NEUTRON AND X-RAY SMALL ANGLE SCATTERING (S.A.S.) STUDY OF THE AMORPHOUS ALLOY Tb.25Cu.75

HAL Id: jpa-00220341

https://hal.archives-ouvertes.fr/jpa-00220341

Submitted on 1 Jan 1980

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.

NEUTRON AND X-RAY SMALL ANGLE

SCATTERING (S.A.S.) STUDY OF THE

AMORPHOUS ALLOY Tb.25Cu.75

B. Boucher

To cite this version:

JOURNAL DE PHYSIQUE CoZZoque C8, s u p p l h e n t au n o 8 , T m e 41, aoct 1980, page C8-135

NEUTRON AND X-RAY SMALL ANGLE SCATTERING (S ~ A . S rn ) STUDY OF THE AMORPHOUS ALLOY T b _-7s

B. Boucher

DPh.G. PSRM, Ome des Merisiers, CEN-SacZay, 91190 Gif-sur-Yvette, France.

The magnetic properties of amorphous alloys REx

M ( 4 . E . =heavy rare earths, Ef = Cu,Ag,Au) have

been widely studied [1,2,3,4,5]. They are of the speromagnetic type for x

>

0.33 and are mictomagne- tic for x

5

0.33. As the magnetic order is only short range, it seemed interesting to study these alloys by small angle scattering with X-rays and neutrons.

FJe chose alloys of Terbium and Copper for the follo- wing reasons : Tb and Cu have very different X-ray scattering amplitudes (~29,65) and similar neutron scattering amplitudes (0.76 x

10-"cm) .

Also the ato-

03

mic volume of Terbium ( ~ 3 3 A ) is almost three

"

3

times that of Copper (11.8 A ) and Cu is less absor- bant than Ag or Au. Tb alloys exhibit hish magnetic ordering temperatures and important moments in con- trast to the majority of other alloys of the same family. One inconvenience with Terbium, however, is the large (X-ray) fluorescence ( XCu).

In order to confirm some interpretations of S.A.S., we were obliged to determine some physical parame-

ters such as the density and porosityandto examine the sample with microscope. These results are also given here.

7 . S m p l e b a n d e x p W e n t d d e v i c e s

7 . P h Q - p j ? a h d ~ n : The alloys are prepared by sputte- ring on a substrate at 77IC.The targets are made in a levitation furnace with Tb and Cu having purities, respectively,99.99 and 99.96X. :Je obtained sheets of amorphous alloys of 1 x 7 cmand about23-40microns thick. These samples were analysed both chemically and with a microprobe. As we observed that the S.A.S. depend on the heat treatment, we also stu-

died's sample Tb 25C~.75 annealed under vacuum at 240'~ for 4 hours.

2 . X-Ray S.A.S. These measurements sere made at the "Laboratoire de MLtallurgie Physique" at Poitiers

using the device built by A. Maudon, M. Jaulin and

A.M. Flank [6,7]. This devices uses a convergent

linear beam

(A

Cu). The sample is in a furnace under a primary vacuum. TJe use a detectbr at linear loca- Lisation which allows simultaneous measurements to be made at angular intervals of 20". Ire were also able to obtain quickly the whole central scattering or a large part of Debye pattern. Also, the tempe- rature evolution of the sample can be precisely followed. 9e sere able to do measurements for k > 0.06

i-'

(k =

lgl

= 4r sinB/X).

3. FIeu;Dron S.A.S. We used several devices located at the I.L.L. (Srenoble).

-

D l which operates at 0.002 < k < 0.02

i-'

with

X

= 5.94

8,

or

X

= 11.77

8.

-

D17 :

X

= 7.65 and 0.01 ckcO.14

i-'

Both apparatus have a resolution AX/X-gX. A point source was used.

-

DIB :device built for diffraction with a multi-

0

detector,

X

= 2.52 A. Ifeasurements were carried out for 0 . 8 B_<40°, that is f o r k > 0.07

W-'.

The sample was composed of 40 p thick platelets stacked to form a sample of total thickness 0.75m.

2

The irradiated area was about 1 cm for Dl] and DI7 and 0.5 x 5 cm for DIB. This surface was perpendicu- lar to the indident beam. The measurements were carried out at different temperatures between 3 and 300 K. In all cases, corrections due to the back- ground and sample absorption were applied. The neu- tron energy was not analyzed. The measurements obtained with the different devices were normalised

0

together. With

X

= 11.77 A, we were able to measure the incident flux without using an attenuator or beam stop. He were thus able to-obtain absolute va- lues of the scattered intensities with 11.77

i,

and, by normalising, to obtain an order of magnitude of the absolute values for the spectra measured with the smaller wavelength.

JOURNAL DE PHYSIQUE C8-136

4 . Ofhe4 mcanuhmentn. Magnetic susceptibilities 1.7ere measured with a Faraday balance in a field of

*

3-4 kOe between 7 and 300 K.

The densities were determined by pycnometry with toluene [ C ] . The porosity was measured with Hg under-pressure [9] and also by the method of nitro- gen absorption at 77 K 1101. In these three cases,

0

**

the liquid penetrates into voids up to about 158. In addition we examined transverse cuttings by

***

transmission with an electron microscope

.

The surEace was observed with a scanning microscope by

3. Luzet at the CNRS (Lab. of Bellevue) in collabo- ration with J.C. !!artin.

1) The chemical analysis gave in atomic fraction H 4%, N 3%, 0 1.5% and C 0.8%. Microprobe ana-

**it;*

lysis shows a reduction of the R.E. content in comparison with the target. The average composition is Tb.225C~-775. The R.E. reduction tends to increase between the beginning and the end of the preparation. But this variation

cannot significantly change the following results. The measurements also showed Che presence of Argon

( - 2 7 ) .

2 ) X-iiay S.A.S. show a large maximum at about

1: = 0.17 (fig. I ) . ,ken the sample temperature increases (-10°C per minute)the pattern does not change up to 200°C. Above 200°C an additional cen- tral scattering irreversibly occurs. $,%en the tem- perature increases, the scattering becomes narrower and the profile of the initial loti temperature scat- tering reappears (fig.2). The first indications of crystallisation occur at about 360°C.

3) The ratio of the scattering amplitudes are not the same for X-rays and for neutrons. However, the neutron and X-ray S.A.S. measurements are similar for k > 0.09 ;-I (fig.3) The neutron pattern is a little broader due to instrumental effects (AX/?,

,

divergence effect).

*

I!easurements done by 11.G. Clerc (DPh.SP21-CEMS)

**

lleasurements done by Messrs. Yvars, Schnedecker Laloux, Rasneur (DGL-CEMS)

***

Examined by Mr. Maire and Mrs. Auge (DPC-CENS)

****

The analysis were carried out by Mrs.Zemkoff (DYECM-CENS) and f!rs

.

C. Jehanno (DPh

.

SEP CENS) independanmently on different sauples. The results: are in goodagreement.

'Fiq

.2. X-nag6 ncattening doh -the nmpLe. heated at

A-

di&mentzt'tempma;t~en

For weak values of k (k <0.08

id')

we observe in- tense scattering, which varies abruptly (approxima-

-3

tely as k ) . Indeed at low temperatures, a magne- tic scattering is superimposed on the nuclear scattering. We obtain the magnetic intensities from the difference between the patterns (figs .4,5). The magnetic order subsists up to 50 K.

4) The magnetic susceptibility (fig.6) obeys the Curie !.Teiss law for T >6OI< ; neffi 9.89

ug

for

= 9.72 yg for the ~ b ~ ' free ion) ;

x=0.25 (neEf

8 = 23 K. The measurements olan annealed sample showed the same behavior and lead to the same va- lues of the constants.

prepared using the same method. The values obtained were:first sample 9.48, second sample 9.32 third

sample : 9.09,9.17,9.70. The measurements are sensi- tive to degasing time. Despite the dispersion ofthe results, we are obliged to accept that the alloys areseen to be very dense. The values are in fact higher- than for Cu (8.92) or Tb (8.23) and also for every crystallized adl'lmx of TbCu whose density has been reported in the literature (:< 9).

..

These unex- pected results are being fmther-examined.

F i g . 4 . F,lagn&CLc intenniLLen u e m u k aA diddehent

-

Zempehatwe

.

expenirreW- value

-

valueb c d c u t t e d

at$te,k

( 2 I

.

7Je also observed a high porosity (> lp), but there are not many voids smaller than 1p.Hith Hg, the mea- surements gave a 2X porosity. The N absorption

2 0

leads to 4% of voids with sizes from 500 to 2000 A and shows that no smaller voids are present. 1,Je did not observe any voids with the electron microscope

on the transversal cutting. The scanning microscope shows surface which is in contact with the subtrate and relatively rough, whereas the other face is very smooth.

FLg.6. Recipnocd ~ w c e p ~ b U y

111. gincunhion

It is possible to accounf f0,r scattqing for

0

k=O.17 A by assuming very dense spheres being embed- ded in a less dense amorphous matrix., the spheres being close enough together so that there is scatte- ring interference. Vith a rough model in which the presence probability of a sphere is 0 from the ori- gin to a distance R and 1 for r > Ro, we obtain the equation [ 1 1 1 :

L L

k R ~ r

3

I(k) = A exp

-

-

1

I

-3

(sinkZo

-

kRocoskRo)] (I)

- k Po

were R is the gyration radius. With RrN8 A and

G o 2

Ro-20 A, good agreement is obtained (fig.1). So the alloy can be considered as being made up to dense spheres separated by an average distance of

0

40 A. But we do not know at present what is the na- ture of these spheres, for example, local fluctua- tion of the concentration Cu/Tb or fluctuation of the number of atoms per unit voluue.

JOURNAL DE PHYSIQUE

The scattering at very low angles is connected to heterogeneities of about 1000

i,

that is to say, roughly the same size as the voids. If we attribute this scattering to the porosity, we can write for k=O, do/dw = TIN^] ,' where n is the number of voids,

V their size, N the number of atoms per unit volume and b the scattering amplitude for these atoms. It is impossible to obtain do/& for k=O, but we can set a lower limiting value by taking k = 0.0024

i-'

-1

which gives do/dw=672 cm /Ster. So rue are able to 2

set a lower value of nV

.

The porosity measurements 603

give nV. Thus we have V

-

9.10 P. as a 200

i.

This lower limiting value is in good agreement with the value determined by porosimetry. It seems reasona- ble to attribute this scattering to the porosity. When we heat the sample a new scattering appears indicating that some order appears in small domains. Possibly in this case, the impurities, and in par-

ticular hydrogen, play a role. We know that at 200°c hydrogen reacts with rare earth metals and it could lead to anew distribution of hydrogen atoms. The scattering given by the sample before heating and which subsists up to the crystallization tem- perature could arise in the same way. It is to be noted that these "domains" do not produce Bragg peaks and do not modify the susceptibility (T>7K). The value of neff indicate that the Tb ion is in 3+ state and that the composition is not far from Tb.25C~a75. 9e have 0

-

23K, so we would expect a magnetic ordering temperature slightly'lower. But the scattering measurements clearly show that a short range order appears at about 508 and there is no indication that this order is noticeably modi- fied below 20K. This behavior @an suggest a large relaxation time [I2

1.

It is also possible to account for the scattering intensity (figs.4,5) by the empirical formula :

2 2 I (k) = A exp

-

18,50 k exp

-

0.00233 T (2)

t1

The difficulty to obtain precisely the background (fig.3) does not permit to obtain an absolute value of

A'

and thus of the magnetization. By considering the Fourier inverse, we have 'the correlation func- tion between local magnetizations responsible for for the scattering (fig.7). So we establish that these magnetizations are correlated over about 201; this size is above that of statistical agregates [51 and is also visible for values much higher than

that corresponding to

.

However this distance is of the same order of magnitude as the diameter of

0

dense spheres (16 A). It is not impossible that both the phenomena are correlated.

Neutron S.A.S. has shown that magnetic short range ordering occurs at temperatures much higher than the asymptotic Curie temperature and has enabled the correlation length of the local magnetization to be found

.

However, a microscopic atomic struc ture does existsand is not yet fully understood. It corresponds to a dense stacking of atoms.

We thank all those mentioned in the text who have helped with the measurements and ?I. Luzet who prepared the samples. We are very gratefully to Messrs.P.Convert, A.Naudon,M.Rothfor help with S.A.S. measurements and in the interpretation of the results. We also thank Mr. G. Jehanno for useful discussions.

[l] Neil Heiman, N. Kazawa, IBM Res. Lab. San Jose Res.Rep.RJ2148 (29561)12.12.77 Sol. State Phys. [21 B. Boucher, IEEE Trans. Xagn. U G G , n05

Sept. 1977 p.1601.

131 B. Boucher, Phys. Stat. Sol. (a)42,K175 (1977) [4 1 S. von Molnar

,

C

.

N. Guy

,

B. J

.

Gambino, T. 3. Mc

Guire, Proc. Inter. Conf. t4agn. Munich 1979.

[51 K. Pappa ThSse 3e cycle,Paris 1979.

[61 A. Naudon, M. Jaulin, Rev. Phys. Appl.3 (1968) 152.

171 A.M. Flank, ThSse 3e cycle, Poitiers(l976) [8] P. Mottet, M. Yvars, Ph. Gerlinger, Rapport

C E A - D P C / S P C A / ~ ~ ~ ~ ~ du 18-12-1968. [9] G. Schnedecker, DGI/SEPCP/SSDP. CEA [10]B. Rasneur, Bull. Soc-Fr. de nO1O1,

Oct-Dec.1973.

[ll]J. Allain, ThSse 18 juin 1979, Poitiers [ 12]Boucher, B. Barbara, J. Phys. F

2

nO1