Tu Chen, J.C. Mikkelsen, and G.B. Charlan, J.Crystal Growth 13, 5 (1978) (2) R. Leubolt, Ph.D.Thesis, University of Vienna (1985) (3) Tu Chen, D. Rogowski, and R.M. White, J.Appl.Physics 49, 1425 (1978) —

R. Leubolt, Institut f. Anorganische Chemie der Universitat, WahringerstraBe 42, A-1090 Wien, Austria p 4 a

MAGNETIC TRANSITIONS, SITE OCCUPANCIES AND STRUCTURAL VARIATIONS 111 (Fel-xMnx)2P (0-01

Y. Andersson Institute of Chemistry, University of Uppsala, Box 531, S-751 21 Uppsala, Sweden

Fe2P and Mn2P crystallize in the hexagonal Fe2P-type structure. Fruchart et_ aJL. (1) found that this structure type is maintained in the ternary alloys (Fe.._xMnx)2 P for xi0,31 and x>0.62 while an intermediate phase of the orthorhombic CO2P-type structure exists in the approximate composition range 0.31

P 01 x Different alloys of (Fe.j_xMn „) ? (°- ^ S0•85) were prepared by mixing appropriate amounts of Fe2P and Mr^P and heating in evacuated silica tubes at temperatures between 1120K and 1270K. The products were examined by X-ray powder diffraction techniques. (Fe-]_xMnx)2P alloys in the composition range 0.27ixs;0.67 were found to crystallize in the orthorhombic Co2P-type structure and the alloys with x<0.23 and x>0.71 in the hexagonal Fe2p-type structure.

Mossbauer spectra of (Fe-| _xMnx)2 P with x = 0. 01 ,0. 03i0 .1 5 , 0 . 30 , C.60, 0.85 were recorded in absorber mode using a 57coRh source in the temperature range 10-400K, including paramacnetic and mac- net ically ordered states. The data obtained were analyzed in terms of atomic and magnetic ordering, and the hyperfine field parameters were determined.

References

(1) R. Fruchart, A. Roger and J.P. Senateur, J. Appl. Phys., 40, 1250 (1969).

Y. Andersson, Institute of Chemistry, University of Uppsala, Box 531, S-751 21 Uppsala, Sweden P 4 B 6

A STUDY OF THE HOMOGENEITY RANGES OF THE KAPPA-PHASES IN THE Hf-Mo-{Si,P,S,Ge,As,Se} SYSTEMS

A. Harsta Institute of Chemistry, University of Uppsala, Box 531, S-751 21 Uppsala, Sweden

In previous studies new ic-phases were discovered in the ternary systems Hf-Mo-{Si,P,S,Ge,As,Se}, Hf-W-{P,S,As,Se}, Hf-Re-{Si,P,S, Ge,As,Se} and Hf-Os-{Si,P,S,Ge,As,Se) (1-3). The unit-cell volumes of these phases seemed to be connected with the nonmetal atomic species involved but could not be immediately explained in terms of atomic size differences between the nonmetal atoms. A closer investigation of the K-phase in the Hf-Mo-P system has shown that this compound has an extended range of homogeneity, which is due to a variable Hf/Mo substitution in the metal atom sublattice (4). An analogous situation might prevail also in other K-phases. The present study reports the results from an investigation of the unit-cell volume variations for the K-phases in the ternary systems Hf-Mo-{Si,P,S,Ge,As,Se} and from X-ray single-crystal structure refinements of the <-phases in the Hf-Mo-Se and Hf-Mo-Ge systems. Some results on the various phase equilibria involving the ic-phases will also be presented.

(ij A. Harsta and E. Wennebo, Acta Chem. Scand. A35, 227 (1981). (2) A. Harsta and E. Wennebo, Acta Chem. Scand. A36, 547 (1982). (3) A. Harsta/ Thesis. Acta Univ. Ups. (Faculty of Science) No. 767 (IT: (4) A. Harsta, Acta Chem. Scand. A36, 535 (1982). P 4 B 7

K2PtH4, SYNTHESIS AND STRUCTURE

W. Bronger, G. Auffermann, K. Hofmann and P. MUller Institut fiir Anorganische Chemie der RWTH Aachen, Prof.-Pirlet- Strafle 1, 5100 Aachen (F.R.G.) H.F. Franzen Ames Laboratory, Iowa State University, Ames, Iowa 50011, (U.S.A.)

Ternary metal hydrides A M.H with A = alcalimetal and M = transi- x y z tion metal are hardly known. Our investigations on platinum compounds revealed the existence of LiPtHQ gg [1] and Na-PtH, [2]. The first compound shows a metallic behaviour, the second is a salt-like hydride in which the platinum atoms are planar coordinated by four hydrogen atoms. Within the system K/Pt/H we synthesized the compound K-PtH, by the reaction of potassium hydride with platinum sponge in a H--atmosphere. The temperature range was 280° to 400°C.

Fig.: on the left side: arrangement of the potassium (dashed circles) and platinum (full circles) atoms; on the right side: positions of the deuterium atoms, which occupy the corners of the octahedra partially (2/3). P 4 B 7

X-ray investigations and neutron diffraction experiments on the deuterated compound revealed its structure. The atomic arrangement was found to be similar to the K2PtClg-type with partially occupied clorine positions (a » 798.0(1) pm for K2PtD4) (cf. fig.).

Susceptibility measurements showed diamagnetism, characteristic for a planar coordinated d -ion. At lower temperatures we found an ordered tetragonal structure (a = 555.8(1) pm and c = 803.5(1) pm). The phase transition is discussed.

References :

1. B. Nacken und W. Bronger, Z. anorg. allg. Chemie 439 (1978) 29. 2. W. Bronger, P. Miiller, D. Schmitz and H. Spittank, Z. anorg. allg. Chemie, in press.

Prof. Dr. W. Bronger, Institut fttr Anorg. Chemie der KWTH Aachen Prof.-Pirlet-Str. 1, 5100 Aachen (P.R.G.) P 4 B 8

AHD EIECTHOff IRRADIAIIOH EFFECTS XV OBBERED P-PiH(D) AB0U9D THE RRSISTTTTPT AHOMALT HEAR 50 Z. J.P. BuTf«r, J.5. Daon* A. Lucaaaon and P. 7a.ida

EHA N9 720 du C.N.H.3., Bit. 350, Oniversite"BiFis-Sadf F-91405 Orsay, Trance

(3-PdH(D) exhibits, for 0.6 -t r £ 0.7, a well-known isotope dependent anomaly near T = 50 K which had been observed by specific heat and electrical resistivity measurements; it has been identified by neutron scattering as due to a complicated ordering process involving short-range ordered (SRO) and long-range ordered (LRO) configurations of hydrogen on octahedral interstitial sites and vacancies (for a review, see £lj). We have investigated the low-temperature behaviour of the electrical resistivity of various |3-PdH(D) specimens after irradiation with electrons of subthreshold energy with regard to the metal lattice (S = 350 to 800 keV) and compared it with that of quenched and of relaxed specimens. The following significant features were observed (see Pig. l): (1) A quench across the anomaly lowers the residual resistivity by -Ap , in [2] q agreement with earlier measurementsu *, (2) Electron irradiation at T ss 20 K also lowers the residual resistivity, in- ducing an energy dependent -Ao. ; (3) Effects (l) and (2) are additive, with only minor saturation effects until

-Ao/oo of several ;

U)An isotope effect is observed under irradiation, -&qi/0 (PdH)« -Z&o./tf (PdD), for a given electron fluence $; (5) The induced resistivity changes, -£o and -£$2-.,* anneal both in the anomaly region, maintaining the isotope effect on the recovery temperature. The phenomena are understood when assuming a partial freezing-in of the low-o SRO-regions under quench; electron irradiation, on the other hand, creates Pren- kel-type defects in the hydrogen sublattice of the IBO structure at low temperature (with an isotope-mass dependent energy transfer). These defects seem to be analogs of the high, temperature SRO regions.

(1) 0. Blaijchko, J. Less-Conson Met. 100, 307 (1984) (2) T.B. Ellis, Thesis, Oniv. of Illinois at Urbana (USA), 1978 P 4 B 8

Pig. 1: (0 10 103 K Normalized resistivity, o/q (T), taken with increasing temperature for PdH(D) M after Tarioos treatments. The broken ?/^ line represents in all cases the relaxed (slowly cooled) specimen, the fall line was obtained for: (a) PSD, quenched .» across the anoaaly; (b) PdD and (c) MS, electron irradiated at 20 K in the same conditions; (d) PdD, quenched and subsequently electron irradiated. The small arrows indicate the T . and T of the anomaly for the specimens after various treatments, helping to visualize the isotope effect.

70 90 K

Dr. Peter 7ajda Laboratoire de Cbiaie Fhysjiqne, Bit. 350, Onlrersit* ?arl»-Sad, ?-91405 Orsay, Trance P 4 B 9

SUFERCOHDUCTIHG PHASES OF

S.O. Janz, A.J. Pindor and F.D. Manchester Department of Physics, University of Toronto, Toronto, Ontario M5S 1A7

Hecent work has shown (1)(2) that there are two superconducting phases of PdH , both of them Type II superconductors with differing Tc's evident at hydrogen concentrations above x - H/Pd a 0.8k. On present evidence, one structure is an ordered form of PdH consistent with the presence of an ordering region for the concentration range 0.7 < x <$ 0.9, and below 100K, in the PdH, phase diagram. The effect of the ordering process on samples of PdH cooled to temperatures where they can become superconducting, is to introduce an uncertainty in the relative concentration of these two phases in a given sample. Thus the true form of the Tc(x) relation for each phase has yet to be determined, and thus the estimate of Tc(x) for any sample of PdHx, measured in the past, is suspect. At present, the experimentally measured difference between the Tc's for disordered and ordered PdH is considerably less than the X difference between the best available theoretical estimate of Tc(x) and the most reliable measured Tc(x) for PdH , which, we now know to contain a mixture of the ordered and disordered phases. Thus, apart from an experimental effort to unambiguously determine Tc(x) for the disordered and ordered phases of PdH , a further refinement of calculations to determine Tc(x) for each of the phases from theory is required, if the superconductivity of PdHx is to be accounted for satisfactorily. A discussion of possible ways of improving the calculation of Tc(x) values for disordered PdH is presented.

(1} I.S. Balbaa and F.D. Manchester, J. Phys. F: Met. Phys, 13, 395-^Oh (1983). (2) I.S. Balbaa, A.J. Pindor and F.D. Manchester, J. Phys. F: Met. Phys. lU_, 2637 (198J+). P 4 B lo

TRANSITION METAL-HYDRIDES : ELECTRONIC STRUCTURE, STRAIN EFFECT AND H-H BINDING ENERGY. J. Khalifeh Department of Physics, The University of Jordan, Amman (Jordan) G. Moraitis Department of Physics, University of Antananarivo, Antananarivo (Madagascar) Ch. Minot Laboratoire de Chimie Theorique, Bat. 490, Universite Paris-Sud 91405 Orsay Cedex (France) M.A. Khan and C. Demangeat L.M.S.E.S. (U.A. au CNRS n° 306) - Universite Louis Pasteur 4, rue Blaise Pascal 67070 Strasbourg Cedex (France)

The behaviour of Hydrogen in transition metals remains a complicated subject for two reasons. Firstly, the impurity provides a large perturbation which leads to a strong hybridization with the metallic electrons. Secondly, the impurity couples strongly to the host ion coordinates and leads to strain effects.

In the first part of this abstract we will proceed along the lines proposed recently by Mackintosh and Andersen (t) for relaxation in pure metals and extend it to dilute alloys of transition metals (2) and to Hydrogen in Palla- dium (3). The lattice displacement u.(R) of a metallic atom at site R, resulting from the presence of a Hydrogen atom at site A, can be obtained from the knowledge of the forces acting on site R and of the lattice Green function (4). These forces and force constants can be obtained formally from the first and second derivatives of the total energy of the system calculated in the tight binding approximation in which the variation u.(R) is included. From the relation between the dipole force tensor P, measured experimentally, and the field of forces F(R) acting on site R, we have deduced F(R) on the three neighbouring shells of the interstitial Hydrogen (5). Also, force constants in the-alloy can be obtained from an expansion of the total energy of the system up to second order in u.(R) (6). However, as discussed in (6), it is by no means trivial to perform numerical estimation of these complicated forms of the force constants (7). Up to now, only calculation of this field of displacement for H in Pd (8) using rough approximation for the force constants have appeared in the literature. Application to H in Ni is under present investigation (9). Also, Hydrogen-Hydrogen binding energy, in this very low concentration limit, has been obtained (10) and P 4 B lo

shows a strong repulsion up to second nearest neighbouring position.

The second part of this abstract is devoted to a recent calculation of lattice location, heat of formation, Hydrogen-Hydrogen repulsion (11) and trapping effect in a-Fe. This calculation uses the tight binding method (12) of band structure calculations based on extended Huckel (13) formalism. To compute the binding energy of a pair of Hydrogen atoms we consider an unit cell containing two Iron atoms and two Hydrogen atoms. The Hydrogen atoms are expected to be at tetrahedral positions (14) ; they generate two bands whose energies are close to -13.6 eV. The combination with d band metal leads to stabilize the Hydrogen 1s bands and to destabilize the d bands, the later being pushed up. The upraisal of the d bands is responsible for the positive adsorption energy. The geometry with two Hydrogen atoms at third nearest neighbour position emerges from the results as the most probable. This appears in agreement with the recent work of Boureau (15) concerning the description of phase diagrams of hydrides of bcc transition metals in the framework of the blocking model. Presently, this method is extended to the calculation of the lattice location of Hydrogen in a-Fe and to the estimation of the energy at tetrahedral and octahedral positions ; hence the migration path of the interstitial. Migration may be influenced by point defects such as substitutional (Ti, V, Co, Ni...) or interstitial (C, N, 0...)• One of our purposes will be the estimation of their binding energies with Hydrogen, in order to classify them into traps or repellers (16).

(1) A.R. Mackintosh and O.K. Andersen, in Electrons at the Fermi Surface, ed. by M. Sprinford (Cambridge University Press) 149 (1980). (2) G. Moraitis, B. Stupfel and F. Gautier, J. Phys. F _M_, L 79 (1981) and to be published. (3) J. Khalifeh, G. Moraitis and C. Demangeat in Electronic Structure and Properties of Hydrogen in Metals, P. Jena and C.B-. Satterthwaite eds. (Plenum Press) 119 (1983). (4) V.K. Tewary, Adv. Phys. 22, 757 (1973). (5) J. Khalifeh, G. Moraitis and C. Demangeat, J. Phys. (Paris) 43_, 165 (1982) ; J. Less Com. Met. 85, 171 (1982). (6) G. Moraitis, J. Khalifeh and C. Demangeat, J. Less Coram. Met. 101, 203 (1984). (7) G. Moraitis, A. Mokrani and C. Demageat, to be published. (8) J. Khalifeh, G. Moraitis and C. Demangeat, 7th Conf. on Solid State P4 Bio

Science and Applications (Cairo, 198-t), to be published by Egyptian Journal of Solids. (9) Z. Badirkhan and J. Khalifeh, to be published. (10) J. Khalifeh, G» Moraitis, M.A. Khan and C. Demangeat, Studies in Physical and Theoretical Chemistry_32, 445, ed. F. Lacombe, Elsevier (1984). (11) C. Minot and C. Demangeat, Int. Symp. on Hydrogen in Metals, Belfast (1985). (12) M.H. Whangbo and R. Hoffmann, J. Am. Chem. Soc. 100, 6093 (1978) ; C. Minor, M.A. Van Hove and G.A. Somorjai, Surf. Sci. 127, 441 (1983). (13) R. Hoffmann, J. Chem. Phys. 39, 1399 (1963). (14) E. Ligeon, 1984, private communication. (15) G. Boureau, J. Phys. Chem. Sol. (1984) under press. (16) G.M. Pressouyre, Met. Trans. A U_, 2189 (1983).

C. DEMANGEAI LMSES - DLP 4, roe Blaise Pascal B 11

ON THE HYDHIDING KINETICS OF LagMg^,-, AND la^^G M. Khrussanova, M. Terzieva and P. Peshev Institute of General and Inorganic Chemistry, Bulgarian Academy or Sciences, 1040 Sofia, Bulgaria

The hydriding kinetics of the alloys LagMg^r,, La,, .aCao.2Ms17 and La,, cp^-n tiP&*n were studied using experimental results from previous papers (1, 2). The rate-controlling stage of the process was determined on the basis of the theoretical equations proposed by Park and Lee (3) and Lim and Lee (4), according to which the experimental data should satisfy one of the following equations: (i) for hydrogen chemisorption on the alloy surface

(ii) for three-dimensional hydrogen diffusion through the product layer 1 + 2(1 - F) - 3(1 - ?)2/3 = K,(py2 - ?l(2)t /2/

(iii) for the chemical reaction between hydrogen and the alloy

where F is the conversion degree; A, a constant ( A = = otf(e-Hirr2/C2*MR)'1/'2 ); ac , the condensation coefficient; M, the molecular weight of the gas; f(0-), a function of the surface coverage; R, the gas constant; T, the absolute temperature; E , the activation energy of chemisorption; r , the initial 3. O radius of the alloy particles; EQ, the pressure of the hydriding process; P eq, the equilibrium gas pressure determined from the P-C-T curves; K^, a coefficient of the diffusion rate depending on T alone; K, a temperature-dependent constant. The conversion degree versus time curves at various pressures and a temperature of 325°C have lineaEC parts for all three alloys when 0< E<0.4. Hence, within this range of F values, hydrogen chemisorption on the surface of alloy particles is the rate-controlling stage of the process. P 4 B 11

Equation /1/ can be written as —'E /RaT ^ Vfr /V where K- is a function of temperature only. The activation energy of the alloys under investigation is determined from the plots of the lnKL = f(1/T) dependence. The E values obtained are presented in Table 1 which also contains data on the alloys and Mg-2O#LaNi,- calculated by other authors (4, 5) and the corresponding values of hydrogen capacity. Table 1

Activation energy Absorption capacity No Alloy of chemisorption, of the alloy at 325 C EQ (cal/mol) and EQ = 30 bar cL

1. Mg-2O*LaNi5 98Q0 3.0 2. La1.8Ga0 2Hg17 5940 3.5 3. La1 &CaQ 4Mg1? 4950 4.2 4. La2Mg1? 3960 4.7 5. GeHg12 3560 4.7

Obviously, the increase in activation energy of cheadsorption is associated with adecrease in absorption capacity. This can be explained taking into account that the higher the concentration of hydrogen ions formed during hydrogen chemisorption on the alloy surface, the higher the absorption capacity of the latter. At conversion degrees of 0.4< F< 0.65 the experimental data satisfy both equation /2/ and equation /3/. In this case the calculated E values indicate that the three-dimensional diffusion Ok should be rate-controlling, which is also accepted by other authors for other alloys (6 - 10). For the alloy La^Mg,.,,, E = = 3460 cal/mol if the chemical reaction is the rate-controlling stage and E& =* 6250 cal/mol if hydrogen diffusion through the M boundary layer is rate-limiting. For La^ gCa0 2 &i7 these values are 2000 and 5350 cal/mol, respectively.* On the basis of the results obtained it magr be concluded that at low conversion degrees hydrogen chemisorption on the alloy surface is the rate-controlling stage of the hydriding of the pure La Ga M alloy LagMg^ and. the alloys 2_x x &T-o (where part of the lanthanum is replaced by calcium), "the introduction of calcium increasing the activation energy of the cheaisorption. At 0.4 < F< 0.65, the three-dimensional hydrogen diffusion through the hydride phase formed is rate-limiting. In this case calcium facilitates the diffusion process by increasing the interface. This is indicated by the higher values obtained for the coefficient of the diffusion rate K^ (equation 2) with Ca-substituted alloys compared to that for the pure alloy

(1) M. Khrussanova, M. Pezat, B. Darriet and P. Hagenmuller, J. Less-Common Metals, 86, 153 (1982) (2) M. Khrussanova, M. Terzieva, P. Peshev, K. Petrov, M. Pezat, B. Darriet and J. P. Manaud, Int. J. Hydrogen Energy (in press) (3) C. N. Park and J. Y. Lee, J. Less-Common Metals, 83, 39 (1982) -* (4) S. H. Lim and J. I. Lee, Int. J. Hydrogen Energy, 8, 369 (1983) (5) M. I. Song and J. I. Lee, Int. J. Hydrogen Energy, 8, 363 (1983) (6) P. S. Rudman, J. Less-Common Metals, 89, 93 (1983) (7) C M. Stander, 2. Phys. Chem. (N.F.), 104, 229 (1977) (8) M. Mintz, Z. Gavra and Z. Hadari, J. Inorg. Nucl. Chem., 40, 765 (1978). (9) M7 Mintz, S. MalJciely, Z. Gavra and Z. Hadari, J. Inorg. Nucl. Chem., 40, 1951 (1978) (10) E. Aiciba, K. Nomura, S. Qno and S. Suda, Int. J. Hydrogen Energy, 7_, 787 (1982)

Prof. Dr. Pavel PESHEV, Institute of Genera&L and- Inorganic Chemistry, Bulgarian Academy of Sciences, 1040 SOFIA, Bulgaria P -4 B 12

INFLUENCE OF INERT GASES UPON HYDROGEN ISOTOPES ABSORPTION AND DESORPTION BY MmNi Al 4.5 0.5 E. Tuscher and P. Weinzierl

Institut fur Experimentalphysik der Universit*a*t Wien, Strudlhofgasse 4, A-1090 VIENNA, Austria

O.J. Eder and E. Lanzel

Osterreichisches Forschungszentrum, A-2444 SEIBERSDORF, Austria

Introduction

In advanced nuclear technology, the use of transition metals and their alloys is of some interest in collecting and storing of hydrogen isotopes [1-3]. With re^-ird to the feasibility of hydride forming metals in scavenging hydrogen isotopes from admixed inert gases, investigations performed recently proved the fact that even small amounts of helium or argon were very effective in blocking the hydrogen absorption under static test conditions [4-7]. This

"blanketing effect" which may be attributed to diffusion-controlled concentration gradients in the gas phase could be reduced by circu- lating the gas mixture over the metal surface [6].

Experimantals and Results

The influence of different inert gas concentrations on hydrogen

(deuterium) sorption characteristics of MmNi Al (Mm = Misch- 4.5 0.5 metall) was investigated in a single container system using the gasvolumetric method. Absorption and desorption of H (D ) were 2 2 thermally induced by cycling the temperature of the absorbent between 286 K and 348.5 K. X , the hydrogen (deuterium) concentra- t tion in the metal at tine t, was determined via the gas pressure measured at this time. An increase of the inert gas amount present P4 BLi

in the system resulted in a decreased hydrogen (deuterium) sorption rate, as shown in fig. 1 for MmNi Al /D /Ar: The deuterium to 4.5 0.5 2 metal ratio X is larger for the desorption process and smal- t,n(Ar) ler for the absorption process than X (deuterium to metal ratio t,0 in the absence of inert gas), due to the decreased sorption rate in the presence of n moles argon. Therefore, the ratio X /X t,n(Ar) t,0 is larger than 1 for the desorption process (upper half of 'ig. 1) and lass than 1 for the absorption process (lower part of fig. 1). The observed "inert gas effect" diminishes with reaction time t. NTo dependence of the equilibrium data (pressure, hydrogen capacity) upon the inert gas concentration could be detected. Curves similar to these shown in fig. 1 were found substituting D for H and Ar 2 2 for He.

References [I] J.R. Powell, Rep. BNL 20563, 1975 (Brookhaven National Labora- tory, Upton, N.Y.). [2] J.L. Maienschein, Nucl. Techn. 2S. (1973) 387. [3] K.D. Fischmann and H.D. Rohrig, in: Metal-Hydrogen Systems (T.N. Veziroglu, ed.), Pergamon Press, Oxford, 1932; p.691. [4] J.L. Anderson, T.C. Wallace, A.L. Bowman, C.L. Radosevich, and M.M. Courtney, Rep. LA-5320-MS, 1973 (Lo<: Alamos Scientific Laboratory, Los Alamos, N.M.). [5] R.S. Carlson, in: Radiation Effects and Tritium Technology for Fusion Reactors (J.S. Watson and F.W. Wiffen, eds). Oak Ridge National Laboratory, Tennessee, 1976 (CONF-750939); p.IV-36. [6] J.M. Yaraskavitch and W.J. Holtslander, in: Metal-Hydrogen Sys- tems (T.N. Veziroglu, ed.), Pergamon Press, Oxford, 1982; p.619. Rep. AECL-7151, 1981 (Chalk River, Ontario, Can.). [7] T. Fujimaki and Y. Tanaka, J. Less-Common Metals §9 (1983) L34.

Acknowledgement This work was supported by the Bundesministerium fur Wissenschaft und Fsrschung, Vienna, Austria, within the fusion research program coordinated by the Austrian Academy of Sciences. 4 B 12

1 uof Xt>fl z

* IJJ7 o 2O2S T 7U7 o 160.9S a 2SB3)

Fig. 1; Influence of different amounts of argon. n(Ar), on

deuterium absorption and -desorption by MmNi Al : 4-5 0.5 X : deuterium to metal ratio at reaction time t t,n(Ar) in the presence of n moles argon X : deuterium to metal ratio at reaction time t in t.O the absence of argon

Dr. E- TUSCHER Inst.f. Experimentalphysik der Universitat Wien Strudlliofgasse 4 A-1090 WIEN P 4 B 13

LOW-TEMPERATURE STRUCTURE OF EVIDENCE FOR MICROTWINNING P.ZolliJcer and K.Yvon Laboratoire de Cristallographie aux Rayons X, Université de Genève, 24 Quai E.Ansermet, CH-1211 Genève 4 (Switzerland)

Ch.Bärlocher Institut für Kristallographie und Pétrographie, ETH, 8092 Zürich (Switzerland)

The hydrogen storage material Mg2NiH4 undergoes a structural phase transformation from a disordered, cubic high-temperature (HT> modification into a presumably ordered low-temperature (LT) modification of lower symmetry. Conflicting results on the existence of possibly different LT-modifications, their symmetry and their detailed atom arrangement have been reported (for a recent review see (D). According to these reports at least two different LT-phases exist, one (called LTD of monoclinic symmetry and another (called LT2) of orthorhombic symmetry . Approximate atomic coordinates have sofar only been reported for the LT1 phase. The structure refinements have been performed on powder diffraction data and have yielded relatively poor agreement factors between observed and calculated intensity profiles (for example R =0.16, 22 refined atomic parameters in space group Cc (2)). Likely reasons for these bad fits were the necrlect of anisotropic line broadening and the inadequate attribution of certain diffraction maxima to the LT2 phase. In this work various powder diffraction patterns of

LT-Mg2NiH4 were analysed in terms of a model which did not require the assumption of a second phase. The model consisted of a monoclinic metal atom arrangement similar to that of the LTl-phase (a=6.491A, b= 6.407A, c*13.207A, ß=93.21°, space group 12/a.) which contained various concentrations of planar faults. As shown in Fig.l the faults consisted of twin planes oriented parallel to (001) which were separated along c by multiples of c/2. A structure which contains a low probability of twinning («so) has mainly monoclinic symmetry» whereas a structure which contains a high probability of twinning («si) has mainly orthorhombic symmetry. P 4 B 13

.25 a= .50

a- .90 a=1.00 Fig.l: Monoclinic containing various probabilities of twinning (a).

- o- .25 A

- a- .50

- a- .90

10.0 20.0 30.0 40.0 50.0 60.0 70.0 29-degrees Co Ka

Fig.2: Simulattd X-ray powd«r diffraction patttrn* for ^ f assuming various probabilities of twinning (a) as shown in Fig.l. Theoretical X-ray powder diffraction intensity profiles were calculated for various probabilities of twinning- by numerical integration of expressions given by Cowley and Au (3). A3 shown in Fig. 2 already small values of cc lead to significant peak broadening- (see peaks marked by arrows in the profile for a=0.10). At intermediate values of « these peaks are further broadened and gradually disappear whereas new relatively diffuse peaks appear (for example those marked by stars for a=o.25, 0.50 and 0.90). At high values of « these new peaks increase in hiaht and aradually become sharp. A comparison between these calculated intensity profiles and the experimentally observed powder patterns confirmed these features. A structure refinement with consideration of microtwinning was performed on a pattern which resembled a intensity profile calculated for a twinning probability of s^O.l*. The refinement yielded a significantly better fit (R =0.12, space group C2/c, 12 atomic parameters) and more reliable metal atom parameters than those obtained in previous refinements without consideration of microtwinning.

Interestingly Mg2NiH4 samples which were rapidly cooled

through the HT-LT phase transformation (Tpi235°C) crave significantly different powder diffraction patterns. The patterns resembled the intensity profiles calculated for higher twinnincr probabilities (0.20

References: (1) K.Yvon, J.Less-Common Met.,103 (1984) 53. (2) H.Hayakawa, Y.Ishido, K.Nomura, H.Uruno and S.Ono to be published in J.Less-Common Met. (3) J.M.Cowley and A.Y.Au, Acta Cryst.,A34,(1978),738.

P.Zolliktr, Labozitoir* 9.m Cristailographia aux Rayons X, Vnivmxsiti Am G«a#v« 24 Quai £. Ansarort, CH-1211 Gtatv 4 (Switxtrland) P 4 B 14

PREPARATION, STRUCTURE AND PROPERTIES OF M?2CoH5 CONTAINING SQUARE-PYRAMIDAL CCoHg]4" ANIONS

P.Zollifcer and K.Yvon Laboratoire de Cristallographle aux Rayons X, Universite de Geneve, 24 Qua! E.Ansermet, CH-1211 Geneve 4 (Switzerland)

P.Fischer and J.Schefer Labor f(lr Neutronenstreuung, ETH ZUrich, CH-5303 WUrenlingen (Switzerland)

During a search for new hydrogen storage materials which contain 3d transition elements we have recently discovered a ternary metal hydride of composition Mg2FeH6 in which the iron atoms were surrounded octahedrally by hydrogen atoms (1). Here we report the synthesis, structure and properties of a new related hydride of composition Mg2CoH5 in which the cobalt atoms are surrounded by hydrogen atoms in a square-pyramidal configuration.

Mg2CoH5 and its deuteride were prepared from the elements by a sintering technique similar to that described previously (1). Pellets containing 2:1 mixtures of Mg and Co powders were heated during several days in a steel autoclave at temperatures between 390° and 420°C (380°-400°C for the deuteride) and hydrogen (deuterium) pressures between 40 and €0 bar. X-Ray and neutron powder diffraction data were recorded at room temperature and 230°C. The data at room temperature suggest a tetragonal distorted CaF2 type metal atom structure (deuteride: a=4.463(l)A, c=6.594(l)A; space group P4/nmm, Z=2). The D atoms surround the Co atoms in an ordered square-pyramidal configuration (dCCo-D3»l.590(17)A (apical), 1.515(3)A (basal)). At 230°C the structure transformed into a disordered cubic modification (a«6.460(2)A). Square-pyramidal complex ions based on cobalt are relatively rare, and the present compound appears to be the first example for such a complex which contains Cod) Ions. An example for another square-pyramidal da system having a very similar geometry is the complex Ion Ni(CN)_ As shown In Fig.l Mg2CoH5 can be considered as an intermediate member in the isoelectronic structural series

Mg^FeHg - Mg2CoH5 ~ Mg^iH*- In the latter compound the H atoas surround the nickel atoas in a deformed square-planar configuration (2). All three compounds obey the 18-electron rule.

With respect to the physical properties Mg2CoHs resembles closely its Ni and Fe based neighbours. Its heat of dissociation as measured from pressure-composition isotherms is 86(5) kJ/mole

H2, which is lower than that of Mg2FeH6 (98(3) JcJ/mole H2) and

higher than that of Mg2NiH4 (64 kJ/mole H2). Its volume and weight efficiencies for hydrogen storage (4.5 wt.%, 7.6xlO 22 H 3 22 atoms/cm ) are higher than those of Mg,NiH4 (3.6 wt.%, 5.4xlO

H atoms/cm ) and lower than those of Mg2FeH6 (5.4 wt.1t, 9.1x10

H atoms/cm >• Mg2CoHg shows non-metallic behavior.

(a) (b) (c)

,4- 4 CCoHK3 ~ 5 Fig.l: Hydrogen environment of the 3-d transition elements in Mg2NiH4 (a), Mg2CoH5 (b) and Mg2FeH& (c)

References:

(1) J.-J.Didisheim,P.Zolliker,K.Yvon,P.Fischer,J.Schefer, M.Gubelmann and A.F.Williams, Inorg.Chen.21,(1984),1953. (2) D.Noreus and P.-E.Werner,J.Less-Common Met.97,(1984),215.

P.Zollikir, Uboratoir* d* Cri«tallogr*phii auz Rayons X, Uniwsit* dt Gtn*v* 24 Quai E. Aannit, CH-1211 G«ntv 4 (9»»itx«rland) A U THOR I N 0 E X

A60STINELLI* E. PI A 1 CARCALY* C. PI A 2 P3 A 1 CARROLL* P.J. PI B 5 AMELINCKX* S. 0 6 CASCALE5» C. PI B 9 ANDERSSQN, Y. P4 B 5 CENZUAL* K« P3 B 2 ANDRESEN* A.F. P3 A 2 CHABOT* B. P3 S 2 P3 A 4 CHANDRA* S. PI A 6 ANNE* M. P4 A 2 CHATTOPADHYAY* T. 0 2 AUFFERMANN* G. P4 B 7 CHAUDOUET* P. 019 8ABEL* 0. ?Z A13 P3 A 5 BACH, H. ai* P3 B 7 BACMANU* H. P4 A 2 P4 A 2 P4 A 3 P4 A 3 P4 A k P4 A13 BADUREK, G. P3 Al 4 CHAUSSY* J, P2 A 4 BAtRLOCHÉR* CH. P4 B13 CHEN* C.H. PI B 5 BAERMsR* K. P3 A 2 CHENAVAS, J. P3 aio BARNIcR* S. P3 A 3 CHENËVIER* B. P3 A 5 BARTHELEMY* E. PI A 2 P4 A 2 BARTuLOME* J. P3 A 5 P4 A 3 P3 AIO P4 A 4 BAUER/ J. P2 A 1 P4 AIO BAUEk-JROSSE* E. ?Z A 2 CHRISTENSEN* A.N, PZ à 1 BECKHAM Ü. P3 A 7 CLARK* A. P2 A 5 BEN LArlINÊ* A. 016 COLINET* C. Pc Blù P3 B 1 P2 Bll BERGER* R. PI A 3 COLLIN* G. 0 1 PI AIO CORBETT* J.O. PL10 BERNHARD* J. P3 A 7 COROIER* G. P4 A 5 BHAN* S. PI A 4 COURTOIS* A. P2 Ail bJARHA.l* j. PI AIO CUOMO* J.J. P2 B 8 BCACHNIK* R. 020 CYBULSKI* Z. PI A 7 BLAHA* H. •PI Alò O'HEURLE* F.M. P2 B 9 BLOCK* G. P2 A 3 DABKOWSKA* H. P2 A14 BLOCK, H. P4 A 1 DABKOWSKI* A. P2 A14 BÖLLER* H. PI A16 DAHN, J.R. 0 4 P3 A13 OAQU, J.N. P4 ä 8 BGMBIK* A. P3 A16 OE NOVION* C.H. PL 5 BCURJIER* 0. P4 A13 OEMANGEAT* C. P4 BIO BRGNGER* W. PL 1 DEMAZEAU* G* P3 A17 P4 B 7 DEPPE* P. PI 3 7 BROZEK* V, 0 8 OHAHRI* E. 019 P2 A 7 DI SALVO* F.J. 015 BUCHER* E. PI B 3 PI 3 4 P*t A12 Pi B 5 6UEHRER* W. PI B 1 DIETRICH* L.H. P4 A 6 BÜFFET* B. P3 Al 7 OORMANN* J.L. P3 Ail BURGER* J.P. P^ 8 8 DUCASTELLEj F. a 9 BUSCHQW* K.H.J. P2 BIO EDER* O.J. P4 B12 BUTZ* T. PI A 5 EIBLER* R. P2 B 2 PI A17 EIBLER* R. P3 A15 CABANEL* R. P2 A t* EL-BORAGY* M« P4 A 7 CALLENAS* A. P2 8 1 ÊLLNER* ». P4 A 7 CARABATOS, C. PZ 8 4 ENGELNANN* W. PI B12 ENGST*OÉM/ I. P3 815 GRUBER/ H. Pl A 9 ERICSSON* T. Pl P2 A 6 P3 Aa n6 GUERIN/ R. P* A15 P* B 5 P* M 6 EZZAdUIA/ H. a 3 GUITTARD» M. Pl A 8 FILIPPIDIS/ A. Pl Bll P3 A 3 FIORIANI/ 0. P3 A 1 GURIN/ V.N. P3 SI* FISCHER/ P. P* B14 HAAS/ C. P3 A12 FJELLVAG/ H. a t HAEGGSTROEM> L. Pl AIO 021 P3 B 5 P3 A <* P* B 5 P* A 8 HAGENMULLERf P. P3 A17 P* A 9 HAJEK, B. 0 8 P* B 3 P2 A 7 FLAHAUT/ J. Pl A 8 P2 A 8 Pl A15 HAMAR/ R. 018 P3 A 3 HAMAR-THIBAULT, S. 018 FLlYUU/ H. 023 HAMEDOUN, il. P3 Ail FLUCH, A. Pl B12 HARPER/ J.M.E. P2 B e FGISc/ J. 0 3 HARSTA* A. P* B 6 FRANCOIS/ M. P3 B13 HAUCKz J. 022 FRANTZ; C. ?Z A 2 HAUSER/ J.J. 011 FRANZcN/ 1.F, P* B 7 HENTZELL/ H.T.G. P2 B C FREEIUN/ A.J. PL * HERZIG/ P. P2 B 2 P2 8 5 HICT6R/ P. P2 Bll FRELTOFT/ T. P3 A 7 HIEBL/ K. P3 A18 FRIEüT/ J.M. 01* HILSCHERz G. P3 A15 FkUCHART/ D. P3 A 5 HOOEAU/ J.L. P3 310 P3 A 8 HOENLE/ W.- P* A 9 P3 AIO HOFMANN/ K* P* B 7 P3 B 3 HOLLECK/ H. a 7 P3 B 6 HOVESTREYDT, E. 017 % P3 8 8 HUANG/ M. P2 fl 5 P* A 2 HUGEL/ J. P2 B * P* A 3 IPSER» H. Pl 8 6 DA À L. r " A *t r H 0Q HA P* AIO JAKU80WSKI/ U. p* Ail FRUCHART/ R. 019 JANZ/ S.O. p* B 9 P3 B 6 JASIOLEK/ G. P2 Al* P3 B 7 JEITSCHKO/ W. P2 A 3 FURUJETH/ S. P* A 8 P* A 1 GAMSJAEGER/ H. Pl B12 P* A ô GANTOIS/ M. P2 Ail P4 Ail P3 Bll P4 Al* GARCIA CASADO/ P. Pl B 9 P* A17 bARCIA/ J* P3 A 5 P* B 1 P3 AIO P* B 2 GARDETTE/ M.F. 0 1 JELLINEK/ F. Pi A 3 GAST«LDI» L. Pl A 1 JEZIERSKI/ A. ?Z B13 Pl A 8 JOHANSSON/ B.Q. P2 B 8 GHEDIRA/ lit P3 BIO JOHANSSON/ L.I. ?Z 8 1 GONZALEZ; D. P3 A 5 JOUBERT/ J.C. P2 A * GOROCHOV/ 0. 0 3 KALDIS/ E. PL t GRESSIiR/ P. Pl B 8 KAREN/ P. 0 8 GRIN/ YU. P3 B * P2 A 7 GRÛÉS3INGER/ R. P3 Al* P2 A 8 P3 A15 KELLER-BESREST, P. 0 1 KHACHFI/ M. ?Z Ail MAOAR/ R. P4 A13 P3 Bll MAHY/ J. 0 6 KHAliEEH/ J. P4 BIO MALAMAN/ 3. P3 A 8 KHAN/ M.A. 023. P3 B 8 P4 BIO P3 B 9 KHRUSSANOVA/ H. P