Magnetovolume effect in Ce(Ni1−xCux)₅ alloys

Magnetic susceptibility χ of the isostructural Ce(Ni1−xCux)₅ alloys (0 ≤ x ≤ 0.9) was studied as a function of the hydrostatic pressure up to 2 kbar at fixed temperatures of 77.3 and 300 K, using a pendulum-type magnetometer. A pronounced pressure effect on χ is found to be negative in sign and stro...

Повний опис

Збережено в:
Бібліографічні деталі
Дата:2007
Автори: Grechnev, G.E., Logosha, A.V., Panfilov, A.S., Svechkarev, I.V., Musil, O., Svoboda, P.
Формат: Стаття
Мова:English
Опубліковано: Донецький фізико-технічний інститут ім. О.О. Галкіна НАН України 2007
Назва видання:Физика и техника высоких давлений
Онлайн доступ:http://dspace.nbuv.gov.ua/handle/123456789/70290
Теги: Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
Назва журналу:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Цитувати:Magnetovolume effect in Ce(Ni1−xCux)₅ alloys / G.E. Grechnev, A.V. Logosha, A.S. Panfilov, I.V. Svechkarev, O. Musil, P. Svoboda // Физика и техника высоких давлений. — 2007. — Т. 17, № 1. — С. 59-66. — Бібліогр.: 22 назв. — англ.

Репозитарії

Digital Library of Periodicals of National Academy of Sciences of Ukraine
id irk-123456789-70290
record_format dspace
spelling irk-123456789-702902014-11-03T03:01:46Z Magnetovolume effect in Ce(Ni1−xCux)₅ alloys Grechnev, G.E. Logosha, A.V. Panfilov, A.S. Svechkarev, I.V. Musil, O. Svoboda, P. Magnetic susceptibility χ of the isostructural Ce(Ni1−xCux)₅ alloys (0 ≤ x ≤ 0.9) was studied as a function of the hydrostatic pressure up to 2 kbar at fixed temperatures of 77.3 and 300 K, using a pendulum-type magnetometer. A pronounced pressure effect on χ is found to be negative in sign and strongly (non-monotonously) dependent on the Cu content, showing a sharp maximum at x ≅ 0.4. The experimental results are discussed in terms of the valence instability of Ce ion in the studied alloys. For the reference CeNi₅ compound the main contributions to χ and their volume dependence are calculated ab initio within the local spin density approximation (LSDA), and appeared to be in close agreement with experimental data. 2007 Article Magnetovolume effect in Ce(Ni1−xCux)₅ alloys / G.E. Grechnev, A.V. Logosha, A.S. Panfilov, I.V. Svechkarev, O. Musil, P. Svoboda // Физика и техника высоких давлений. — 2007. — Т. 17, № 1. — С. 59-66. — Бібліогр.: 22 назв. — англ. 0868-5924 PACS: 71.20.Eh, 75.30.Mb, 75.80.+q http://dspace.nbuv.gov.ua/handle/123456789/70290 en Физика и техника высоких давлений Донецький фізико-технічний інститут ім. О.О. Галкіна НАН України
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
collection DSpace DC
language English
description Magnetic susceptibility χ of the isostructural Ce(Ni1−xCux)₅ alloys (0 ≤ x ≤ 0.9) was studied as a function of the hydrostatic pressure up to 2 kbar at fixed temperatures of 77.3 and 300 K, using a pendulum-type magnetometer. A pronounced pressure effect on χ is found to be negative in sign and strongly (non-monotonously) dependent on the Cu content, showing a sharp maximum at x ≅ 0.4. The experimental results are discussed in terms of the valence instability of Ce ion in the studied alloys. For the reference CeNi₅ compound the main contributions to χ and their volume dependence are calculated ab initio within the local spin density approximation (LSDA), and appeared to be in close agreement with experimental data.
format Article
author Grechnev, G.E.
Logosha, A.V.
Panfilov, A.S.
Svechkarev, I.V.
Musil, O.
Svoboda, P.
spellingShingle Grechnev, G.E.
Logosha, A.V.
Panfilov, A.S.
Svechkarev, I.V.
Musil, O.
Svoboda, P.
Magnetovolume effect in Ce(Ni1−xCux)₅ alloys
Физика и техника высоких давлений
author_facet Grechnev, G.E.
Logosha, A.V.
Panfilov, A.S.
Svechkarev, I.V.
Musil, O.
Svoboda, P.
author_sort Grechnev, G.E.
title Magnetovolume effect in Ce(Ni1−xCux)₅ alloys
title_short Magnetovolume effect in Ce(Ni1−xCux)₅ alloys
title_full Magnetovolume effect in Ce(Ni1−xCux)₅ alloys
title_fullStr Magnetovolume effect in Ce(Ni1−xCux)₅ alloys
title_full_unstemmed Magnetovolume effect in Ce(Ni1−xCux)₅ alloys
title_sort magnetovolume effect in ce(ni1−xcux)₅ alloys
publisher Донецький фізико-технічний інститут ім. О.О. Галкіна НАН України
publishDate 2007
url http://dspace.nbuv.gov.ua/handle/123456789/70290
citation_txt Magnetovolume effect in Ce(Ni1−xCux)₅ alloys / G.E. Grechnev, A.V. Logosha, A.S. Panfilov, I.V. Svechkarev, O. Musil, P. Svoboda // Физика и техника высоких давлений. — 2007. — Т. 17, № 1. — С. 59-66. — Бібліогр.: 22 назв. — англ.
series Физика и техника высоких давлений
work_keys_str_mv AT grechnevge magnetovolumeeffectinceni1xcux5alloys
AT logoshaav magnetovolumeeffectinceni1xcux5alloys
AT panfilovas magnetovolumeeffectinceni1xcux5alloys
AT svechkareviv magnetovolumeeffectinceni1xcux5alloys
AT musilo magnetovolumeeffectinceni1xcux5alloys
AT svobodap magnetovolumeeffectinceni1xcux5alloys
first_indexed 2025-07-05T19:34:53Z
last_indexed 2025-07-05T19:34:53Z
_version_ 1836836815537963008
fulltext Физика и техника высоких давлений 2007, том 17, № 1 59 PACS: 71.20.Eh, 75.30.Mb, 75.80.+q G.E. Grechnev1, A.V. Logosha1, A.S. Panfilov1, I.V. Svechkarev1, O. Musil2, P. Svoboda2 MAGNETOVOLUME EFFECT IN Ce(Ni1−xCux)5 ALLOYS 1B. Verkin Institute for Low Temperature Physics and Engineering 47 Lenin Ave., 61103 Kharkov, Ukraine 2Charles University, Faculty of Mathematics and Physics, DES Ke Karlovu 5, 121 16 Prague 2, The Czech Republic E-mail: panfilov@ilt.kharkov.ua Magnetic susceptibility χ of the isostructural Ce(Ni1−xCux)5 alloys (0 ≤ x ≤ 0.9) was studied as a function of the hydrostatic pressure up to 2 kbar at fixed temperatures of 77.3 and 300 K, using a pendulum-type magnetometer. A pronounced pressure effect on χ is found to be negative in sign and strongly (non-monotonously) dependent on the Cu content, showing a sharp maximum at x ≅ 0.4. The experimental results are discussed in terms of the valence instability of Ce ion in the studied alloys. For the reference CeNi5 compound the main contributions to χ and their volume dependence are calculated ab initio within the local spin density approximation (LSDA), and appeared to be in close agreement with experimental data. Introduction Many of Ce intermetallics are characterized by a strong hybridization of the magnetic 4f electrons with the conduction electron states resulted in delocalization of the 4f level and a change of its occupancy, and hence the Ce valence. As is evi- dent from measurements of X-ray absorption and lattice parameters [1], together with the magnetic [2,3], electric and thermoelectric properties [3] in the isostruc- tural Ce(Ni1−xCux)5 alloys, the Ce valence decreases steadily from Ce4+ to Ce3+ with increase of the Cu content, and the system undergoes a series of transitions from the nonmagnetic metal with empty 4f level (x = 0) through the intermediate valence (IV) state combined with a nonmagnetic dense Kondo state (0.1 ≤ x ≤ 0.8) to the magnetic 4f metal (0.9 ≤ x ≤ 1). Thus, the reference CeNi5 compound is the exchange-enhanced itinerant paramagnet [1,4,5] with the temperature dependent magnetic susceptibility exhibiting a broad maximum around 100 K, similar to those observed in YNi5 and LuNi5 [4,6]. On the other hand, the CeCu5 compound is a Kondo lattice antiferromagnet with TN = 3.9 K and TK = 2.2 K [7]. The mag- Физика и техника высоких давлений 2007, том 17, № 1 60 netic susceptibility in CeCu5 at T ≥ 50 K obeys a Curie−Weiss law with the effec- tive magnetic moment value close to that expected for Ce3+ state [7–9]. Due to a direct relation between magnetic properties and the rare earth (RE) valence state, and also the strong correlation between the valence itself and RE ionic volume, the RE compounds with unstable f shell exhibit a large magnetovolume effect. Therefore, a study of pressure effect on magnetic properties of the systems with variable RE valence are of great interest to gain insight into a nature of the IV state. Here we report results of our study of the pressure effect on the magnetic sus- ceptibility of Ce(Ni1−xCux)5 alloys in a wide range of Cu concentrations. The ex- perimental results are supplemented by calculations of the magnetovolume effect value for the reference CeNi5 compound, using a modified relativistic full poten- tial linearized «muffin-tin» orbital method (FP-LMTO). Experimental details and results The polycrystalline samples of Ce(Ni1−xCux)5 alloys (0 ≤ x ≤ 4.5) were pre- pared by arc-melting of a stoichiometric amount of initial elements in a water- cooled crucible under protective argon atmosphere. The study of X-ray powder diffraction at room temperature revealed that all samples crystallize in CaCu5-type hexagonal structure, and obtained data on their lattice parameters agree closely with that published in literature. Any other phases were not detected within the resolution of the X-ray technique. The pressure effect on magnetic sus- ceptibility was measured under helium gas pressure up to 2 kbar at fixed tem- peratures, 77.3 and 300 K, using a pen- dulum magnetometer placed into the nonmagnetic pressure cell [10]. The relative errors of our measurements, per- formed in the magnetic field H = 1.7 T, did not exceed 0.05%. The pressure de- pendence χ(P) appeared to be linear in all samples studied. For each tempera- ture, the values of χ at ambient pressure and their pressure derivatives, dlnχ/dP, were corrected for a weak field depend- ence of χ caused by ferromagnetic impu- rities. The corresponding corrections were less than 5%. In Fig. 1 the typical pressure depend- encies of the magnetic susceptibility for Ce(Ni1−xCux)5 alloys demonstrate a mag- 0.97 0.98 0.99 1.00 b a χ( P) /χ (0 ) χ( P) /χ (0 ) 0.0 0.5 1.0 1.5 2.0 0.98 0.99 1.00 P, kbar Fig. 1. Pressure dependence of the mag- netic susceptibility of Ce(Ni1−xCux)5 al- loys at T = 77.3 (a) and 300 K (b), nor- malized to its value at P = 0: • − x = 0, ■ − 0.3, ▼ − 0.4, ▲ − 0.5 Физика и техника высоких давлений 2007, том 17, № 1 61 nitude of the pressure effect and its linear behavior. The negative sign of the effect is consistent with anticipa- tion that high pressure has to increase the valence, since the Ce ion in the higher valence (less magnetic) state has a smaller volume. Of particular interest is a strong and non-monotonous concentration dependence of the pressure effect which shows a sharp maximum in the vicinity of x ≈ 0.4 for both tempera- tures, 77.3 and 300 K (Fig. 2,a). A comparison between the obtained ex- perimental results and the data on concentration dependence of the lat- tice parameter a and the effective Ce valence ν from [1] (Fig. 2,b) indicates that the maximum in dlnχ(x,T)/dP correlates with a drastic change of a (and ν) around x ≈ 0.4 (ν ≈ 3.5). It is interesting to note that a similar peculiarity in dlnχ/dP versus valence was observed for various Yb compounds at room temperature [11]. As was shown, the relative change of χ with pressure is the most pro- nounced also at the half-integer value of valence, ν = 2.5, but contrary to Ce com- pounds, it has a positive sign, as can be expected. Theory for CeNi5 Ab initio calculations of the electronic structure were carried out for the refer- ence compound CeNi5 by employing a modified FP-LMTO method [12,13]. The exchange-correlation potential was treated in the LSDA approximation [14] of the density functional theory. To analyze the observed magnetovolume effect value in CeNi5, the magnetic susceptibility and its volume dependence were calculated within the modified method, wherein the external magnetic field H was taken into account by means of the Zeeman operator, H(2s + l). The latter was incorporated in FP-LMTO Hamiltonian [15] for calculations of the field-induced spin and or- bital magnetic moments. The corresponding contributions to magnetic suscepti- bility were derived from the field-induced moments, which have been calculated in an external magnetic field of 10 T. The electronic structure calculations were performed for a number of lattice parameters close to the experimental one. The –20 –16 –12 –8 –4 0 a d ln χ /d P, M ba r–1 0.0 0.2 0.4 0.6 0.8 1.0 –0.10 –0.08 –0.06 –0.04 –0.02 0 CeNi5–xCux ∆a , Å x 4.0 3.8 3.6 3.4 3.2 3.0 b C e va le nc e Fig. 2. (a) Pressure derivative of the mag- netic susceptibility dlnχ/dP in Ce(Ni1−xCux)5 alloys at 77.3 K (•) and 300 K (○). (b) De- viation of the a lattice parameter, ∆a (○), in Ce(Ni1−xCux)5 alloys from the a(x) depend- ence for the Ce ion assumed to be in a tri- valent state (left scale) and the Ce valence (•) deduced from X-ray absorption studies (right scale) at room temperature versus Cu content x (according to the data of Ref. [1]) Физика и техника высоких давлений 2007, том 17, № 1 62 equilibrium lattice spacing a and corre- sponding theoretical bulk modulus B were determined from the dependence of the total energy on the unit cell volume (the ratio c/a was fixed at its experimental value), and appeared to be a = 8.96 a.u. and B = 1.9 Mbar (to be compared with experimental a = 9.2 a.u. [1] and B =1.43 Mbar [16]). The differences between theory and experiment on bulk properties of CeNi5 are presumably related to the overbonding tendency of the LSDA ap- proach [12]. The strongly volume dependent spin contribution to χ originates predomi- nantly from the 3d states of Ni. Regard- ing the orbital contribution to χ, it comes mainly from electrons in atomic sphere of Ce and amounts to about 20% of total susceptibility. At the theoretical lattice parameter, the calculated total suscepti- bility at T = 0 K (2.9·10−3 emu/mol) ap- peared to be very close to the experi- mental one (3.0·10−3 emu/mol [4]). The corresponding volume derivative of sus- ceptibility, dlnχ/dlnV = 4.2, is in agreement with that resulted from the experi- mentally observed pressure derivative for CeNi5 at T = 77.3 K, dlnχ/dlnV = 3.9 ± 0.4. Thus, it has been demonstrated that LSDA provides an adequate description of the strongly exchange enhanced magnetic susceptibility of CeNi5 and its pres- sure dependence. Discussion As is shown, the LSDA allows to describe the magnetovolume effect in the reference CeNi5 compound that gives grounds for future application of LSDA ab initio approaches to some Ce(Ni1−xCux)5 alloys. Here, however, we shall restrict our consideration of the experimental data in alloys within a phenomenological approach. A. Concentration dependence Anticipating the pressure effect on the magnetic susceptibility to arise mainly from the change of Ce valence, or the fractional occupation of the 4f 1 magnetic state n4f (ν = 4 – n4f), the pressure effect can be analyzed within a simple relation 0 1 2 0.45 x = 0.4 0.7 a 0.6 ln χ 0.4 0.6 0.8 1.0 0 2 4 6 –d n 4f /d P, M ba r–1 b n4f = 4 – ν Fig. 3. Values of lnχ at 77.3 K (a) and dn4f /dP (b) both plotted against n4f for Ce(Ni1−xCux)5 alloys. Circles and trian- gles in (b) denote the data obtained with Eq. (1) and Eq. (4), respectively. Points for x = 0.45 are interpolation of the ex- perimental data on concentration depend- ence of χ and dlnχ/dP Физика и техника высоких давлений 2007, том 17, № 1 63 4f 4f d ln ln d d d n P n P χ ∂ χ ≈ ∂ (1) in terms of the pressure dependence of n4f (or ν). The most reliable results of such analysis would be expected in the Cu-rich alloys at low temperatures where the 4f contribution χ4f becomes dominant. In Fig. 3,a the χ versus n4f dependence is shown for Ce(Ni1−xCux)5 alloys (0.4 ≤ x ≤ 1) at 77.3 K, which is obtained by using the experimental χ(x) values and ν(x) data of Fig. 2,b. A substitution of the resulted from Fig. 3,a derivatives ∂lnχ/∂n4f and experimental data on dlnχ/dlnP at 77.3 K into Eq. (1) gives dn4f /dP value which strongly depends on n4f (Fig. 3,b). As is seen, the maximum value of dn4f /dP is expected at n4f = 0.5 (ν = 3.5) to be about −6.5 ± 1.5 Mbar−1. The cor- responding estimates of the valence change under pressure, dν/dP = −dn4f /dP, are of the same order as those resulted from the study of magnetovolume effect in SmB6 (2 Mbar−1 [17]) and from the measurements of resonant inelastic X-ray emission in YbAl2 under pressure (~ 5 Mbar−1 [18]). B. Temperature dependence In a simple empirical model which includes interconfiguration fluctuations between f n+1 and f n levels [19], the contribution of the 4f 0 (J = 0) and 4f 1 (J = 5/2) states of Ce to magnetic susceptibility is given by 2 4 4( ) ( ) / 3 ( )f f fT N n T k T Tχ = µ + . (2) Here N is the Avogadro number, µ − effective magnetic moment of the 4f 1 state, Tf – the characteristic temperature (valence fluctuation temperature, or Kondo temperature, or heavy-fermion bandwidth). It should be mentioned that a quantitative analysis of the χ4f (T) dependence using Eq. (2) requires the complete data on n4f(T) (and probably on Tf(T) as well) which are actually unavailable. Furthermore, to separate the χ4f(T) term from the experimental data on χ(T) one needs to know a background contribution χ0, which generally can not be ne- glected. A simplified analysis of the experimental data can be performed assuming n4f, Tf and χ0 to be temperature independent. Then the magnetic susceptibility obeys a modified Curie−Weiss law, 0 4f 0( ) ( ) /( )T T C Tχ = χ + χ ≡ χ + −θ , (3) with C = Nµ2n4f /3k and θ = −Tf. For the representative Ce(Ni0.5Cu0.5)5 alloy, the best fit of Eq. (3) to the experimental data at T ≥ 50 K (Fig. 4,a) [2] is obtained with χ0 = 0.6·10−3 emu/mol, C = 0.48 K·emu/mol and θ = −79 K. It should be pointed out that the estimate n4f = 0.6, resulted from C, is in a reasonable agree- ment with the value of 0.8 that follows from the data in Fig. 2,b for x = 0.5. Физика и техника высоких давлений 2007, том 17, № 1 64 As is evident from Eqs. (2) and (3), the pressure effect on the 4f susceptibility is governed by changes of n4f and Tf with pressure, as 0 50 100 150 200 250 300 0.0 0.2 0.4 0.6 0.8 T, K (χ -χ 0)–1 , 1 03 m ol /e m u) 0 1 2 3 4 –16 –12 –8 –4 0 d ln χ 4f /d P, M ba r–1 χ4f, 10–3 emu/mol а б Fig. 4. Temperature dependence of the magnetic susceptibility χ (a) and pressure deriva- tive dlnχ4f/dP plotted against χ4f (b) for Ce(Ni0.5Cu0.5)5 alloy 4 4 4d ln ( ) d d ln ( ) dd ln 1 d d d d d f f f f f f T T n T TC P P T T P P C P χ χ = − ≡ − + , (4) being a linear function of 1/(T + Tf) or χ4f (T). The data on dlnχ4f /dP for Ce(Ni0.5Cu0.5)5 alloy were derived from the measured effect, dlnχ/dP, in the framework of Eq. (3) using a value dlnχ0/dP ~ −1 Mbar−1 as a rough estimate for the pressure dependence of the background [20] which is assumed to originate from 3d(4d) itinerant electrons. The obtained values dlnχ4f /dP are plotted in Fig. 4,b as a function of χ4f (T). Its linear approximation in accordance with Eq. (4) gives 4 1 1d ln dd ln 3.2 0.7 Mbar , 1650 250 Mbar d d d f fn TC K P P P − −= = − ± = ± ⋅ . The resultant value dn4f /dP = − 2.5 ± 0.5 Mbar−1 is in line with the value dn4f /dP = = − 2.0 ± 0.3 Mbar−1 obtained previously for x = 0.5 from analysis of the concen- tration dependence of the pressure effect within Eq. (1). From the pressure de- pendence of Tf the corresponding Grüneisen parameter, Ω, is estimated as d ln d ln 31 5 d ln d f fT T B V P Ω ≡ − = = ± (5) using the experimental bulk modulus B = 1.5 Mbar [21]. The Anderson impurity model provides the Kondo temperature and its pressure derivative to be described in terms of n4f [22]: 4 4K K 4 4 1 d lnd ln 1, d 1 d f f f f n nTT n P n P − ∝ = − . (6) Физика и техника высоких давлений 2007, том 17, № 1 65 Then, assuming Tf ∝ TK and using in Eq. (6) the values dn4f /dP = −3.2 ± 0.7 Mbar−1 and n4f = 0.8 evaluated above for the alloy with x = 0.5, one obtains ln T 24 5 ln V K f K d d Ω = Ω = − = ± , (7) in reasonable agreement with the direct estimate (5). For Ce(Ni0.4Cu0.6)5 alloy, the analogous analysis in the framework of Eq. (3) and Eq. (4) yields the following Curie−Weiss parameters: C ~ 0.806 K·emu/mol, χ0 ~ 0, Tf = −θ = 26 K, and their pressure derivatives: 4 1 1d ln dd ln 1.7 0.5 Mbar , 620 100 Mbar d d d f fn TC K P P P − −= = − ± = ± ⋅ . The latter results in the Grüneisen parameter Ωf = 35 ± 6, assuming the bulk modulus value B = 1.5 Mbar, as in the Ce(Ni0.5Cu0.5)5 alloy. The similar estimate follows from Eqs. (6) and (7) with n4f = 0.93 derived from the data in Fig. 2,b. The reasonable description of the Grüneisen parameter for alloys with x = 0.5 and 0.6 with Anderson model [21] allows to consider the Cu-rich alloys studied in the present work as the nonmagnetic Kondo lattices. Conclusions The pressure effect on magnetic susceptibility of Ce(Ni1−xCux)5 alloys has been observed for the first time. This effect is negative in sign, and also strongly and non-monotonously dependent on the Cu content. For the reference CeNi5 compound, the pressure effect value is successfully described within LSDA ap- proximation, using the modified full potential relativistic FP-LMTO method. For Ce(Ni1−xCux)5 alloys the effects of pressure and alloying on the valence state of Ce ion are the most pronounced around x ~ 0.4, which corresponds to the half- integer valence ν ~ 3.5. In other words, the fractional occupation n4f ~ 0.5 with the nearly degenerate f 0 and f 1 configurations of electronic states is favorable for the valence instability. It is also found that the main contributions to the pressure ef- fect on magnetic susceptibility for the Cu-rich alloys are i) the decrease of the ef- fective Curie constant and ii) the increase of the characteristic temperature Tf The latter exhibits a large and positive value of the Grüneisen parameter, which can be apparently described within the Anderson impurity model. Both of these contri- butions have their origin in the change of the Ce valence state caused by depopu- lation of the f state under pressure due to its shift relative to the Fermi energy. The work of P.S. and O.M. is a part of the research program MSM 0021620834 financed by the Ministry of Education of the Czech Republic. The authors thank S.N. Dolya for discussions and V.A. Desnenko for help in magnetic measurements. Физика и техника высоких давлений 2007, том 17, № 1 66 1. D. Gignoux, F. Givord, R. Lemaire, H. Launois, F. Sayetat, J. Physique 43, 173 (1982). 2. O. Musil, P. Svoboda, M. Diviš, V. Sechovsky, Czech. J. Phys. 51, Suppl. D, D311 (2005). 3. N.B. Brandt, V.V. Moshchalkov, N.E. Sluchanko, E.M. Savitskii, T.M. Shkatova, Solid State Phys. 26, 2110 (1984). 4. M. Coldea, D. Andreica, M. Bitu, V. Crisan, J. Magn. Magn. Mater. 157/158, 627 (1996). 5. L. Nordström, M.S.S. Brooks, B. Johansson, Phys. Rev. B46, 3458 (1992). 6. E. Burzo, V. Pop, I. Costina, J. Magn. Magn. Mater. 157/158, 615 (1996). 7. E. Bauer, M. Rotter, L. Keller, P. Fisher, M. Ellerby, K.M. McEwen, J. Phys.: Condens. Matter 6, 5533 (1994). 8. I. Pop, R. Pop, M. Coldea, J. Phys. Chem. Solids 43, 199 (1982). 9. I. Pop, E. Rus, M. Coldea, O. Pop, J. Phys. Chem. Solids 40, 683 (1979). 10. A. S. Panfilov, High Pressure Physics and Technology 2, № 2, 61 (1992) (in Russian). 11. W. Zell, R. Pott, B. Roden, D. Wohlleben, Solid State Commun. 40, 751 (1981). 12. J.M. Wills, O. Eriksson, Electronic Structure and Physical Properties of Solids, H. Dreysse (ed.), Springer, Berlin, 247 (2000). 13. K.H.J. Buschow, G.E. Grechnev, A. Hjelm et al., J. Alloys and Compounds 244, 113 (1996). 14. U. von Barth, L. Hedin, J. Phys. C5, 1629 (1972). 15. G.E. Grechnev, R. Ahuja, O. Eriksson, Phys. Rev. B68, 64414 (2003). 16. B. Butler, D. Givord, F. Givord, S.B. Palmer, J. Phys. C13, L743 (1980). 17. A. S. Panfilov, I.V. Svechkarev, Yu. B. Paderno, E.S. Konovalova, V.I., Lazorenko, High Pressure Physics and Technology issue 20, 3 (1985) (in Russian). 18. C. Dallera, E. Annese, J.-P. Rueff et al., J. Phys.: Condens. Matter 17, S849 (2005). 19. B.C. Sales, D. Wohlleben, Phys. Rev. Lett. 35, 1240 (1975). 20. V. Heine, Phys. Rev. 153, 673 (1967). 21. U. Staub, C. Schulze-Briese, P.A. Alekseev et al., J. Phys.: Condens. Matter 13, 11511 (2001). 22. O. Gunnarsson, K. Schonhammer, Phys. Rev. B28, 4315 (1983).