High-pressure magnetization of the heavy-Fermion system CeRu2Si 2 .

HAL Id: jpa-00210835

https://hal.archives-ouvertes.fr/jpa-00210835

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.

High-pressure magnetization of the heavy-Fermion system CeRu2Si 2 .

J. Voiron, J.-M. Mignot, P. Lejay, P. Haen, J. Flouquet

To cite this version:

J. Voiron, J.-M. Mignot, P. Lejay, P. Haen, J. Flouquet. High-pressure magnetization of the heavy-Fermion system CeRu2Si 2 .. Journal de Physique, 1988, 49 (9), pp.1555-1560.

�10.1051/jphys:019880049090155500�. �jpa-00210835�

High-pressure magnetization ^{of the} heavy-Fermion system CeRu2Si2 .

J. Voiron (1), ^J.-M. Mignot, ^P. Lejay, P. Haen and J. Flouquet

(1) Laboratoire Louis Néel (*), CNRS, ^B.P. 166X, 38042 Grenoble Cedex, ^France CRTBT (*), CNRS, ^B.P. 166X, 38042 Grenoble Cedex, ^France

(Requ ^{le 29}

^1988, accepté ^{le 24 mai} 1988)

Résumé.

L’aimantation de CeRu2Si2 monocristallin à basse température

été mesurée

pression hydrostatique jusqu’à 6 kbar. Les résultats révèlent

décroissance extrêmement rapide ^{de la} susceptibilité magnétique ^~~ (T ~ 0) pour H suivant l’axe _c, à raison d’environ 2014 17 % par kilobar. On ^constate

forte corrélation (~~ (0 ) Tmax

cst.) ^entre la variation

pression ^{de la} susceptibilité ^{à T}

^{0 et la} température ^du

maximum de susceptibilité. ^Les données pour ~~ (T) ^T peuvent ^être représentées par

courbe unique rapportée ^à

variable réduite T/Ts (P),

^accord

les théories Kondo à

impureté. L’anisotropie magnétique ~~ ~ ^décroît

s’y attend pour

système s’approchant ^du régime de valence intermédiaire, mais d’autres explications ^sont également envisagées. Les résultats sont comparés ^{à d’autres}

études du même matériau

haute pression ^ainsi qu’aux effets observés dans certains de

alliages.

Abstract.

The low-temperature magnetization ^of single-crystal CeRu2Si2 ^{has been} investigated ^under hydrostatic pressures up ^to 6 kbar. The results reveal

extremely rapid decrease of the magnetic susceptibility ~~ (T ~0) for ^H applied along ^{the c-axis, at}

^rate of about 201417% per kbar. A strong correlation (~~ (0) Tmax ~ const. ) is found between the pressure-variations ^{of the} susceptibility ^{at T}

^{0 and the} temperature ^{of the} susceptibility maximum. The data for _~~ (T) ^T

^be represented by

single

^{in terms}

of the reduced variable T/Ts (P), in accordance with single-impurity Kondo theories. The magnetic anisotropy ~~/~ decreases,

expected ^for

system approaching the mixed-valence regime, but alternative

explanations

also considered. The results

compared ^{to other} high-pressure ^studies

^the

^material

and to the effects produced by alloying.

Classification

Physics ^Abstracts

75.30M - 62.50

75.40C

75.30G

1. Introduction.

The cerium-based ternary ^silicides [1] belonging ^to

the tetragonal ThCr2Si2 ^{structure exhibit a} large variety ^of low-temperature behaviours, ^{which in-}

clude magnetic ordering (CeRh2Si2) [2], ^{mixed val-}

ence (CeOs2 Si2 ) [3], ^and heavy-fermion (HF) super-

conductivity (CeCu2Si2) [4]. ^The compound ^with

ruthenium CeRu2Si2 ^{can be} categorized ^{as a}

«

medium-heavy-fermion

material from the value of its linear specific-heat coefficient y=

350 mJ mole - K- 2 [5, 6]. ^{This system was} reported

to exhibit quite remarkable magnetic properties [7].

Its susceptibility ^is strongly anisotropic depending

on whether the applied field is oriented parallel ^or

normal to the c-axis. In the former case, XII approxi- mately obeys ^{a Curie law} above - 70 K, then reaches

(*) Laboratoire associd a l’Universit6 Joseph ^Fourier,

Grenoble.

a weak maximum around Tmax = 10 ^{K and for}

T-+ 0 tends to a high ^value (,y 0 11 == 3.5 10-2 emu/mole Ce), characteristic of the correlated

Fermi-liquid ^state. On the other hand, ^the suscepti- bility ^{for H _L c assumes} much smaller values, ^{and in}

a X -1 ^{vs. T} plot, ^{exhibits a} Curie-Weiss behaviour with a large negative 0 ^{below room} temperature, followed by ^a pronounced ^{downward curvature}

around - 100 K. The anisotropy ratio X / X J.. ^in-

creases from - 2.5 at 200 K to 13-15 at 4.2 K, ^{with a} faint maximum near T,,ax-

Whereas the magnetic ^isotherms M (H ) ^{are essen-} tially ^{linear with} applied field down to - 10 K, positive curvatures gradually develop ^{at lower tem-} perature ^for Hllc, leading ^{to a} steplike anomaly

around HM = 8 T ^at T = 1.5 K. This behaviour,

which is suggestive ^{of some kind of} metamagnetism

is further associated with a sharp positive peak ^{in the} magnetoresistance [7]. ^In contrast, X1 remains field

independent up ^to 18 T and _p (H ^-L c ) ^increases only

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphys:019880049090155500

1556

weakly ^{with the} applied field. Recent neutron-dif- fraction experiments [8] provided ^no evidence for an

ordered magnetic ^{structure in this} compound, ^but

rather indicated short-range antiferromagnetic (AF)

correlations with moments along the c-axis to exist among neighbouring cerium sites up ^to about 60 K.

It is interesting ^{to note} that the maximum in

p (Hllc ) persists ^{over the same} temperature range.

These results point ^{to a} rather unconventional behaviour of the Fermi-liquid ^state in the presence of an external field. From what ^we know about this

compound, ^{it seems} unlikely ^that they ^{can be} explained by simply invoking ^the quenching ^{of the}

AF couplings by ^the applied ^{field : even at H}

0,

the system ^is already ^{in a} complicated ^{state where}

the normal development ^{of a} long-range ^structure

has been inhibited by demagnetizing processes (Kon-

do effect, ^s-f mixing, etc.). Furthermore, it is well known that crystalline anisotropy plays ^an important

role in defining ^the ground ^state of non-cubic cerium

compounds. ^An important point is that very similar effects have also been reported ^for UPt3 [9] (though

at much higher fields), ^{so that} they appear ^to characterize a new crossover situation which has not

yet been considered in current theories but might

occur in a larger class of HF systems.

It is therefore an urgent ^{task to} better document the relationship existing between the various Fermi-

liquid parameters ^{at H}

0 and the way the system responds ^{to an external} field. This was undertaken in the present study by measuring ^the magnetization ^of CeRU2Si2 ^{as a} function of pressure up ^to 6 kbar. The results are presented ^{in the} following and character- istic Grüneisen parameters ^{are derived.} Anisotropy

effects and their pressure dependence will receive

particular attention since the measurements were

performed ^{on a} single-crystalline specimen ^for Hllc and H L c. Comparisons will also be made with data obtained for other quantities (p (P, T), ^V (H))

and in the case of alloy systems.

2. Experiments.

The single crystal used for this study ^{was cleaved}

from an ingot grown by ^a Czochralski technique ^{in a}

triarc furnace, ^under purified argon atmosphere.

The constituents were high-purity ^elements (Ce :

99.99 %, Ru : 99.999 %, ^{Si :}

99.999 %). The crys- tal presented large cleavage planes ^{which were}

checked by x-ray diffraction ^to be normal to the c-

direction. In order to allow measurements with H either parallel ^or perpendicular ^{to the} tetragonal axis, additional plane ^{faces were cut} perpendicular

to the basal plane. ^The parallelepiped-shaped speci-

men, with dimensions of approximately ^{0.4 x 0.5 x}

0.5 mm3, ^{was the one} used for the magnetostriction

measurements in a capacitance ^cell reported ⁱⁿ

reference [10]. ^The magnetization data for this

sample (Fig. ^{2 in Ref.} [10]) agreed ^with previous results, including ^the critical field (HM ^{= 8.3 and}

7.9 T at 4.2 and 1.3 K respectively), and the ani- sotropy ^ratio X II /X 1 ^{= 14 at 4.2 K.}

Hydrostatic pressures up ^to 7 kbar were generated

in a copper-beryllium ^cell using compressed ^helium

as a pressure-transmitting medium. The pressure inside the experimental ^{chamber was} measured with

a calibrated thermally-compensated manganin gauge located in the pressure generator which is connected to the cell by ^a capillary tube. At low temperatures, the helium gas solidifies and pressure corrections

were computed using ^the phase diagram ^{of helium.}

Magnetic fields up ^to 7 T were produced by ^a superconducting ^{solenoid at} the Service National des

Champs ^{Intenses, and the} magnetization ^was

measured by the conventional extraction technique.

The background contribution from the pressure cell

was determined in a separate experiment ^{and sub-}

tracted from the data.

3. Results.

The magnetization ^{of our} single crystal ^{was measured}

at various pressures with the field applied ^either parallel ^{or normal to} the c-axis. The corresponding

curves for T

4.2 K are shown in figure 1. Pressure is seen to produce ^a drastic decrease of M(H) ^for

both field directions.

Let us focus first on the case where H is parallel ^to

the easy axis (H// c). ^{In this case, the} magnetization

in 6 T is reduced by ^a factor of 3 with only ^{6 kbar. At}

ambient pressure, ^a positive ^{curvature of} M(H) ^is observed, ⁱⁿ agreement ^with previous data, ^{due to} the approach ^{of the} steplike transition (pronounced

maximum of 8M/8N) ^at HM = 8 T. At 6 kbar, ^on

the other hand, ^no deviation from a straight ^{line can}

be detected any ^more. This difference indicates that the transition has shifted to higher magnetic fields,

as already inferred from previous magnetoresistance

measurements [11] : in the latter study, HM ^was

found to increase to 18 T for P

5 kbar.

The low-field susceptibility XII deduced from the linear part ^{of the} magnetic ^{isotherms at 4.2 K is} displayed ⁱⁿ figure ^{2 as a} function of pressure.

Relying ^{on the} shape ^of the XII ^{vs. T curves} (see below), ^{we will assume} these values to approximate

the Pauli susceptibility Xo II of the compound ^at

T

0. The data can be represented quite accurately by ^an exponential ^{law with a rate} of variation of

rjj(X)i ^a ln XoII /aP ^{= - 171 Mbar-1.}

Applying the field in the basal plane results in much weaker effects. Indeed, ^{it was noted} previously

that the heavy-fermion character of CeRu2Si2 ^mani-

fests itself in the magnetic properties (large ^Pauli susceptibility X o, maximum in X (T),

metamagnet- ism ») predominantly for Hllc, ^whereas weak, ^struc-

tureless variations are observed for H 1.. c [7]. ^How-

Fig. 1.

Magnetization

^of single-crystal CeRU2S’2 ^{at T}

^{4.2 K} under pressure : (a) ^field applied along ^the tetragonal c-axis ; (b) field normal to the c-axis.

Fig. ^2.

^Pressure dependence of the low-field suscepti- bility ^of CeRu2Si2 ^{at 4.2 K} (Hllc ^and H ..L c) ^{and 49 K}

(HI! c).

ever, the point here is that even the relative variation of X , is comparatively ^small (r Ix) = ^a In X ..L / aP

-

70 Mbar-l). Consequently, ^the magnetic ^ani- sotropy X 11 /X -L ^{decreases rapidly from 14 at P}

⁰

to 8 at 6 kbar (Fig. 3). The non-4f terms in

Fig. ^3.

Pressure decrease of the magnetic anisotropy (xll Ix, ) ^of CeRu2Si2 ^{at 4.2 K.}

x 0’ ^as estimated from measurements on LaRU2Si2 [12], ^{would represent} less than one percent ^{of the} total susceptibility, ^{and no} attempt ^{was thus made to} subtract them out.

The temperature dependence ^{of the low-field} susceptibility X II for various pressures is presented ⁱⁿ figure ^{4. One sees} that the decrease of Xo is ^corre- lated with an increase of the temperature Ta,, ^{of the} susceptibility maximum. This behaviour suggests ^to

cast the pressure-dependence of XII ^{in terms of a}

reduced temperature ^variable TS (P ) by writing ^the product X II T as :

Fig. ^4.

Temperature ^variation of the initial suscepti- bility ^of CeRU2S’2 parallel ^to the c-axis at various press-

ures :

closed circles : experiments ; solid lines : fit obtained

by scaling ^{the upper}

with X, ^T

f [T/TS (P ) I (see text).

Such a relation was established by ^Krishna- Murthy et ^al. [13] ^{for a} spin 1 ₂ ^{Kondo impurity, in}

which case the scaling temperature ^is simply ^the

1558

Kondo temperature TK. ^{A nice} experimental ^illus-

tration was found in the alloys ^Ce (Inl - xSnx)3 ^{for a}

wide interval of characteristic energies (50 ^{to 200} K) [14]. ^{Here we} will follow the analysis proposed ^for

the latter system and define Ts ^somewhat arbitrarily

as the temperature ^for which X II T/CM = 2 , cm ²

being ^{the Curie constant of the} (J

5/2) multiplet.

Although ^the change ^{in the} susceptibility with pres-

sure is quite substantial in the present ^{case, the} scaling ^law applies surprisingly ^{well over the entire}

range of parameters investigated (Fig. 5). ^Minor

deviations occurring ^{at the} higher temperatures

(T >. ³⁵ K) may indicate ^a breakdown of the scaling regime, ^but they ^could also be due to imprecisions ⁱⁿ

the background correction (paragraph 2). ^{Note that}

the quality of the fit would degrade only slightly ^if

one chose T max (P) ^{instead of} Ts (P ) ^{as the} scaling parameter.

Fig. ^5.

Temperature dependence of XII ^T

^function

of the reduced variable T/Ts.

The value of Ts changes rapidly with pressure

(Fig. 6) ^{and, as} expected ^from equation (1), ^exhibits

an exponential variation with about the same rate

(Fil(T,S)= ^{a In} TsIaP

^{+ 167} Mbar-1) ^{as the initial}

susceptibility. Accordingly, ^the product

Fig. ^6.

Comparison of the pressure variations of the Pauli susceptibility xll (0) ^{and the} scaling parameter Ts. ^Note the reversed logarithmic scale for Xll (0).

xo II (P ) Ts(P), ^which represents the initial slope ^of

the curves in figure 5, ^remains pressure-indepen-

dent. An interesting consequence is that, knowing

one of the yll (T) ^curves (e.g. ^{for P}

0), ^{it is} possible ^to predict ^the temperature dependence ^of

X II , at least up ^to 6 kbar, simply by interpolating ^the corresponding ^{value of} Ts from the upper ^curve in

figure 6. The result is represented by solid lines in

figure ^{4 and} yields perfect agreement with the data.

As noted above, the absolute value of Ts ^{is not} meaningful, ^but its variation is supposed ^{to be} approximately proportional ^{to that of} TK. ^Thus, assuming ^an initial value of TK (P

0 ) = 24 K ^as reported in reference [5], ^{we find that} TK (6 kbar)

should already be of the order of 75 K.

The present ^{results can be} compared ^{to those}

obtained by substituting ^{Os for Ru} [3, 15]. Typically, replacing 5 % of the transition-metal atoms has ^an effect comparable ^{to that of} applying - ^{6 kbar :}

Xo decreases by ^a factor of about 3 and Tmax goes up

by approximately ^{the same amount. The} X T ^{vs. T}

curves also scale with some concentration-depen-

dent « spin-fluctuation » temperature T*sf. ^{In that}

case however, ^{it was} argued [15] ^{that Os} substitution did not really produce ^a

chemical pressure », but rather affected the cla ratio of the material, thereby changing ^the degree ^of overlap of the Ce orbitals with those of their Si and/or Ru neighbours. ^These

considerations may be of interest ^even for pure

CeRU2Si2 since the linear compressibilities Ka and K c ^{are most} likely anisotropic.

4. Discussion.

The major ^{outcome of the} present experiments ^{is the}

extreme sensitivity ^{of the} low-temperature ^Fermi- liquid ^{state to} moderate external pressures. From the values obtained for Fll(x) ^and rll(Ts), ^{we can infer}

that the characteristic temperature ^{T* associated} with spin fluctuations increases by ^{= 17 %} per kbar.

Although ^the compressibility

^of CeRU2Si2 ^{is not}

yet ^known precisely, ^a reasonable estimate of

10- 6 bar- leads to an exceedingly large value of the electronic Gruneisen parameter !J G II

rll(T*)

f-- 170, which is among the largest ^ever observed for any rare-earth compound. This value is fully ^consist-

ent with that deduced from an analysis of the low- temperature T2 terms in the resistivity (Fll (A2)

168 Mbar-1) [11, 16]. ^{A similar} estimate, f2 G

+ 180, ^{was also obtained} in reference [17] by fitting

a Kondo model to specific-heat ^{data for} polycrystals

in which chemical pressure (either positive ^or nega-

tive) ^was produced by substituting ^Ce by ^{Y or La.}

Grüneisen parameters ^{in excess of 100 are} quite exceptional, ^even among HF systems. ^In UPt3 ^for

instance, whose properties ^{are in some} respects similar to those of the present system

( y

^{420 mJ} mole -1 K- 2 [9, 18]), !JG ^is only ^60.

CeAl3, another well-known HF system does exhibit such large ^values (I nG I ^{= 200} [19]), but in that case, the negative sign of the thermal expansion

below 1 K [20] implies ^{that the Gruneisen} coefficient is initially negative, possibly ^{due to some kind of} magnetic ^order [21]. ^If pressure ^is applied, ^a positive n G ^is readily recovered, but its absolute value then decreases rapidly ^{as the} characteristic temperature increases [22].

Experimentally [23], giant Gruneisen coefficients

are the signature ^{of the} heavy-fermion ^{state, whereas}

their reduction usually ^{indicates a crossover into the}

intermediate-valence (IV) regime. ^The puzzling ^issue

in the case of CeRu2Si2 ^is that f2G remains almost unaffected while TK ^{increases to values} approaching

those observed in IV materials (resistivity ^measure-

ments, however, ^indicate that the pressure coef- ficient begins ^to decrease above 6 kbar [16]). ^The validity ^of T/TS (P ) scaling ^{law over such a} large

range of Ts (P ) also supposes that the characteristic energy of the spin fluctuations always remains smal- ler than the crystal-field splitting L1CF. This condition is probably ^{realized in the} present system ^{where the} latter energy ^was estimated to be of the order of several hundred kelvin [5]. On the other hand, ^we

have already noted that the magnetic anisotropy drops rapidly with pressure, which may ^{seem to} be in contradiction with the preceding assertion. The

important point ^{here is} that this reduction of the

anisotropy results from markedly contrasted be- haviours of X II and X_L ^{as a} function of pressure. The close agreement found between rll(x) ^{and the values}

deduced from specific-heat (with ^Y and La substitu-

tion) ^and transport experiments performed ^{in zero}

field (see above) clearly indicates that the pressure variation of XII is dominated by spin-fluctuation

processes. In contrast, X.L ^{seems to} involve a

different mechanism.

To understand these features, ^{one must consider}

the effect of crystal-field anisotropy. The energy level scheme derived from specific-heat ^measure-

ments [5] consists of a well-isolated ground-state doublet, ^{with the next doublets} lying about 220 and 1 000 K above. By taking ^{into account the} tetragonal symmetry ^{and the} sign ^{of the} anisotropy, ^{one is led}

for the c-axis to a ground ^state of the form :

a I:t 5/2 > - N,/l - a 2IT 3/2). ^{In a} previous study [24], ^{it was} emphasized ^{that the} system ^{was close to a}

case of pure axial symmetry (crystal-field coefficient

B44 = 0, a = 1). However in this limit, X.L arises

entirely ^{from a} Van Vleck mechanism and thus its

experimental temperature dependence [7] ^{cannot be}

accounted for. Turning ^{now to} pressure effects, ^it

can be shown, using for instance the calculation of Hanzawa et al. [25], ^that X.L (T = 0) ^{for a pure}

!±5/2) (or !±3/2)) doublet would be practically

unaffected by ^a change ^of TK up ^to the 50 K range, in contrast to the present ^results.

On the other hand, ^{a value of}

^{= 0.96 was}

derived from the magnetic ^moment (IL II ^{= 1.9} 9B

extrapolated ^{to infinite field} (P. ^Lehmann [17]).

With such a value, Curie-type contribution is added to X-L (T) ^{in the} crystal-field-only limit, ^{and accord-} ingly, ^Hanzawa’s Xj_ (T = 0) decreases if TK goes up. The main observation is that for a realistic choice of the parameters ^and assuming TK -= ²⁵ K, ^{the Van} Vleck term always represents ^a large fraction of X.l. (T) ^{down to} TK X ^has predominantly ^{a Curie}

character. Therefore, the pressure variation of whereas XII (T TK) ^will directly reflect that of

TK (i. e. 1-II(x) - - r (TK ), ^whereas the decrease of X I will be slower.

The arguments given above thus provide ^a simple explanation ^{for the} principal features of our results.

However, ^we will refrain from going ^{into more}

detailed calculations because it is quite obvious that

no realistic description ^of CeRu2Si2 ^{can be worked}

out without including ^the effect of intersite coup-

lings : ^as noted in the introduction, ^AF correlations have been observed to set in at temperatures ^as high

as 60 K, ^{and no more than 8 %} substitution of La for Ce is sufficient for getting ^a long-range magnetic

structure [26]. Unfortunately, ^{no current Kondo} theory incorporates ^this type of effects together ^with

a realistic treatment of the crystal-field anisotropy.

This problem ^{becomes even more serious when}

one comes to the high-field properties. Although ^the

maximum magnetic ^field accessible in the present experiments ^{do not allow us to} investigate ^the polarized ^{state above} HM, ^{we want to} briefly ^mention

the question whether the drop of Xo II under pressure should be ascribed to a reduction of the saturation moment, ^or simply ^{to a} global shift of the magneti-

zation curve to higher magnetic fields. We already

noted above that the disappearance ^{of the} upward

curvature in Mil (H) ^{favours the second interpre-}

tation. This is confirmed by ^the experimental ^obser-

vation that in substituting ^Ce by ^{La, the saturation} magnetization ^remains practically unchanged ^while

the value of HM ^decreases [17]. ^From magnetoresist-

ance measurements [11], HM ^{was found to increase} by - ^{16 %} per kbar, ^so that Xo 11 HM ^remains roughly

constant up ^to 6 kbar. Furthermore, ^{it was shown} recently by magnetostriction experiments [10] ^that M(H, ^{T --+} 0 ) ^{scales as} f [H/Hs (P ) ] , if ^{the relative}

variation of HS (P ) is chosen equal ^{to -} rll(x). ^{For a}

complete discussion of the high-field ^{results, the}

reader is referred ^to the original papers.

In conclusion, the pressure increase of the charac- teristic energy of CeRu2Si2 ^{has been determined}

from static susceptibility measurements. The excel- lent agreement obtained between the values deduced from Xo, Ts, ^{and the} coefficient of the quadratic

term in the resistivity supports ^a single-energy-scale

picture, ^{such as} that which emerges from ^recent

Kondo-lattice models. A direct measurement of

1560

y (P ) ^{is in} project ^to complete ^this analysis. ^From

the theoretical point ^{of view,} further effort is clearly

needed to start from realistic models, including ^the

strong crystalline anisotropies, antiferromagnetic correlations, ^and metamagnetic ^{behaviour in} high fields, ^{and then to work out the} single energy scale

that seems to dominate these competing phenomena.

Acknowledgments.

The authors are grateful ^{to Drs. A.} Meyer, ^K.

Matho and J. Pierre for useful discussions concerning crystal-field calculations.

References

[1] ^{RIEGER, W. and} PARTHÉ, E, Monatsh. Chem. 100

(1969) ^444.

[2] QUEZEL, ^S. ROSSAT-MIGNOD, ^J. CHEVALIER, B., LEJAY, ^{P. and} ETOURNEAU, J., Solid State Commun. 49 (1984) ^685.

[3] HIEBEL, K., HORVATH, C., ROGL, ^{P. and} SIENKO, M. J., Z. Phys. ^{B 56} (1984) ^201.

[4] STEGLICH, F., AARTS, J., BREDL, ^C. B., LIEKE, W., MESCHEDE, D., FRANZ, ^{W. and} SCHÄFER, H., Phys. Rev. Lett. 43 (1979) ^1892.

[5] BESNUS, M.-J., KAPPLER, J.-P., LEHMANN, ^{P. and} MEYER, A., Solid State Commun. 55 (1985) ^779.

[6] ^{THOMPSON, J. D., WILLIS, J. O,} GODART, C., MCLAUGHLIN, ^{D. E.} and GUPTA, ^L. C., ^Solid State Commun. 56 (1985) ^169.

[7] HAEN, P., FLOUQUET, J., LAPIERRE, F., LEJAY, ^P.

and REMENYI, G., ^{J. Low} Temp. Phys. ⁶⁷ (1987) ^391.

[8] REGNAULT, L.-P., ERKELENS, ^{W. A.} C., ROSSAT- MIGNOD, J., LEJAY, ^{P. and} FLOUQUET, J., Phys. ^Rev. B, ^{to be} published.

[9] ^DE VISSER, A., MENOVSKY, ^A. and FRANSE, ^{J. J.}

M., Physica ^147B (1987) ^81, and references therein.

[10] PUECH, L., MIGNOT, J.-M., VOIRON, J., LEJAY, P., HAEN, ^{P. and} FLOUQUET, J., ^{J. Low} Temp.

Phys. ⁷⁰ (1988) ^237.

[11] HAEN, P., LAPIERRE, F., LEJAY, P., MIGNOT, J.-M., PONCHET, ^{A. and} VOIRON, J., J. Magn. Magn.

Mat. 63-64 (1987) ^320.

[12] HIEBL, K., HORVATH, C., ROGL, ^{P. and} SIENKO, ^M.

J., ^J. Magn. Magn. ^{Mat. 37} (1983) 287 ; VOIRON, J. (unpublished results).

[13] KRISHNA-MURTHY, H. R., WILKINS, ^{J. W. and} WILSON, ^K. G., Phys. ^{Rev. B. 21} (1980) 1003 ; Ibid 1044.

[14] LAWRENCE, J., Phys. ^{Rev. B 20} (1979) ^3770.

[15] GODARD, C., UMARJI, ^A. M., GUPTA, ^L. C., and VIJAYARAGHAVAN, R., Phys. ^{Rev. B 34} (1986)

7733.

[16] ^PUECH, L., PONCHET, A., MIGNOT, J.-M., VOIRON, J., LEJAY, P., HAEN, ^P. and FLOUQUET, J., Jpn

J. Appl. Phys. ²⁶ Suppl. ^26-3 (1987) 2103 ; MIGNOT, J.-M., PONCHET, A., HAEN, ^{P. and}

FLOUQUET, ^J. (to ^be published) ;

PONCHET, A., Thesis, University ^{of Grenoble} (1987).

[17] BESNUS, M.-J., LEHMAN, ^{P. and} MEYER, A., ^J.

Magn. Magn. Mat. 63-64 (1987) ^{323 ;}

LEHMAN, P., Thesis, University ^Louis Pasteur, ^Stras- bourg (1987).

[18] ^DE VISSER, A., FRANSE, ^{J. J.} M., MENOVSKY, ^A.

and PALSTRA, ^{T. T.} M., Physica ^127B (1984) 442 ;

STEWART, ^G. R., FISK, Z., WILLIS, J. O. and SMITH, J. L., Phys. ^{Rev. B 52} (1984) ^679.

[19] FLOUQUET, J., LASJAUNIAS, J.-C., PEYRARD, ^{J. and} RIBAULT, M., ^J. Appl. Phys. ⁵³ (1982) ^2127.

[20] RIBAULT, M., BENOIT, A., FLOUQUET, ^{J. and PAL-}

LEAU, J., ^J. Phys. Lett. France 40 (1979) ^L-413.

[21] JACCARD, D., CIBIN, R., JORDA, J.-L. and FLOUQUET, J., Jpn ^J. Appl. Phys. ²⁶ Suppl. ^26-3 (1987) ^{517 ;}

BARTH, S., OTT, ^H. R., GYGAX, ^F. N., HITTI, B., LIPPELT, E., SCHENCK, ^{A. and} BAINES, C., Jpn

J. Appl. Phys. ²⁶ Suppl. ^26-3 (1987) ^519.

[22] BRODALE, ^G. E., FISHER, ^R. A., PHILLIPS, ^{N. E.}

and FLOUQUET, J., Phys. ^{Rev. Lett. 56} (1986)

390.

[23] JACCARD, ^{D. and} FLOUQUET, J., ^J. Magn. Magn.

Mat. 47-48 (1985) ^45.

[24] FISHER, ^{G. and} HERR, A., Phys. Status Solidi B 141

(1987) ^589.

[25] HANZAWA, K., YAMADA, ^{K. and} YOSIDA, K., ^J.

Phys. ^Soc. Jpn ⁵⁵ (1986) ^641.

[26] QUEZEL, S., BURLET, P., JACOUD, J.-L., ^RE-

GNAULT, L.-P., ROSSAT-MIGNOT, J., LEJAY, ^P.

and FLOUQUET, J., submitted to the ICCF6

Conf. (Frankfurt FRG) ^1988.