Pressure-induced three-dimensional ferromagnetic correlations in the giant magnetocaloric compound G d 5 G e 4
C. Magen, L. Morellon, P. A. Algarabel and M.R. Ibarra
Gd 5 (Si x Ge 1-x ) 4 is a unique class of materials where many interesting properties and intriguing behavior have been recently discovered . The unprecedented giant magnetocaloric effect , strong magnetoelastic effects [3, 4], and giant magnetoresistance [5, 6] can be highlighted as the most relevant. This phenomenology has been associated with the intrinsically layered crystallographic structure combined with a magnetic-martensitic first-order phase transformation . The coupled magnetic-crystallographic transition can be induced reversibly by the change of external parameters such as temperature or external magnetic field [1, 4]. Therefore, these alloys are attractive for their potential applications in magnetic refrigeration and/or as magnetostrictive/ magnetoresistive transducers.
Three extended solid solution regions exist in the temperature-composition (T–x) phase diagram [8,9]: the Si-rich solid solution, 0.575 £ x £ 1, has the orthorhombic Gd 5 Si 4 -type structure [ O (I)]; the intermediate phase 0.4<x £ 0.503 has a room temperature monoclinic ( M ) structure; and the Ge-rich region, 0<x £ 0.3 crystallizes in the Gd 5 Ge 4 -type structure [ O (II)]. All three structures are composed of identical two-dimensional (2D) sub-nanometer-thick layers (slabs) interconnected via partially covalent inter-slab X–X bonds (X = Si, Ge). In the O (I) structure, all the slabs are interconnected by X–X bonds; half of these bonds are broken in the M structure and none remain in the O (II) structure. The magnetic-crystallographic transition involves breaking/reforming specific covalent X–X bonds  and the low-temperature ground state for all compositions 0<x £ 1 is always ferromagnetic (FM) with all the slabs being interconnected, i.e. with the O (I) structure. The M structure is always paramagnetic (PM) whereas the O (II) can support either PM or antiferromagnetism (AFM) [4, 9]. The magnetic behavior of the R 5 (Si x Ge 1-x ) 4 compounds can be understood qualitatively in terms of competition between intralayer (within the 2D slabs, conventional indirect 4f- 4f RKKY) and interlayer exchange interactions (between slabs, direct Gd-Si/Ge-Gd superexchange propagated via the X-X bonds) [1, 10].
In sharp contrast with the magnetic behavior of the Ge-rich compounds no FM phase is observed in Gd 5 Ge 4 in zero magnetic field and down to the lowest measured temperature [11, 12]. Gd 5 Ge 4 orders antiferromagnetically at ~ 130 K, this system presenting a very complex magnetic field-temperature (H–T) phase diagram and an interesting magnetoelastic behavior . A fully irreversible field-induced O (II) (AFM) ® O (I) (FM) transformation takes place below » 10 K, this becoming fully reversible above » 20 K. Between » 10 K and » 20 K the field-induced transition is partially reversible and a spatially phase-segregated O(II) + O(I) state is found at low temperatures.
The aim of this work is to investigate the possibility to induce three-dimensional (3D) ferromagnetic correlations in Gd 5 Ge 4 upon application of an external hydrostatic pressure.
In Fig. 1 we display the linear thermal expansion (LTE) of Gd 5 Ge 4 at different values of the applied hydrostatic pressure. Upon application of pressure, a distinct anomaly develops in the LTE curve and a clear jump is detected at higher pressure values. According to our investigation of the magnetoelastic behavior of this compound , this anomaly signals the existence of a pressure-induced low-temperature O(I)–FM phase, the amount of this phase being reflected in the magnitude of the jump.
FIG. 1. Linear Thermal Expansion as a function of temperature under selected values of the applied hydrostatic pressure. In the inset the relative change in volume of the sample at selected temperatures is displayed
In order to correlate the structural behavior with the magnetic properties we have carried out a systematic study of the magnetism of Gd 5 Ge 4 as a function of temperature, applied magnetic field and applied hydrostatic pressure. In Fig. 2 we display the temperature dependence of the magnetization in an applied field of 500 Oe at selected values of the applied hydrostatic pressure: 0, 1.3, and 8 kbar. As expected [11-13], at ambient pressure no FM state is detected down to 5 K (open circles) but application and removal of 50 kOe at 5 K induces a FM state that remains up to » 14 K (open squares). As is clearly seen in Fig. 3, an applied hydrostatic pressure of 1.3 kbar is sufficient to induce a FM signal at » 35 K (heating), the value of which at 5 K is much lower than that obtained when applying and removing 5 T at ambient pressure. Therefore, from this and the results in Fig. 1, we can interpret this data supposing that the sample volume at 1.3 kbar is spatially phase segregated into O(II)–AFM and O(I)–FM regions, i.e. 1.3 kbar is sufficient to enhance the interlayer interactions favoring 3D FM correlations. As far as we know, this is the first time phase-separation phenomena have been observed in a pure 4 f localized-moment system upon application of hydrostatic pressure. If we apply and remove isothermally a magnetic field of 50 kOe at 5 K and 1.3 kbar, a full FM signal is recovered (solid squares). Should this physical picture be correct, a higher pressure should increase the relative volume of the O(I)–FM regions. This is indeed the case since the FM signal at 8 kbar is almost maximum. We therefore propose that a hydrostatic pressure of 8 kbar is able to induce a O(I)–FM ground state in the majority of the sample with a T C @ 65 K (heating).
||FIG. 2. Magnetization in an applied field of 500 Oe as a function of temperature under selected values of the applied hydrostatic pressure. Black arrows indicate the direction of temperature change.
The relative percentage of the pressure-induced O(I)–FM phase can be estimated from magnetization measurements. The final values are displayed in Fig. 3 (solid circles) together with the results obtained from the LTE data, see Fig. 1, assuming that the jump in the LTE at 10.9 kbar corresponds to a 100% of O(I)-FM transformed phase at low temperatures. As can be seen the results are in very good agreement, this being a strong evidence of the physical picture presented to interpret the LTE and magnetization results as a function of pressure. From the results in Fig. 3 we can be quite confident that at pressures above » 10 Kbar the entire volume of the sample at low temperatures below T C (10 Kbar) @ 53 K is FM crystallizing in the O(I) structure.
In conclusion, we have explored the possibility to induce 3D ferromagnetic correlations in the giant magnetocaloric alloy Gd 5 Ge 4 by means of LTE and magnetic measurements at different applied pressures. As a main result, application of hydrostatic pressure induces a spatially phase-segregated ground state where O(I)–FM and O(II)–AFM regions coexist within the sample volume. It has been predicted that at pressures of over 10 kbar, the low-temperature O(I)–FM ground state is stabilized in the entire sample volume.
FIG. 3. Percentage of pressure-induced O(II)-FM
phase as estimated from LTE results (open circles) and magnetization
measurements (solid circles). The line is a guide for the eye
We propose a physical picture where the effect of pressure
is to reduce the interatomic distances, thus enhancing the interlayer interactions
in this naturally nanolayered material. This should favor the formation of
specific covalent bonds, inducing a first-order pressure-induced magnetic-crystallographic
transformation from an O(II)–AFM to an O(I)–FM structure.
C. Magen, Z. Arnold, L. Morellon, Y. Skorokhod, P. A. Algarabel,
M. R. Ibarra, J. Kamarad , Phys. Rev. Lett 91 , 207202
The finantial support of Grant No. MAT2000-1756 is acknowledged.
 V. K. Pecharsky and K. A. Gschneidner, Jr., Adv. Mater.
13 , 683 (2001).
 V. K. Pecharsky and K. A. Gschneidner, Jr., Phys. Rev. Lett. 78
, 4494 (1997); Appl. Phys. Lett. 70 , 3299 (1997);
J. Magn. Magn. Mater. 167 , L179 (1997).
 L. Morellon et al., Phys. Rev. B 58 , R14721 (1998).
 L. Morellon et al., Phys. Rev. B 62 , 1022 (2000).
 L. Morellon et al., Appl. Phys. Lett. 73 , 3462 (1998);
J. Magn. Magn. Mater. 237 , 119 (2001).
 E. M. Levin, V. K. Pecharsky, and K. A. Gschneidner, Jr., Phys. Rev. B
60 , 7993 (1999).
 W. Choe et al., Phys. Rev. Lett. 84 , 4617 (2000).
 V. K. Pecharsky, and K. A. Gschneidner, Jr., J. Alloys and Compounds 260
, 98 (1997).
 A. O. Pecharsky et al. , J. Alloys and Compounds 338
, 126 (2002).
 E. M. Levin, V. K. Pecharsky, and K. A. Gschneidner Jr., Phys. Rev. B
62 , R14625 (2000).
[ 13] E. M. Levin, K. A. Gschneidner, Jr., and V. K. Pecharsky, Phys. Rev.
B 65 , 214427 (2001).
 E. M. Levin et al. , Phys. Rev. B 64 , 235103
 C. Magen et al., J. Phys.: Condensed Matter 15 , 2389