Structures and magnetic properties of Fe1+xNb3-xSe 10-ySy series

HAL Id: jpa-00210440

https://hal.archives-ouvertes.fr/jpa-00210440

Submitted on 1 Jan 1987

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished 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.

Structures and magnetic properties of Fe1+xNb3-xSe 10-ySy series

A. Ben Salem, P. Molinie, A. Meerschaut

To cite this version:

A. Ben Salem, P. Molinie, A. Meerschaut. Structures and magnetic properties of Fe1+xNb3-xSe 10-ySy series. Journal de Physique, 1987, 48 (2), pp.277-284. �10.1051/jphys:01987004802027700�.

�jpa-00210440�

Structures and magnetic properties ^of Fe1+xNb3-xSe10-ySy ^series

A. Ben Salem, ^P. Molinie and A. Meerschaut

Laboratoire de Chimie des Solides, U.A. CNRS n° 279, Université de Nantes, 2, ^rue de la Houssinière, ⁴⁴⁰⁷²

Nantes Cedex, ^France

(Reçu ^{le 7} juillet 1986, révisé le 9 octobre, accepté le 13 octobre 1986)

Résumé.

Une étude structurale

rayons ^X

été entreprise ^sur des monocristaux de Fe1 +xNb3 -xSe10 ^et Fe1 +xNb3-xSe10-ySy. Les facteurs de confiance tendent vers

minimum pour les valeurs

0,25 ^et y ~ 1,4

accord

l’existence d’une non-st0153chiométrie précédemment annoncée. Nous mentionnons également ^les

résultats des mesures magnétiques ^et de R.P.E. ainsi qu’une ^tentative d’interprétation.

Abstract. 2014 X-ray structural studies have been performed

^both Fe1 +xNb3-xSe10 ^and Fe1 +xNb3 -xSe10- ySy ^single

crystals. ^R factors converge ^to minimum values when

0.25 and y ~ 1.4 which confirms the non-stoichiometry previously mentioned. We also report magnetic and E.S.R. measurements together ^with

^tentative explanation ^of

those results.

Classification

Physics ^Abstracts

72.15E

71.30

75.20E

1. Introduction.

MNb3Selo derivatives (with ^M

^{Fe, V,} Cr) ^{have been}

the subject ^{of numerous} studies with regard ^to charge density ^wave (C.D.W.) phenomena [1]. They all exhibit the FeNb3Seio structural type [2] which is characterized

by the presence of ^two distinct chains (Fig. ^la, b) :

-

a trigonal prismatic [NbSe3] chain similar to that

exhibiting the shortest (Se-Se) pairing ⁱⁿ NbSe3 [3]. ^It

has to be noticed that CDW phenomenon ^was assigned

to this kind of chain.

-

a double chain of edge-shared [ ^{(Fe, Nb)} Se6]

octahedra.

Nb and Fe atoms are randomly distributed within the octahedral chain. This feature gives ^an explanation ^of

the drastic resistivity rise observed at low temperatures.

Indeed, Hillenius et al. [4] suggested ^that this behaviour could be associated to an Anderson type localization in the _presence of a C.D.W.

X-ray scattering experiments [5] ^{gave evidence of a}

Peierls distortion with a value of q,

(0.,0.270(3),0.) ^near 140 K for FeNb3Selo. ^A

more recent study [6] indicated, ^{in addition to the} previous ^{ones, new diffuse} scattering ^{features in the}

form of rods with reduced wavevector components q2 = ( 0.5, 0.33, 1 ) . ^{Taking into account a possible}

non-stoichiometric _range as illustrated by ^{the modified}

Fel +xNb3 -xSelo formulation (with ^{0.25 x} 0.40),

the origin ^{of the} q2 ^{vector was assumed to} correspond

to a partial ordering ^of Fe/Nb ^{in an} occupancy ratio

Fig. ^la.

Projection ^{of the structure}

^{to the} (010) plane.

b

Perspective view of the structure showing ^octahedral

and trigonal prismatic ^chains.

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphys:01987004802027700

278

( 2l3 ^Fe-113 Nb ) leading ^{to a} tripling in the b direction.

This agreed very well with the average composition reported ^{above, but needed to} be ascertained by ^{an X-}

ray structural study. ^{This is} part ^{of this} report.

Besides, X-ray structural investigations ^are given ^for

the Fel +xNb3 -xSelo - ySy ^{series from a} single crystal study. Finally, magnetic measurements on these new

series performed ^{on a} Faraday ^{balance are} discussed. A tentative explanation ^is provided.

2. Experimental.

2.1. Fel,.,,Nb3 -xSelO REINVESTIGATION.

Starting

from previous ^data collection which was used to

determine the crystal ^structure [2], ^{we have done some}

least square refinements by varying ^the Fe/Nb ^ratios

within the octahedral chain only. ^{A minimum R value is}

reached for the ( Fe1.25Nb0.75 ) OCr ( Nb2 ) TP Selo ^for-

mulation as can be seen when plotting ^{R versus x where}

x is the iron content (Fig. 2) [2]. ^This corresponds ^to

the lowest limit of Fe concentration mentioned by Whangbo ^{et al.} (7) ^{who found} Fel +xNb3 -xSelo ^with

0.25 -- x -- 0.40. Of _course, all these refinements are

performed ^{under the same} conditions by varying only

the correlated (x,1- x ) occupancy values of both

Fe/Nb ^{atoms on} the octahedral sites. The positional

and thermal parameters (Beq ) ^are considered as

being identical for both species. ^It should be noted that

no change ^{in the} positional parameters is observed in

going ^{from x}

^{0.25 to x}

^{0.33 with the} exception ^of

the Beq values which slightly ^differ (Beq ^{= 1.61 A2 or}

1.54 A2 respectively). ^The experimental procedure ^for

data collection, along with the refinement conditions, ^is summarized in table I. The positional ^{and thermal} parameters ^are given ^{in table II ;} main interatomic distances are reported ^{in table III.}

Fig. ^2.

^The agreement ^{factor versus iron content. The} agreement factor is defined as

2.2 Fel +xNb3.-xSelO - ySy ^SERIES.

^{As already men-}

tioned in a recent study [8], ^{selenium atoms can be}

substituted for sulphur up ^to the highest composition

limit of Fel +xNb3 -xSegS2. ^{A single crystal structure}

determination has been performed. ^The experimental

conditions used to collect the data are summarized in table IV.

Referring ^to non-substituted compounds, ^{the iron}

content was fixed to the value Fe 1.25. In a first stage, the structure was refined ignoring ^the sulphur ^sub-

stitution. This revealed two Se sites (among ^{the five} ones) with Beq values much higher ^than the others. This

was the criterion for the Se-S occupancy variations which would make Ben ^{values more} homogeneous.

Thus, ^{Se-S rates} of occupancy ^were correlated to

(x,1- x ) relationships, whereas the positional ^and

thermal parameters ^were constrained to be identical.

This led to a minimum R value ( R

0.048) ^when considering ^the following conditions:

Thereby, ^the chemical formulation for this crystal corresponds ^to (Fe1.25Nb0.75) Nb2Se8.6S1.4. ^{At this}

point, ^no significant ^{feature was} observed in the Fourier difference _map.

Positional and thermal parameters ^are given ⁱⁿ

table V.

2. Discussion.

X-ray structural determination on Fel.25Nb2.75Selo-YSY

indicates that selenium atoms could only ^be substituted

on Sel and Se5 positions (Fig. 3). These sites corre-

Fig. ^3.

Projection ^{of the structure on to the} (010) place.

Table I.

Experimental conditions used to collect the data for Fe1.2SNb2.7SSe10.

Table II.

Positional and thermal parameters for Fe1.2SNb2.7sSe1o.

280

Table III.

*

Interatomic distances (A) ^{in the} trigonal prismatic ^chain

spond ^{to isolated X2 -} species. ^In contrast, Se2, ^Se3 and Se4 are not substituted by ^{S atoms. Se2 and Se3} belong ^to (Se2 ) 2 - ^{pairs and hence are} Se- ions. The

more electronegative ^{S atom} preferentially substitute in the Se2 - sites and therefore both Se2 and Se3 are left

pristine. ^{For the Se4} position, ^we should like here to note simply that S substitution for both Se4 and Se5 would have diminished the metal-metal repeat ^distance along the octahedral chains (b parameter) which would have been incommensurate with the metal-metal dis- tances along ^the trigonal prismatic ^chains.

Nevertheless, ^the question ^of why ^{Se4 atom does not} undergo substitution is a more open ^one which we will

investigate ^more fully ^{in a} later work.

The effect of the S substitution is then to compact slightly ^more the double octahedral chain and to increase weakly ^the interaction between trigonal pris-

matic chains (NbI-Selc ^{distances are reduced from}

2.742 [1] ^{to 2.695} [2] A).

But the main effect is to highly reinforce the disorder which, in turn, will affect the transport properties. ^This

is well demonstrated by drawing ^Ln R/Ro ^versus 103/T

for various sulphur ^{contents between} 0 , y , ² (Fig. 4)

and through the activation energies (see ^Table VI).

Fig. ^4.

Logarithm ^of normalized resistance versus 103/T

for Fe1,33Nb2.6T’ela - ysy

3. Magnetic measurements.

Magnetic susceptibility ^was measured with a Faraday

balance (resolution ^{of 1} ug) in the 4.2-300 K tempera-

ture range. ^{Data were} collected at external field of 7.8 kG. The susceptibility ^{values were} corrected from the diamagnetic contribution of the core electrons.

Figure 5 shows the magnetic susceptibility ^{as a func-}

tion of temperature ^for Fel.33Nb2.67Selo. ^{One can}

Table IV. - Experimental conditions used to collect the data for Fel +xNb3-xSe8.6sl.4*

Fig. ^5.

Magnetic susceptibility ^versus temperature ^for

Fel.33Nb2.6-,Selo.

observe that the magnetic susceptibility ^is essentially

constant at high temperatures ^{while it} rapidly ^increases

when the temperature ^decreases.

To account for the observed paramagnetic ^contri-

bution we give ^two possible interpretations. ^{Both are}

based on the presence of niobium in the octahedral chain.

3.1 ASSUMPTION 1.

The random distribution of Nb in the octahedral chain creates a random potential causing the localization of the conduction electrons.

In a previous ^article [9], ^we discussed the large

increase of resistivity ^as being ^{due to an Anderson}

localization.

282

Table V. - Positional and thermal parameters for Fe1.2sNb2.7SSeS.6S1.4.

Thus we have a system with localized states, and each Anderson localized state is occupied by ^a single ^electron

as far as it is close to the Fermi level due to Coulomb interaction between electrons in the same state. States

far below the Fermi surface are occupied by ^two

electrons and are non-magnetic. ^The magnetic suscepti- bility of the Anderson localized states system ^was calculated by Kobayashi ^{et ale} [10] ^{as :}

U

effective Coulomb interactions between electrons in the same state.

A comparison ^{with the} experimental result is shown in figure 6. As is evident from the figure, the best fit of the data is. ^{obtained for U = 240 K and}

N(0) = 1.2 [9].

3.2 ASSUMPTION 2.

The existence of Nb in the octahedral chain interrupts ^{the Fe} distribution and thus clusters of Fe atoms are formed as assumed by ^{Cava et}

al. [11] to account for the magnetic behaviour of

Fel.33Nb2.67Se10.

The paramagnetic contribution arises from isolated

spins ^or incompensated spins ^at the ends of broken Fe chains.

Table VI.

Variation of energy gap ^{as a} function of sulphur ^content. Fe1.33Nb2.67Sel0-xSx.

Fig. ^6.

Magnetic susceptibility ^versus temperature ^for

( Fe, Ta ) Nb2Selo. The solid line is calculated from the

equation given by Kobayashi ^{et ale}

Following ^{such a} hypothesis ^for Fel.33Nb2.67Se10,

Cava et al. have shown that 4 % of the total amount of Fe could be considered as paramagnetic.

Magnetic measurements are also performed ^{on the}

Fel.33Nb2.6-,Selo - YSY ^{series. Figure 7} shows the variation

of the susceptibility ^{as a} function of temperature ^for

various sulphur ^contents.

Fig. ^7.

Magnetic susceptibility ^versus temperature ^and sulphur ^{contents for} Fel.33Nb2.67Selo_,S,.

It appears that X increases ^{when the S content is}

increased. This is ^{true over} the whole _range of sulphur

content. It may be reasonably thought that the S substitution along the chains is another source of disorder leading ^{to an} enhancement of the localized character of the electrons along ^{the Nb chains.}

Since such a substitution is not expected ^{to increase}

the number of paramagnetic-like Fe, the observed increase in the paramagnetic contribution to the total

susceptibility ^{has thus to be} accounted for by ^consider- ing assumption ^{1 and not} assumption ^2.

We have plotted Ln (x - Xo) ^{versus Ln T}

(Xo

susceptibility ^{at room} temperature) ⁱⁿ figure ^8.

This plot ^reveals that the low temperature part ^{of the}

curves cannot be fitted properly according ^{to a} CT-1 law. A better fit of the data is obtained by ^a

C T- " ( a 1 ) law which is typical of disorder effect within low-dimensional conductors. a and C values for various S contents are given ^{in table VII. a values are}

all less than unity ⁱⁿ agreement with results reported by

Korin-Hamzic et ale [12]. ^{The C} coefficient, ^{which is} indicative of the amount of localized electrons, ^first increases and then decreases with the S concentration.

The decrease can be interpreted ^{in terms of re-}

combination of spins newly ^created. Indeed, ^{the in-}

crease of sulphur concentration leads to an enhance-

ment of the disorder which increases the amount of localized electrons and leads to new spins spatially

closed to those already existing. ^The antiferromagnetic exchange ^between spins ^{can become} strong enough ^to pair them and thus subtract them from the total

susceptibility ^{of the} material. The temperature ^below which the susceptibility is described by ^a C T- " law

corresponds ^to that where resistivity follows the re-

lation :

this corroborates assumptions ^1.

Fig. ^8.

Magnetic susceptibility

temperature ^and sulphur ^contents

Log-Log scale. Solid lines are fits to CT-a law.

Table VII.

Variation of ^a and C values (X ^oc CT- «)

versus sulphur ^content.

4. E.P.R. measurements.

E.P.R. measurements have been performed ^{at room} temperature ^on sulphur substituted phases, ^{as well as}

for Fel.33Nb2.67Selo. ^For the latter compound, ^a very broad line appears. ^For the selenium-sulphur ^substi-

tuted compounds, the line is narrower (Fig. 9).

An increasing disorder is expected ^to give ^{rise to a}

decrease of the relaxation time, ^{and thus to broaden} the peak ^{line as for a common metal;} however, ^the contrary is observed. Such a situation is similarly

encountered for irradiated organic conductors. For these compounds the results _{may be} interpretated using

both the Weger relationship and the model of one-

dimensional localization [13, 14].

284

Fig. ^9.

^The peak-to-peak ^{width versus} sulphur ^concen-

tration at room temperature.

From the model of localization, ^{it can be} predicted

that the frequency ^of transverse hopping should de-

crease when the defect concentration increases

The Weger relation indicates that the width of the

peak ^{varies as the inverse of} the transversal relaxation time. Then, ^{we have}

this explains ^the exponential decrease of the width in relation with the S content (y). ^{In this} example, ^an

additional term H has to be added.

E

0.037 eV and åHoo ^{= 1 600 G. E, which} represents the bandwidth, ^{has a value similar to} that obtained from transport measurements [8].

5. Conclusion.

The substituted Se-S phases ^have permitted ^{us to}

understand better the effect of disorder ^on the magnetic properties ^and subsequent consequences ^{on E.P.R.}

signal.

Acknowledgments.

We would like to thank Prof. Jean Rouxel, Stephen

Lee and Pierre Colombet for useful discussions and comments. This work was partly supported by ^{a NATO}

Research Grant n* 860438.

References

[1] MONCEAU, P., Electronic properties of quasi

dimensional compounds (D. ^Reidel Publishing Company) ^edited by ^P. Monceau, part ^{I and} II, 1985.

[2] CAVA, R. J., HINES, V. L., MIGHELL, ^{A. D. and} ROTH, ^R. S., Phys. ^{Rev. B 24} (6) (1981) ^3634.

MEERSCHAUT, A., GRESSIER, P., GUÉMAS, ^{L. and} ROUXEL, J., Mat. Res. Bull. 16 (1981) ^1035.

[3] MEERSCHAUT, ^{A. and} ROUXEL, J., J. Less Common Met. 39 (1975) ^197.

HODEAU, ^J. L., MAREZIO, M., ROUCAU, C., ^AYRO-

LES, R., MEERSCHAUT, A., ROUXEL, ^{J. and} MONCEAU, P., ^J. Phys. ^{C 11} (1978) ^4117.

[4] HILLENIUS, S. J. and COLEMAN, ^R. V., Phys. ^{Rev. B}

25 (4) (1982) ^2191.

[5] HILLENIUS, ^S. J., COLEMAN, ^R. V., FLEMING, ^{R. M.}

and CAVA, ^R. J., Phys. Rev. B 23 (1981) ^1567.

[6] MORET, R., POUGET, P., MEERSCHAUT, ^{A. and} GUÉMAS, L., ^J. Physique ^{Let. 44} (1983) ^L93-L97 (1983).

[7] WHANGBO, ^M. H., CAVA, ^R. J., DISALVO, ^{F. J. and} FLEMING, R. M., Solid State Commun 43 (4) (1982) ^277.

[8] ^BEN SALEM, A., KIKKAWA, S., MOLINIÉ, P., MEERSCHAUT, A. and ROUXEL, J., C.R. Heb.

Séan. Acad. Sci. 301 II (14) (1985) ^1003.

[9] ^BEN SALEM, A., MEERSCHAUT, A., SALVO, H., WANG, Z. Z., SAMBONGI, T., J. Physique ⁴⁵ (1984) ^771.

[10] KOBAYASHI, D., FUNAGAWA, Y., IKEHATA, ^{S. and} SASAKI, W., ^J. Phys. ^Soc. Jpn ⁴⁵ (1978) ^1276.

[11] CAVA, ^R. J., DI SALVO, ^F. J., EIBSCHUTZ, ^{M. and} WASZCZAK, J. V., Phys. ^{Rev. B 27} (1983) ^7412.

[12] KORIN-HAMZIC, B., MILJAK, ^{M. et} COOPER, ^J. R., Proceedings of the international Conference on

low-dimensional conductors, Boulder, 1981, ⁱⁿ

Mol. Cryst. Liq. Cryst. ⁸⁵ (1982) ^1567.

[13] SANQUER, M., ^Thèse (rapport CEA-R-15335) (1985).

[14] FORRO, L., BOUFFARD, ^{S. et} ZUPPIROLI, L., ^J.

Physique, Colloq. ⁴⁴ (1983) ^C3-927.