POLARIZATION ANALYSIS OF THE MAGNETIC DIFFUSE SCATTERING OF Na3Cr2P3O12

(1)

HAL Id: jpa-00229123

https://hal.archives-ouvertes.fr/jpa-00229123

Submitted on 1 Jan 1988

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.

POLARIZATION ANALYSIS OF THE MAGNETIC

DIFFUSE SCATTERING OF Na3Cr2P3O12

N. Fanjat, O. Schaerpf, J. Soubeyroux, A. Dianoux, G. Lucazeau

To cite this version:

(2)

JOURNAL DE PHYSIQUE

Colloque C8, Supplement au no 12, Tome 49, decembre 1988

POLARIZATION ANALYSIS OF THE MAGNETIC DIFFUSE SCATTERING OF Na3cr2P3012

N. Fanjat (" 2), 0 . Schaerpf (2), J. L. Soubeyroux (2), A. J. Dianoux (2) and G. Lucazeau ( I )

(I) Univ. Paris-Nord, Av. J. B. C l h e n t , F-93430 Villetaneuse, France ( 2 ) Institut Laue-Langevin, 156X, F-38042 Grenoble, France

Abstract. - Neutron scattering by the superionic conductor Na3Cr2P3012 has magnetic contributions from the cr3+ ion. Crystallographic, integral and energy resolved measurements show antiferromagnetic order below 10 K, which is identified as magnetic by spin analysis.

Introduction

Neutron scattering on the sodium superionic conductor Na3Cr2P3012 (Nasicon) at high temperatures has two quasielastic contributions, one due to the con- ducting ion ~ a + and another due to the magnetic ion cr3+ [l-31. With a time of flight (T.O.F.) spectrometer like IN6 they cannot directly be separated. Sep- aration can only be obtained by a T.O.F. instrument with spin analysis like D7 or by temperature variation. Spin analysis also allows to identify the scattering as magnetic. We describe this separation and identify the magnetic contribution at 2 K and 150 K by spin analysis.

Experimental p a r t

INSTRUMENTS. - Data are collected on three instruments a t I.L.L.: diffractograms are measured on D1B and used to extract spin correlation functions [7] taking the difference between the diffractograms measured at 1.5 and 55 K. Quasielastic scattering is measured both on IN6 and D7. IN6 is used with the incident wavelength X = 5.1

A.

The T.O.F. spectrometer with polarization analysis D7 is used with the incident wavelength X = 4.82

A.

The neutrons are polarized along Z (vertical) and X and Y successively by spin rotating coils; 32 detectors are equipped with super- mirrors analysing polarized neutrons after back rota- tion in the Z direction. It is shown in 14-61 that it is possible to single out elastic and inelastic pararnag- netic scattering using the dependence of the polarization direction of the scattered beam on the scattering vector Q, given by P = -Q. (Q.Pincident). The pure paramagnetic scattering I,, can be determined from the spin-flip intensities I,, I,, I* recorded for the incident beam polarized in three orthogonal directions respectively X,

Y,

Z by

duration: 62 h). The T.O.F. analysis with polarization analysis is carried out at 2 K with a resolution of 300 peV (duration: 150 h) and at 150 K with a resolution of 400 peV (duration: 45 h). The spectra of IN6 give the total scattering (coherent, incoherent and magnetic), obtained at 1.5 and 175 K.

Results

a) D7. - Integral measurements show that the observed Bragg peaks are only of nuclear origin. In agree- ment with susceptibility measurements [3], the exis-

tence of an antiferromagnetic short range ordering at low temperatures (Fig. 1) is demonstrated by the pres- ence of a broad pure magnetic peak at Q = 0.9

A-'.

This maximum disappears above 20 K and is replaced by a scattered intensity having the same behaviour as the magnetic form factor of cr3+. This confirms that

-

there is an antiferromagnetic short range order below 10 K.

Fig. 1. - Integral magnetic scattering intensity Imag, in

arbitrary units, measured on D7 as a function of Q at three different temperatures: T = 2 K, 20 K, 200 K. Antifer- romagnetic short range order is progressively established when the temperature decreases. (Continuous lines are given as a guide for the eyes).

MEASUREMENTS. - We measure on D7 in the inte- T.O.F. measurements: T = 2 K : the T.0.F.- g a l mode and with polarization analysis the scatter- spectrum of figure 2b was obtained by applying ing of the sample at 2, 10, 20, 30, 50, 100, 200, 300 K, (1) for each channel. The continuous curve is and the necessary calibrations according to [6] (total a fit with a Gaussian curve including the factor

(3)

C8 - 200 JOURNAL DE PHYSIQUE

Fig. 2. - (a) IN6 S ( Q , w ) spectrum, in arbitrary units,

at T = 175 K and Q = 0.61 A-1 fitted with a broad

Gaussian (H.W.H.M. = 1.8 meV) and a narrow Lorentzian

(H.W.H.M, = 0.3 meV) (Continuous line). (b) Magnetic

scattering spectrum, in arbitrary units, obtained on D7 at

T = 2 K and Q = 1.0

A-l

with T.O.F. and spin po-

larization analysis and fitted with a Gaussian (H.W.H.M.

= 0.8 meV) (Continuous line).

w

/

(1

-

exp (-bw)) where

p

= 1

/

kT.

The H.W.H.M. of the fitted Gaussian is equal to 0.8 meV.

T = 150 K : the antiferromagnetic short range order disappears. The purely magnetic contribution can be fitted as at 2 K with a Gaussian with a H.W.H.M. of 1.1 meV, except for three angles 0.2

₍

<

Q

<

0.4

A-

where the H.W.H.M. is much smaller (0.2 to 0.3 meV). b) IN6. - T.O.F. measurements given in figure 2a show two quasielastic contributions, one which can be fitted with a Gaussian of H.W.H.M. 1.8 meV and one with a Lorentzian of H.W.H.M. 0.3 meV. This Lorentzian corresponds to the feature attributed in the previous high temperature measurements to magnetic scattering, while the Gaussian was identified as a background component [I-21. The present experiments show that both the broad Gaussian and the narrow Lorentzian are of magnetic origin. In addition, a weak inelastic feature at about 1.1 meV has been observed on both spectrometers and has been attributed to incoherent scattering at low temperature on D7 by the combination Ii,, = I ,

+

I ,

-

31,. To compare the in-

tegral and T.O.F. measurements of D7 and the results of IN6, figure 3 shows the variation of the integral of

Fig. 3. - Variation of the quasielastic intensity (Q.E.I.) (in

arbitrary units) versus Q of: a) ( m ) Gaussian used to fit the

purely magnetic spectra obtained on D7 at T = 2 K ; b)

( 0 ) sum of Lorentzian and Gaussian used to fit the spectra

obtained on IN6 at T = 1.5 K. (For a) and b) continuous

line is given as a guide for the eyes.)

the quasielastic scattering measured by T.O.F. on D7 (Fig. 3a) and on IN6 (Fig. 3b). We notice the same behaviour in each case indicating antiferromagnetic short range order.

c) DlB. - Two diffractograms are recorded at 1.5 and 55 K with the wavelennth 2.52

-

A.

The difference spectrum obtained between these two temperatures shows a modulated background without any Bragg peak, confirming the existence of a magnetic short range scattering at low temperature. The data are analysed using the spin correlation determination technique [7]. The distances between chromium atoms are fixed parameters obtained from a previous crystallographic determination [I]. The calculation gives access to the sign of the interactions between neigh- bouring magnetic ions and to the intensity of these interactions. In order to reproduce perfectIy the magnetic scattering five parameters, representing the magnetic interactions of five shells of chromium, are used. The Cr-Cr distances obtained are 4.47, 5.07, 5.69, 6.24 .-

and 8.71

A

with respectively the spin function (not normalized

-

530, 680,

-

660, 110 and 190. These data are consistent with an antiferromagnetic lattice of chromium atoms as expected for cr3+ oxygenated octahedra sharing a vertex.

Conclusion

Our experiments on DlB, IN6, D7 confirm that there is an antiferromagnetic short range order below 10 K and that the quasielastic signal observed is purely magnetic at low temperatures. The most suprising observation is that the quasielastic scattering at small Q' s is maintened at temperatures as high as 300 or 400 OC [I-21. This unexpected observation comes from the fact that Cr3+ ions are not directly coupled. We pJan to check its magnetic nature by a new forthcoming experiment on D7.

[I] Lucazeau, G., Barj, M., Soubeyroux, J. L., Dia- noux, A. J. and Delmas C., Solid State Ionics 18, 19 (1986) 959.

[2] Barj, M., Thesis Paris (1987).

[3] Beltranporter, D., Olazcuaga, R., 'Delmas, C., Cherkaoui, F., Brochu, R. and Le Flem, G., Rev. Chim. Miner. 17 (1980) 458.

[4] Moon, R. M., Riste, T. and Koehler, W. C., Phys.

Rev. 181 (1969) 920.

[5] Meizei, F. and Murani, A., J. Magn. Magn.

Mater. 14 (1979) 211.

[6] Schaerpf, O., I.A.E.A. Vienna (1985) p. 85. 171 Colombet, P., Danot, M. and Soubeyroux, J. L.,