Phase relationships and structural investigations in TmSe and alloys

(1)

HAL Id: jpa-00218916

https://hal.archives-ouvertes.fr/jpa-00218916

Submitted on 1 Jan 1979

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Phase relationships and structural investigations in TmSe and alloys

E. Kaldis, B. Fritzler, E. Jilek, A. Wisard

To cite this version:

E. Kaldis, B. Fritzler, E. Jilek, A. Wisard. Phase relationships and structural investigations in TmSe and alloys. Journal de Physique Colloques, 1979, 40 (C5), pp.C5-366-C5-369.

�10.1051/jphyscol:19795130�. �jpa-00218916�

(2)

Phase relationships and structural investigations in TmSe and alloys

E. Kaldis, B. Fritzler, E. Jilek and A. Wisard

Laboratorium fuer Festkoerperphysik ETH, CH-8093 Zurich, Switzerland

Résumé. — La valence du thulium varie dans le TmSe de + 2,71 à + 3,0 quand la non-stœchiométrie varie entre 1,04 > x > 0,91 (x = moles Tm/moles Se). Une variation de + 2,9 à + 2,0 est obtenue pour x variant entre 0 ^ x =S 1 dans le TmSe^^Te.,. Des transitions de phase induites par la pression montrent des instabilités de réseau pour 0,2 s£ x < 0,5. Dans le TmSe et le TmSej.^Te, ces instabilités sont accompagnées d'une augmentation considérable de lacunes (jusqu'à 10 %). Le diagramme de phase de Tm-Se a été examiné jusqu'à 2 300 °C et la chaleur de la réaction de TmSe avec HC1 a été étudiée en fonction de la stœchiométrie.

Abstract. — The valence of Tm can be controlled in TmSe between + 2.71 and + 3.0 by adjusting the non- stoichiometry between 1.04 ^ x > 0.91 (x = mole Tm/mole Se). Variation of the Tm valence between + 2.9 and + 2.0 is achieved in the mixed crystal system TmSej _x Tex by varying x in the range of 0 < x =S 1 ; pressure induced phase transitions show lattice instabilities in the 0.2 4 x < 0.5 range. In both TmSe and TmSe1_;[Tex

lattice instabilities are accompanied by strong increase of the vacancies (up to 10 %). The phase diagram of Tm-Se was investigated up to 2 300 °C and the heat of reaction of TmSe with HC1 has been studied as a function of the stoechiometry.

Most investigations of TmSe up to now have been made with inadequately characterised samples. For a systematic study of the valence fluctuation phenomena, the chemical parameters influencing the scattering of the lattice constants, reported in the literature [1], must be determined. Further the question of homogeneity and the phase relationships must be clarified.

In this work we have tried to give some first answers to these difficult questions. Further, we have investigated the controlled change of Tm-valency and tried to use calorimetric measurements for criteria of the stability of TmSe as a function of stoichiometry. A more detailed presentation appears elsewhere [2].

Figure 1 shows the dependence of the lattice cons- tant a of TmSe and GdSe on stoichiometry. The dramatic change of Aa/a = 1.7 % for TmSe, as compared to 0.4 % for GdSe, is due to the valence change of Tm. Extrapolation of the lattice constants of the RE-selenides shows that the valence changes from + 3.0 to + 2.71 and that the stoichiometric TmSe has an average valence of + 2.75. Density measurements [2] show that solution of an excess component in its sublattice takes place under formation of vacancies in the other sublattice. Similar assumptions have been made in the past for most RE-chalcogenides [2]. The shape of the curve of GdSe in figure 1 indicates clearly this behaviour. On the other hand, the absence of a decrease in the lattice constant of TmSe due to the formation of Se-vacancies in the Tm-rich region is possibly the result of increasing carrier concentration. The latter seems to decrease the average valence of Tm.

Fig. 1. — Lattice constant us composition in the nonstoichiometric range of TmSe and GdSe.

The results of our density measurements [2, 3]

can be interpreted with a high concentration both of Schottky- (approx. 1-2 % for all compositions) and nonstoichiometric-vacancies. The latter increase with the deviation from stoichiometry so that the total

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:19795130

(3)

PHASE RELATIONSHIPS AND STRUCTURAL INVESTIGATIONS IN TmSe AND ALLOYS C5-367

amount of vacancies reaches at the Se-rich phase boundary approx. 10

%

of the lattice sites. This extremely high concentration of vacancies should be taken into consideration when discussing the physical properties of TmSe.

Most important for the interpretation of the physical properties in connection to the valence fluctuation is the question of the homogeneity of the nonstoichiometric TmSe. As discussed in more detail elsewhere [2, 41 X-ray powder and single crystal investigations as well as electron diffraction indicate the homogeneity of the samples and show a correlation of vacancies in the Se-excess range. Scanning electron microscopy investigations reveal a small concentration of very fine particles (1 ,u) precipitated in the matrix of some crystals, possibly belonging to a high temperature phase (see further). An open question at present is if all concentrations in the nonstoichiometric range 0.91 d x d 1.04 can be actually synthesised.

The valence of Tm can be changed in a much wider range (between

+

^{2.9 and}

+

2.0) by alloying TmSe with semiconducting compounds having NaCl struc- ture and larger atomic radii in order to decrease the internal lattice pressure on the Tm cation. For substitution of the cation, alloys with the selenides of Sm2+, Eu2+ and Yb2+ or Ca2+, Sr2+, Ba2+ are suitable [5]

and for substitution of the anion alloys with TmTe.

All the investigated compositions of Tml-,Eu,Se show mixed crystal behaviour. Some properties of Tmo.5Euo.5Se and TmSe, -,Te, are discussed in another contribution [6]. Here we report in more detail on some other properties of the TmSel-,Te, system.

X-ray powder diagrams indicate mixed crystal formation. The phase diagram measurements existing up to now indicate a solid solution behaviour at least in the region 0.25 ^tx < 0.85. At low Te-concentrations, however, a sharp decrease of the free volume in the cell takes place which makes possible the existence of a miscibility gap or a new phase formation.

Figure 2 shows the dependence of the lattice constant on composition for TmSe,-,Te,. To draw the isovalent Vegard-law lines, estimated values for Tm2+Se (5.95

A)

and Tm3+Te (6.05

A)

were used. Lattice constants, determined from powder diagrams pro- duced from uncrushed single crystals (Gandolfi camera) with compositions 0.2

<

x

<

0.5, have much larger values than those obtained from powdered specimen. These results indicate that in this composition range the lattice becomes unstable and a pressure induced phase transition takes place during pulverisation under the pressure of the mortar. A relaxation of the lattice constant is hindered by the large surface of the powder. Contrary to what is normally expected, the valence of stoichiometric TmSe changes from

+

^{2.75 to}

+

2.9 by the dissolution of small amounts of TmTe (x

<

0.2). As lattice pressure probe the cell parameter is probably reflecting the effect of decreasing free cell volume. The colour

of the as grown material changes to the grey of the semiconducting TmTe at 0.3 < x < 0.4. Density measurements show a large concentration of Schottky defects throughout, this system and for x ⁼0.5 almost every 1/10 sites of the lattice remain vacant.

Similar effects are not observed in the TmSe-EuSe system. In Tmo.5Euo.5Se the concentration of Schot- tky pairslmole (2.87 x 10'') is approx. one order of magnitude smaller than both in TmSe (1.38 x loz2) and EuSe (1.17 x loz2). In the Eu-rich part, the valence of Tm, extrapolated from the lattice constants, remains equal to

+

^{2.0 up to}x ^z0.5.

6.40 f

[?I

LATTICE CONSTANT

6.30

Tm Se . Tm Te

/ ' A/'

5.70

k:An'

^A^DEBYE

4 A

Tm Se

-

^MOLE^O/O ^{Tm Te} ^{Tm Te}

Fig. 2. - Lattice constant vs TmTe-content in the System TmSe,-,Te,. The isovalent Vegard-law lines allow a quick esti- mation of the change in valence. The Gandolfi camera produces Debye-Scherrer patterns from single crystal specimen.

T o clarify the phase width of TmSe and the phase relationships in the Tm-Se system, we have started with the investigation of the phase diagram up to 2 300 OC by means of differential thermal analysis (Mettler Thermoanalyser, sealed W-crucibles, 4.5 mm diam), chemical analysis, microprobe and X-ray measurements. The unexpected phase relationships, the high temperatures, the small crucibles and the strains of the surface (pressure induced phase transitions) during grinding (which diminish the effectivity of metallography) make this investigation a formidable task. An attempt to interpret the results existing up to now is shown in figures 3a and 3b. Dotted lines denote only assumed phase equilibria. Grey surfaces indicate

(4)

Fig. 3a. - Phase diagram Tm-Se near the TmSe. Dotted lines are only suggested phase equilibria.

Fig. 3b. - Magnification of the high temperature part of figure 3a.

regions where up to now unreproducible DTA peaks appear, showing kinetically controlled effects (diffusion). The Tm-rich phase boundary is located at 1.04 < x ⁼mole Tm/mole Se < 1.05 and the Se- rich at 0.86 < x < 0.91. It is interesting that with increasing temperature several phase transformations of TmSe appear : for the Se-rich samples around 1 700 OC, for the Tm-rich samples around 1 100 OC and for the stoichiometric sample at both temperatures ; all samples show at 2 003 OC a phase transfor- mation. Their nature is at the moment unknown.

High-temperature susceptibility measurements [7] up to nearly 1 700 OC show increasing values of the magnetic moment as a function of temperature and a deflection at approx. 1 100 OC. The change of the susceptibility with the temperature is not linear.

The melting curve at 0.91

<

x < 1.04 can be interpreted as eutectic, the eutectic temperature being 2 020 OC and the eutectic composition x ⁼1.0. The intensity of the DTA signal at x ⁼1.0 is larger than that of x ⁼0.91. These results indicate that the homogeneity of the nonstoichiometric TmSe phase deteriorates itself with increasing temperature, possibly by means of the phase transformations. Accord- ing to the above, at least between 2 003 OC and 2 020 OC, two phases must exist in the range of 0.91 < x < 1.04 ; one with the composition x ⁼0.91 (corresponding to Tm3 'Se at room temperature) and the highest melting point found up to now in the Tm-Se phase diagram (2 03 1 OC), the other with x = 1.04 and m.p. 2 028 OC. This phase segregation may be of electronic or structural origin.

Knudsen evaporation experiments showed that TmSe evaporates incongruently, the arrows 1 and 2 indicating compositions achieved after 6 and 8 days of evaporation respectively. At 1 700 OC (arrow 2) the congruently evaporating composition is near x ⁼0.5 (Fig. 3a).

In order to investigate the relative stability of TmSe as a function of nonstoichiometry, we have studied the heat of reaction with 4 N HCl. The elementary Se resulting from this reaction remains undissolved.

Figure 4 shows the results. Due to lack of similar measurements in the literature (resulting from the difficulty to adjust the sample stoichiometry) only a tentative interpretation can be given at the moment, under the assumption that the chemistry of the reaction

mol Tm

-

^[GEl

Fig. 4. - Enthalpy b f reaction of TmSe with 4N HCI vs compo- sition. Up to 50 % changes appear in the nonstoichiometric region.

(5)

PHASE RELATIONSHIPS AND STRUCTURAL INVESTIGATIONS IN TmSe AND ALLOYS C5-369

is not strongly influenced by the nonstoichiometry.

Contrary to the heat of formation, the.less negative heat of reaction corresponds to the most stable composition. Therefore, one is tempted to argue that the most stable at room temperature is Tm3+Se (x = 0.91) ; stoichiometric Tmz,75+Se (x = 1.0) is less stable ; the pronounced instabilities at both sides of the stoichiometric composition could be due to the large number of vacancies ; at even larger devia- tions from stoichiometry the correlations of vacancies

(found by electron diffraction) may stabilize the lattice.

Concluding it must be pointed out that, as the above results show, the solid state chemistry of TmSe is very complex. Unfortunately before we can under- stand it completely, there is a great danger of misinter- preting the physical measurements. The fascinating field of the solid state research can be best explored by a close collaboration of solid state physics with solid state chemistry.

References

[I] B u c m , E., ANDRES, K., DISALVO, F. J., MAITA, J. P., GOS- sARD, A. C., COOPER, A. S., HULL, G. W. Jr, Phys. Rev.

B 11 (1975) 500.

[2] KALDIS, E., FRITZLER, B., PEELER, W., 2. Naturforsch. 34a (1979) 55.

[3] FRITZLER, B., KALDIS, E., PETELER, W., WISARD, A., Proceedings EUCHEM-Conference Chemistry of the Rare Earths p. 129, L. Niinisto Editor, May 1978, Helsinki Institute of Technology.

[4] BATLOGG, B., OTT, H. R., KALDIS, E., THONI, W., WACHTER, P., Phys. Rev. (in publication).

[5] KALDIS, E., FRITZLER, B., JILEK, E., ref. [3] p. 125.

161 BATLOGG, B., KALDIS, E., WACHTER, P. (this conference), J. Physique Colloq. 40 (1979) C5.

[7] MULLER, K., GUNTHERODT, H. J. (unpublished results).