Journal of ELECTRONIC MATERIALS

https://doi.org/10.1007/s11664-019-07027-7
Ó 2019 The Minerals, Metals & Materials Society

Structural, Magnetic, Magnetocaloric and Mössbauer

Spectrometry Study of Gd2Fe17x Cux (x ¼ 0, 0.5, 1 and 1.5)
Compounds

M. SAIDI,1,2 K. NOURI,1,2,5 S. WALHA,2 E. DHAHRI,3 A. KABADOU,2

M. JEMMALI,2,4 and L. BESSAIS1

1.—ICMPE (UMR 7182), CNRS, UPEC, Université Paris Est, 94320 Thiais, France.
2.—Laboratoire des Sciences des Matériaux et de l’Environnement, Faculté des Sciences de
Sfax, Université de Sfax, BP 1171, 3018 Sfax, Tunisia. 3.—Laboratoire de Physique Appliquée,
Faculté des Sciences de Sfax, Université de Sfax, BP 1171, 3000 Sfax, Tunisia. 4.—Chemistry
Department, College of Science and Arts at Ar-Rass, Qassim University, P.O. Box 53, Buraydah,
Saudi Arabia. 5.—e-mail: [email protected]

The structure, magnetic and magnetocaloric properties of arc-melted

Gd2 Fe17x Cux (x ¼ 0, 0.5, 1 and 1.5) solid solution have been studied. The
Rietveld refinement shows that these compounds crystallize in the rhombo-
hedral Th2 Zn17 -type structure with the R3m space group, and the substitution
of iron by copper leads to a decrease in the unit cell volume. The Curie tem-
perature (TC ) of the prepared samples depends on the copper content. The
reduction of the ferromagnetic phase transition temperature from 475 K (for
x ¼ 0) to 460 K (for x ¼ 1:5) is due to the substitution of the magnetic element
(Fe) by non-magnetic atoms (Cu). The magnetocaloric effect was determined in
the vicinity of the Curie temperature TC . By increasing the Cu content, an
increase in the values of magnetic entropy (DSM ) in a low applied field is
observed. The preferred inequivalent crystallographic site of Cu atoms was
determined by Mössbauer spectrometry analysis based on the isomer shift
parameter correlation with the Wigner–Seitz cell volumes.

Key words: Rare-earth alloys and compounds, magnetization, Mössbauer

spectrometry, magnetocaloric effect

sublattice, and (3) the intersublattice 3d–4f

INTRODUCTION
exchange interaction (JRM ). It is noteworthy that
A lot of research has been devoted to magnetic the interactions between the rare-earth spins 4f–4f
materials due to their complexity and the large are supposed to be negligible in comparison with the
number of applications in which they could be other two types of interactions.4–12
used.1–3 Long-range magnetic order is mainly found These rare-earth transition metal compounds
in intermetallic compounds that are based on 4f have received much attention in recent years and
rare-earth metals (R) and 3d transition metals (M), many review articles were written about them for
and in which occur three types of exchange inter- these two reasons.5,13–17 Firstly, rare-earth metals
actions, namely: (1) the 3d–3d exchange interaction have large magnetic moments, while transition
(JMM ) between the magnetic moments of the M metals possess high Curie temperatures. These
sublattice, (2) the 4f–4f exchange interaction (JRR ) two properties are very interesting when found in
between the magnetic moments within the R the same material, which could be used as a
permanent magnet. Secondly, the chemical similar-
ities of rare-earth elements and transition metals
permit substitution of themselves in a given
(Received September 19, 2018; accepted January 31, 2019) material.
 Saidi, Nouri, Walha, Dhahri, Kabadou, Jemmali, and Bessais

The existing intermetallic compounds in the rare- powder diffraction, magnetic measurements and
earth-iron system are RFe2 , RFe3 , R6 Fe23 and Mössbauer spectroscopy analysis.
R2 Fe17 . Due to coexistence of complementary fea- The structural characterization was done at room
tures of itinerant (3d) and localized (4f) electrons in temperature through x-ray diffraction patterns
these compounds, they exhibit useful magnetic using a Bruker diffractometer with a monochroma-
properties. Because the R2 Fe17 compound is the tor providing the CuKa radiation (k ¼ 1:5406 Å).
richest in Fe among these intermetallic com- Data for Rietveld refinement of the samples were
pounds,18,19 it has been intensively studied with collected in the range of 2h from 20:000 to 80:000
the aim of improving its magnetic characteristics for with a step width of 2h ¼ 0:015 and a counting time
potential practical applications.4–6,8,10 The R2 Fe17 of 13.5 s per scanning step. X-ray powder diffraction
compound exhibits a rich variety of magnetic was performed to examine the crystallographic
behaviors such as ferro-, antiferro- or ferri-magnetic structure, to identify the lattice parameters and to
ones, and more complex magnetic orderings of the check the phase purity and homogeneity of the
rare-earth element.20–22 prepared samples. Structural analysis was carried
Recent interest in these R2 Fe17 intermetallic out using the Rietveld powder profile method36–40
compounds has been renewed due to the magne- applying the FullProf software.41–43
tocaloric properties shown by these ferromagnetic The magnetization of the samples was performed
compounds Pr2 Fe17 , Nd2 Fe17 , Er2 Fe17 , Gd2 Fe17 , at high temperature using a DSM-8 MANICS
Sm2 Fe17 and Tb2 Fe17 .23–28 The selected ferromag- differential sample magnetometer, working on the
netic compound Gd2 Fe17 studied in this research same principle as a Faraday balance. The thermo-
has a magnetocaloric effect with a maximum mag- magnetic curve M(T) of each sample was measured
netic entropy DSM equal to 0.89 J/kg K, for an in the 300–800-K temperature range, under a low
external field change from 0 T to 1.5 T at room applied magnetic field of 0.12 T.44 The magnetic
temperature.27 It is noteworthy that the magne- transition temperature TC was deduced from the
tocaloric properties are particularly useful in room- maximum slope in dM=dT plots. In addition, the
temperature magnetic refrigeration technology isothermal magnetization curves M(H) were
which is more efficient, compact and environmen- recorded in the proximity of the Curie temperature.
tally safe than conventional gas refrigeration.29–35 The change of magnetic entropy was determined as
The goal of this paper is to study the structural, a function of the temperature and the change of
magnetic and magnetocaloric effect of Cu substitu- applied magnetic field through numerical integra-
tion in the iron-rich Gd2 Fe17x Cux intermetallic tion of the isothermally measured curves M(H).
compounds. In addition, due to the similarity The room-temperature 57 Fe Mössbauer spectra
between the Fe and Cu structure factors, which with absorbers containing 10 mg/cm2 of natural
makes it difficult to localize the Cu atoms in the iron were collected on a constant-acceleration spec-
structure using x-ray diffraction, we used Möss- trometer (Weissel), in transmission geometry using
bauer spectrometry to localize accurately the Cu 25 mCi 57 Co in Rh matrix source. The spectrometer
substitution site. Furthermore, the Mössbauer tech- calibration gives a line width of 0.25 mm/s for a-Fe.
nique explores the microscopic magnetic properties The spectra were least-squares fitted according to
for the four inequivalent crystallographic Fe sites. the procedure discussed below with estimated
errors of ± 0.1 T for hyperfine fields HHF ,
EXPERIMENTAL METHODS AND DATA ± 0.005 mm/s for isomer shifts d and quadrupole
ANALYSIS splitting 2. The Mössbauer spectroscopy was used
Polycrystalline samples of Gd2 Fe17x Cux (x ¼ 0, to identify iron-containing phases45 and to study
0.5, 1 and 1.5) were synthesized by means of arc- magnetism on a microscopic level.46
melting of stoichiometric starting elements of high
purity (Gd, Fe and Cu: 99.9%) in a water-cooled RESULTS AND DISCUSSION
copper crucible under a high-purity protective argon Structure Analysis
gas atmosphere. To ensure good starting homogene-
ity of the samples, the ingots were turned and It is well known that there are two different
melted several times. An excess amount of rare- crystal structures in the R2 Fe17 compound, the
earth element was added to compensate for the rhombohedral Th2 Zn17 -type structure (R3m  space
evaporation losses of this element during melting. group) for light rare earths, and the hexagonal
This resulted in samples with less than 0.5% weight Th2 Ni17 -type structure (P63 =mmc space group) for
loss in most of the compounds. To reach equilibrium heavy rare earths.1,47 Both structures are deriva-
and/or homogeneity, the arc-melted ingots were tives from the basic CaCu5 -type structure with one-
wrapped separately in tantalum foil, sealed under third rare earth atoms replaced by a pair of iron
vacuum in quartz tubes and then annealed at 800 C atoms. In this R2 Fe17 series, the rhombohedral
for 1 week.11 After annealing, the samples were crystal structure is found for R ¼ Ce, Pr, Nd and
quenched in water. All the final products were Sm.1,26,48–50 The compounds with R ¼ Dy, Er, Tm
ground into powder and characterized by x-ray and Lu crystallize in the hexagonal one.51 However,
Structural, Magnetic, Magnetocaloric and Mössbauer Spectrometry Study of Gd2Fe17  xCux
(x = 0, 0.5, 1 and 1.5) Compounds

the compounds with R ¼ Gd, Tb and Y, crystallize in Interlayer Fe(6c) ions at the dumbbell positions
both the hexagonal or the rhombohedral structures lower the symmetry of some iron ions in the nearest
depending on the annealing temperature.52 intermediate layers from 18h to 9d (above and below
The Rietveld analysis of x-ray diffraction patterns the dumbbell). Therefore, these intermediate layers
shows that the Gd2 Fe17x Cux (x ¼ 0, 0.5, 1 and 1.5) without Gd contain ions Fe (9d), as well as Fe (18h)
compounds are single-phase and crystallize in the in the hexagons.
rhombohedral Th2 Zn17 -type structure (space group, Figure 3 shows the coordination polyhedra for

R3m). Furthermore, the absence of the a-Fe phase each iron site. The Fe (6c) is surrounded by six Fe
in the samples is due to the excess amount of Gd (18f), three Fe (18h), three Fe (9d) and one Fe atom
that was added to compensate for the evaporation at the 6c site. Furthermore, the polyhedron around
losses of this element during melting. the atom at site 18h forms two hexagons in the
Based on the x-ray powder diffraction, we observe structure. The neighboring sites contain two Fe
that Cu substitution for Fe does not change the atoms 9d, two 18h, four 18f and one iron atom at the
structure of Gd2 Fe17 and that all the studied 6c site. Besides, the atom at site 18f is located at the
Gd2 Fe17x Cux compounds crystallize in the center of a icosahedron, and its corners are occupied
Th2 Zn17 -type structure. The calculated and exper- by two Fe atoms 9d, two 18f, four 18h and two iron
imental pattern curves for Gd2 Fe17 and atoms at the 6c site. Finally, the nearest atoms for
Gd2 Fe16:5 Cu0:5 are shown in Fig. 1 as examples. the iron atom at site 9d (four atoms at site 18h, four
The Rietveld fittings reveal a good matching of the atoms at site 18f and two atoms at site 6c) form a
observed and calculated profiles for all the systems. icosahedron.
The values of the crystallographic unit cell param- In this study, Rietveld refinement was also used
eters, the iron atomic positions as well as the to investigate the preferential substitution site.
reliability factors obtained from the best refinement Firstly, the refinement started by allocating the
are illustrated in Table I. As will be shown in the Cu atoms randomly in the four sites 18h, 18f, 9d and
following sections, these accurate, refined unit cell 6c. Secondly, the distribution of Cu was done for the
parameters will be used to calculate the interatomic 18h, 18f, 9d site and 6c separately. The best
distances, the near-neighbor environments and the agreement factor RB of the Gd2 Fe17x Cux com-
Wigner–Seitz cell volumes. pounds was obtained with iron located in 6c, 9d,
Figure 2 illustrates the crystal structure of the 18f and 18h sites, while Cu occupies site 18h. The
rhombohedral Gd2 Fe17 . The rare earth atoms refinement results of the Gd2 Fe17x Cux (x ¼ 0, 0.5, 1
occupy one type of crystallographic site 6c, while and 1.5) series show that copper atoms prefer to
the iron atoms are located in four inequivalent substitute the iron atoms at site 18h rather than at
lattice positions 6c, 9d, 18f and 18h as described in other sites. In addition, the Mössbauer spectroscopy
Ref. 53. In this type of structure, the main struc- was used as a complementary measurement method
tural blocks are iron hexagon building layers. Most in order to corroborate or refute the Cu site
of these hexagons have empty centers, while others substitution.
contain gadolinium ions or iron ions in dumbbell It is worthy to note that Pokharel et al.54 have
positions in their centers. The Fe (18f) hexagons indicated that in the Gd2 Fe17x Tix solid solution, Ti
contain the gadolinium ions in positions 6c at the tends to avoid the 9d and 6c sites, while preferen-
center of each hexagon, as shown in Fig. 2. tially occupying sites 18h and 18f. Moreover, it was

Fig. 1. Observed (dots) and calculated (solid line) conventional x-ray powder diffraction patterns collected at room temperature with Cu radiation
of Gd2 Fe17 and Gd2 Fe16:5 Cu0:5 compounds. The vertical bars indicate the position of the Bragg diffraction reflections. The observed-calculated
difference is depicted at the bottom of the figures.
 Saidi, Nouri, Walha, Dhahri, Kabadou, Jemmali, and Bessais

Table I. X-ray diffraction analysis results obtained by Rietveld method on the Gd2 Fe17x Cux (x ¼ 0, 0.5, 1 and
1.5) compounds
x¼0 x ¼ 0:5 x¼1 x ¼ 1:5

a (Å) 8.5430 (7) 8.539 (6) 8.534 (4) 8.533 (5)

c (Å) 12.448 (3) 12.3436 (7) 12.433 (3) 12.426 (6)
V (Å3 ) 784.57 (5) 785.554 (5) 784.254 (2) 783.748 (4)
v2 4.26 4.66 2.43 1.82
RB 4.85 5.37 11.2 3.54
xf18hg 0.503 0.493 0.487 0.493
xf18f g 0.304 0.307 0.306 0.306
yf18f g 0.496 0.506 0.512 0.509
zf18hg 0.162 0.158 0.159 0.162
zf6cg (Fe) 0.068 0.081 0.075 0.082
zf6cg (Gd) 0.337 0.338 0.336 0.339

0
8.544 12.450

8.542 12.445

8.540
12.440

c(Α)
a(Α)

8.538
12.435

8.536
12.430

8.534
12.425
8.532
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
x content
Fig. 4. Variation of Gd2 Fe17x Cux lattice parameters a and c as a
function of the x content.

The Rietveld refinement proves that the substi-

 crystal tution of Fe by Cu shows that the lattice parameters
Fig. 2. Schematic representation of a rhombohedral R 3m
structure of Gd2 Fe17 . a, c and v decrease when increasing Cu content x.
This result is expected knowing that the copper
(r ¼ 1:45 Å) atomic radius is smaller than that of
the iron (r ¼ 1:56 Å) one. Figure 4 gives the varia-
tion of the cell parameters a and c of the 2:17 phase
versus the Cu content x in the Gd2 Fe17x Cux (x ¼ 0,
0.5, 1 and 1.5) compounds.

Magnetic and Magnetocaloric Properties

The temperature dependence of the magnetiza-
tion M(T) of the Gd2 Fe17x Cux for x ¼ 0; 0:5; 1; 1:5
compounds was measured on a differential sample
magnetometer under a low applied magnetic field of
0.12 T with a heating rate of 5 K/min. All the
compounds show a sharp ferromagnetic-paramag-
netic transition at TC , as presented in Fig. 5. The
value of the transition temperature TC was identi-
fied from the minimum of the temperature deriva-
Fig. 3. The coordination polyhedra for each crystallographic Fe site.
tive of the magnetization dM=dT versus
temperature.
reported that the preferential substituent occu- For Gd2 Fe17x Cux with x ¼ 0, we have found a TC
pancy of Si in Gd2 Fe17x Six and in Sm2 (Fe,Si)17 is equal to 475 K, which is in agreement with the
preferentially at the 18h iron site.53,55,56 results found by Chen et al.23
Structural, Magnetic, Magnetocaloric and Mössbauer Spectrometry Study of Gd2Fe17  xCux
(x = 0, 0.5, 1 and 1.5) Compounds

Table II shows that the Curie temperatures of the magnetic measurements of the isotherms M(H) in
elaborated compounds Gd2 Fe17x Cux (x ¼ 0, 0.5, 1 the vicinity of their Curie temperature with a step of
and 1.5) decrease with the increasing copper content 5 K for each sample. We noticed that for a given
from 475 K (for x ¼ 0) to 460 K (for x ¼ 1:5). This applied field, magnetization increases with decreas-
behavior could be explained not only by the geo- ing temperature. Figure 7 shows the magnetic field
metrical effect involving unit-cell volume diminu- dependence of the isothermal magnetization M(H)
tion, but also by the magnetic dilution that occurs of the Gd2 Fe17x Cux for x ¼ 0, 0.5, 1 and 1.5
while substituting Fe atoms with non-magnetic Cu measured up to 1.5 T at different temperatures.
atoms. Above the Curie temperature, the M(H) curves are
Furthermore, the saturation magnetization MS characterized by a paramagnetic behavior, and,
was deduced from the fit of the isothermal magne- therefore, the magnetizations tend to slowly
tization versus magnetic field curves at room tem- increase with the applied magnetic field.
perature using the saturation approach law57: The magnetic entropy variation DSðTÞ is deter-
  mined as a function of temperature T and the
b applied magnetic field H through the numerical
M ¼ MS 1  2 :
H integration of the isothermally measured M(H)
curves using the following Maxwell relation58,59:
As an example, we show here in Fig. 6 the M(H) Z H 
@M
plot for Gd2 Fe16:5 Cu1:5 at T ¼ 300 K. This compound DS ¼ dH
reaches a magnetic saturation value MS ¼ 0 @T
X Miþ1 ðTiþ1 ; Hj Þ  Mi ðTi ; Hj Þ
68:18 emu/g (Am2 =kg). We have found that the DSðTi ; DHÞ ¼ DHj :
substitution of Fe with Cu atoms has the effect of j
Tiþ1  Ti
decreasing MS as follows: 93:31 Am2 /kg (21:12 lB )
for x ¼ 0; 68:18 Am2 /kg (15:48 lB ) for x ¼ 0:5;
Figure 8 illustrates the magnetic entropy change
61:83 Am2 =kg (13:91 lB ) for x ¼ 1 and 55:82 Am2 /kg of the compounds Gd2 Fe17 and Gd2 Fe16:5 Cu0:5 as a
(12:75 lB ) for x ¼ 1:5. function of temperature and the external field
The magnetocaloric effect of Gd2 Fe17x Cux (x ¼ 0, change. The maximum peak values of DSmax for
0.5, 1 and 1.5) compounds were studied by means of x ¼ 0, x ¼ 0:5, x ¼ 1 and x ¼ 1:5 are equal to 1.14 J/kg
K, 2.14 J/kg K, 2.35 J/kg K and 2.54 J/kg K,

Fig. 5. Temperature dependence of magnetization of Gd2 Fe17x Cux

for x ¼ 0, 0.5, 1 and 1.5 compounds.
Fig. 6. Example of M(H) curve for Gd2 Fe16:5 Cu0:5 at T ¼ 300 K.

Table II. Values of TC , maximum of entropy variation DSmax , and relative cooling power RCP for
Gd2 Fe17x Cux (x ¼ 0, 0.5, 1 and 1.5) compounds
Compounds Gd2 Fe17 Gd2 Fe16:5 Cu0:5 Gd2 Fe16 Cu1 Gd2 Fe15:5 Cu1:5

TC (K) 475 466 462 460

 DSmax (J/kg K) 1.14 2.14 2.35 2.54
RCP (J/kg) 18 23.7 27.2 31.5
 Saidi, Nouri, Walha, Dhahri, Kabadou, Jemmali, and Bessais

Fig. 7. Isotherm magnetization curves M(H, T) of Gd2 Fe17x Cux for (a) x ¼ 0, (b) x ¼ 0:5, (c) x ¼ 1 and (d) x ¼ 1:5.

respectively. We found that the value of DSmax plots is more pronounced when x is changed from 0
increases by increasing the Cu content in the to 0.5 than for x ¼ 1 or x ¼ 1:5. This behavior might
Gd2 Fe17x Cux (x ¼ 0, 0.5, 1 and 1.5). be due to the Cu magnetic dilution.
Another interesting parameter for characterizing In order to confirm the second-order magnetic
magnetocaloric properties and determining the phase transition, the Landau theory can be used.61
suitability of a particular magnetic material for The magnetic free energy F can be developed as a
working bodies in magnetic refrigerators is the function of the magnetization close to the Curie
relative cooling power (RCP). This parameter is temperature.62 When neglecting terms of higher
related to the magnetic entropy variation according power, the free energy F can be given as:
to the following formula:
1 1 1
F ¼ aðTÞ  M 2 þ bðTÞ  M 4 þ cðTÞ  M 6  l0 MH:
RCP ¼ DSMax  dTFWHM 2 4 6
where DSMax is the maximum value of the entropy
change (DS) and dTFWHM is the full width half We have to impose @F=@M ¼ 0 in order to deter-
maximum of DSðTÞ.60 mine the equilibrium condition; as a consequence,
Additionally, in order to study the magnetic we obtain the following equation that describes the
transition nature, we plotted the so-called Arrott total magnetization around TC :
plots, M 2 versus H/M of Gd2 Fe17x Cux compounds
for all Cu content. The positive slope near TC for the aðTÞ  M þ bðTÞ  M 3 þ cðTÞ  M 5  l0 H ¼ 0:
Gd2 Fe17x Cux (x ¼ 0, 0.5, 1 and 1.5) compounds (see
Fig. 9) shows that the transition from a ferromag- The parameters a(T), b(T) and c(T) are the
netic to a paramagnetic state at TC is a second-order Landau coefficients. They are determined by fitting
phase transition. The influence of Cu on the Arrot l0 HðMÞ using the equation above. The minimum of
Structural, Magnetic, Magnetocaloric and Mössbauer Spectrometry Study of Gd2Fe17  xCux
(x = 0, 0.5, 1 and 1.5) Compounds

Fig. 8. Entropy change of Gd2 Fe17x Cux for (a) x ¼ 0, (b) x ¼ 0:5, (c) x ¼ 1 and (d) x ¼ 1:5.

a(T), indicating a maximum of susceptibility, per- density increases, the isomer shift decreases. How-
mits us to confirm the value of TC . Besides, when ever, the increase in the electron density d causes a
the Landau coefficient bðTC Þ is negative, the mag- lower attraction of the s electron by the nuclear
netic transition is of first order. Otherwise, when charge (more shielding of the s electron) and
bðTC Þ is positive, the magnetic transition is of consequently increases the isomer shift.
second order.63,64 Figures 10 and 11 show respec- This is illustrated with the equation:
tively the plots of Landau coefficients for both  DR
Gd2 Fe17 and Gd2 Fe16:5 Cu0:5 compounds, as an 2 2
d ¼ Ze wð0Þa  wð0Þs 2
 R2 :
example. As expected, the minimum of a(T) con- R
firms the values of the Curie temperatures. Addi-
tionally, the magnetic transition is of second order, where R is the effective nuclear radius and DR=R is
given that b(T) is positive for T > TC . the nuclear factor, representing the variation of the
nuclear radius under the effect of the Mössbauer
Hyperfine Parameters transition, wð0Þ2a represents the electron density of
the absorbent and wð0Þ2s is the electron density at
The hyperfine parameters for the Gd2 Fe17x Cux
compounds were studied in order to confirm the Cu the source nucleus.
substitution location obtained by Rietveld refine- Figure 12 illustrates the room-temperature Möss-
ment. This study mainly allows us to determine the bauer spectra with the fit of Gd2 Fe17x Cux (x ¼ 0,
relationship between the isomer shift and the 0.5, 1 and 1.5) compounds. It is worth noting that
Wigner–Seitz cell volumes (WSCVs) and the mag- the complex and magnetically ordered spectrum is
netic moment of each Fe atom site in the structure. the result of the convolution of numerous subspec-
The isomer shift increases due to the interaction tra. This complexity is not only due to the existence
between the nuclear charge density and the sur- of four nonequivalent crystallographic sites, but
rounding s electron cloud. Since the s electron wave also to the existence of Fe-Fe dumbbells (6c site)
function reaches its maximum at the nucleus, it connected to the Gd vacancies. As a consequence, we
affects isomer shift to a large extent. p and d need a pertinent physical model connected to the
electrons ensure screening of the s electron in order structure characteristics in order to analyze this
to reduce its density. Whenever the s electron kind of spectrum. The site assignment of the
 Saidi, Nouri, Walha, Dhahri, Kabadou, Jemmali, and Bessais

Fig. 9. Arrott plots of Gd2 Fe17x Cux for (a) x ¼ 0, (b) x ¼ 0:5, (c) x ¼ 1 and (d) x ¼ 1:5.

Fig. 10. Temperature dependency of Landau coefficients for Gd2 Fe17 compound. (a) Landau coefficient a(T), (b) Landau coefficient b(T).

hyperfine parameters observed for the different multinomial distribution; (2) the correlation
crystallographic iron sites in Gd2 Fe17x Cux was between the WSCV and the isomer shift (d) is
determined by the following rules: (1) the area of calculated using crystallographic data derived from
the sextets subspectra are assessed by a Rietveld analysis;65–68 (3) the correlation between
Structural, Magnetic, Magnetocaloric and Mössbauer Spectrometry Study of Gd2Fe17  xCux
(x = 0, 0.5, 1 and 1.5) Compounds

Fig. 11. Temperature dependency of Landau coefficients for Gd2 Fe16:5 Cu0:5 compound. (a) Landau coefficient a(T), (b) Landau coefficient b(T).

Table III. Interatomic distances and number of

iron near neighbors for the rhombohedral R3m 
structure type for Gd2 Fe17x Cux (x ¼ 0:5; 1:5). The
distances are calculated inside a radius of the
coordination sphere equal to 2.755 Å
x =0
Atom 1 Atom 2 Distance (Å) Number

x ¼ 0:5 x ¼ 1:5

Fe (9d) Fe (6c) 2.685 2.677 2

Fe (18h) 2.821 2.538 4
Fe (18f) 2.461 2.442 4
Megacounts

x = 0.5 Fe (6c) Fe (6c) 1.802 1.646 1

Fe (18f) 2.811 2.812 6
Fe (9d) 2.685 2.677 3
Fe (18h) 2.552 2.829 3
Fe (18f) Fe (18f) 2.625 2.619 2
Fe (18h) 2.568 2.609 4
Fe (9d) 2.461 2.442 2
x=1 Fe (6c) 2.811 2.812 2
Fe (18h) Fe (18f) 2.568 2.609 4
Fe (18h) 2.372 2.323 2
Fe (6c) 2.552 2.829 1
Fe (9d) 2.821 2.538 2

18h sites have the most (13) and the least (9)
x = 1.5
numbers of nearest-neighbor Fe atoms. Therefore,
-10 -8 -6 -4 -2 0 2 4 6 8 10 the most and least hyperfine fields are to be
Velocity (mm/s) associated with the 6c and 18h sites, respectively.
Fig. 12. Room-temperature Mössbauer spectra of Gd2 Fe17x Cux Both the 9d and 18f sites have ten Fe atoms as their
compounds. The black solid line and the red + symbol scatter plot neighbors. However, the relative numbers of site
are the calculated and experimental spectra, respectively. The occupancy on the two sites are 3 and 6, respectively.
colored solid lines are the seven sextets used in the fit (Color figure Therefore, the subspectra for the 9d and 18f sites
online).
can be distinguished by their subspectral areas. The
interatomic distances and the number of iron near-
the hyperfine field and the near-neighbor environ- neighbors are determined for the rhombohedral
 structure type for Gd2 Fe17x Cux (x ¼ 0.5 and
R3m
ments of the four iron sites.
The highest the number of Fe near neighbors, the 1.5), and presented in Table III.
highest its hyperfine field. The assignments of four In order to apply the model based on the WSCV
subspectra corresponding to four Fe sites were made correlation with isomer shift,66–68 the Rietveld
by considering their relative areas together with the refined accurate values of the unit-cell parameters
neighboring environments of the sites. The 6c and with the atomic positions are used for the
 Saidi, Nouri, Walha, Dhahri, Kabadou, Jemmali, and Bessais

Table IV. Room-temperature Mössbauer hyperfine parameters, hyperfine field (l0 HHF ), isomer shift (d),
quadripole splitting (Q) and Wigner–Seitz cell volumes (WSCV) for Gd2 Fe16:5 Cu0:5
6c 9d 18f 18h Wt. ave.

l0 HHF (T) 31.9 (1) 28.4 (1) 26.3 (1) 23.7 (1) 26.5 (1)
d (mm/s) 0.035 (5)  0.040 (5)  0.025 (5)  0.012 (5)  0.020 (5)
Q (mm/s) 0.105 (5)  0.034 (5) 0.263 (5) 0.125 (5) 0.120 (5)
WSCV (Å3 ) 12.42 (2) 11.36 (3) 11.73 (5) 12.01 (5) 11.82 (5)

result in an increase of the isomer shift. In contrast,

if the neighborhood of the 18h site is slightly
0.04
affected by the copper substitution, the s charge
density remains unchanged. There is no change
0.02
observed for the 18h isomer shift. Therefore, the
difference for the 18h site behavior then confirms
Isomer Shift (mm/s)

0.00
the preferential occupation of copper atoms on this
inequivalent crystallographic site.
-0.02
Using the conversion factor of 14.5 T/lB for metal-
-0.04
lic iron we found the average iron magnetic moment
hlðFe atÞi equal to 2.01 lB for x ¼ 0, 1:83 lB for
6c
x ¼ 0:5, 1:75 lB for x ¼ 1 and 1:74 lB for x ¼ 1:5.
-0.06 18h
18f Assuming that the Gd magnetic moment lGd ¼ 7 lB ,
-0.08 9d we have deduced the magnetic moment per the
formula unit of Gd2 Fe17 : l ¼ 20:20 lB /f.u.; Gd2 Fe17
-0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Cu0:5 : l ¼ 16:11 lB /f.u.; Gd2 Fe17 Cu1 : l ¼ 14:05 lB /
Cu content f.u.; Gd2 Fe17 Cu1:5 : l ¼ 12:96 lB /f.u. These results
Fig. 13. Isomer shift versus Cu content for Gd2 Fe17x Cux for x ¼ 0, are in agreement with those derived from magneti-
0.5, 1 and 1.5. zation measurements.

calculation of the WSCV of each inequivalent crys- CONCLUSIONS

tallographic site. The WSCVs were calculated by In summary, the structural characterization,
means of Dirichlet domains and the coordination determination of the magnetic properties and the
polyhedra for each atomic site 6c, 9d, 18h and 18f magnetocaloric effects and the Mössbauer analysis
with the radius values of 1.88 Å, 1.26 Å and 1.35 Å, were performed for the Gd2 Fe17x Cux (x ¼ 0, 0.5, 1
respectively, for Gd, Fe and Cu. The calculated and 1.5) compounds. A single rhombohedral Th2 Zn17
WSCVs are summarized in Table IV. Consequently, phase was obtained after 1 week of heat treatment at
the WSCV sequence is 6c > 18h > 18f > 9d. 800 C for synthesis by means of arc-melting. The
The results of the calculated Mössbauer spectrum crystal structure of the parent Gd2 Fe17 was found to
show that the hyperfine fields obey the following remain unchanged for the Gd2 Fe17x Cux compounds
classification: HHF f6cg > HHF f9dg > HHF f18f g > HHF which crystallize in the rhombohedral structure with
f18hg confirmed by the number of iron neighbors in  Our experimental results show
the space group R3m.
the corresponding site. The sequence observed for that in the Gd2 Fe17x Cux (x ¼ 0, 0.5, 1 and 1.5)
the isomer shift is df6cg > df18hg > df18f g > df9dg compounds, the Curie temperature decreases by
(Fig. 13), which is in agreement with the WSCV increasing the copper content from 475 K for (x ¼ 0)
sequence. The Mössbauer hyperfine parameters and to 460 K for (x ¼ 1:5). This behavior can be attributed
WSCV for Gd2 Fe16:5 Cu0:5 are summarized in Table - to the unit-cell volume decrease and the magnetic
IV as an example. dilution. Moreover, we found that the copper substi-
This behavior could be understood in terms of the tution leads to an increase in DSmax from 1.14 J/kg for
preferential Cu atom occupation. The 6c, 9d and 18f Gd2 Fe17 to 2.54 J/kg K for Gd2 Fe15:5 Cu1:5 . In addi-
have 3, 4 and 4 adjacent 18h nearest neighbors, tion, Arrott plots and Landau parameters allowed us
respectively, and the 18h site atoms have only 2. to prove that the transition from the ferromagnetic
The behavior of the isomer shift of 6c, 9d and 18f Fe state to the paramagnetic one is of second order. The
sites could be attributed to the additional 3d Mössbauer spectra obtained at room temperature
electrons brought by 18h copper atom neighbors. were analyzed using the correlation between the
They increase the shielding of the 4s electrons, isomer shift and the Wigner–Seitz cell volume,
decrease the s charge density at the nucleus, and calculated using crystallographic data derived from
Structural, Magnetic, Magnetocaloric and Mössbauer Spectrometry Study of Gd2Fe17  xCux
(x = 0, 0.5, 1 and 1.5) Compounds

Rietveld refinement and the near-neighbor environ- 24. P. Alvarez, P. Gorria, V. Franco, J.S. Marcos, M.J. Perez,
ments of the four iron sites. Mössbauer spectrometry J.L.S. Llamazares, I.P. Orench, and J.A. Blanco, J. Phys.
Condens. Matter 22, 216005 (2010).
measurement demonstrated that Cu atoms are 25. P. Alvarez, P. Gorria, J.S. Marcos, J.L.S. Llamazares, and
located in the 18h site which is totally in agreement J.A. Blan, J. Phys. Condens. Matter 25, 496010 (2013).
with the result of the Rietveld refinement. 26. R. Guetari, R. Bez, A. Belhadj, K. Zehani, A. Bezergheanu,
In conclusion, it is significant to note that N. Mliki, L. Bessais, and C.B. Cizmas, J. Alloys Compd. 588,
64 (2014).
Gd2 Fe17x Cux (0  x  1:5) compounds are attrac- 27. S. Charfeddine, K. Zehani, L. Bessais, and A. Korchef, J.
tive alloys due to their good magnetic and magne- Solid State Chem. 238, 15 (2016).
tocaloric properties which are particularly 28. C. Djega-Mariadassou and L. Bessais, J. Magn. Magn.
suitable not only for permanent magnet applica- Mater. 210, 81 (2000).
tions, but also, and mostly, for magnetic refrigera- 29. A.M. Tishin, J. Magn. Magn. Mater. 316, 351 (2007).
30. K.A. Gschneidner Jr., V.K. Pecharsky, and A.O. Tsokol, Rep.
tion technology which is more efficient cost-wise, Prog. Phys. 68, 1479 (2005).
and a lot more environmentally safe than conven- 31. B.F. Yu, Q. Gao, B. Zhang, X.Z. Meng, and Z. Chen, Int. J.
tional gas refrigeration. Refrig. 26, 622 (2003).
32. O. Tegus, E. Brük, L. Zhang, O. Dagula, K.H.J. Buschow,
ACKNOWLEDGMENTS and F.R. de Boer, Phys. B 319, 174 (2002).
33. G.V. Brown, J. Appl. Phys. 47, 3673 (1976).
This work is mainly supported by the CNRS, the 34. V.K. Pecharsky and K.A. Gschneidner Jr., Phys. Rev. Lett.
PHC MAGHREB Project 15MAG07 and the Tuni- 78, 4494 (1997).
sian Ministry of Higher Education and Scientific 35. F. Hu, B. Shen, J. Sun, Z. Cheng, G. Rao, and X. Zhang,
Research and Technology. Appl. Phys. Lett. 78, 3675 (2001).
36. H.M. Rietveld, J. Appl. Crystallogr. 2, 65 (1969).
REFERENCES 37. H. Rietveld, Acta Crystallogr. 22, 151 (1967).
38. C. Djega-Mariadassou, L. Bessais, A. Nandra, and E. Burzo,
1. E. Burzo, A. Chelkovski, and H.R. Kirchmayr, Landolt- Phys. Rev. B 68, 24406 (2003).
Bornstein Handbook (Berlin, 1990). 39. R. Bensalem, W. Tebib, S. Alleg, J.J. Sunol, L. Bessais, and
2. S. Chikazumi, Physics of Ferromagnetism, 2nd edn. (Oxford: J.M. Greneche, J. Alloys Compd. 471, 24 (2009).
Oxford University Press, 1997). 40. R. Fersi, N. Mliki, L. Bessais, R. Guetari, V. Russie, and M.
3. P.J. Cregg and L. Bessais, J. Magn. Magn. Mater. 202, 554 Cabie, J. Alloys Compd. 522, 14 (2012).
(1999). 41. J. Rodriguez-Carvajal, Abstract of the Satellite Meeting on
4. D. Givord and R. Lemaire, IEEE Trans. Magn. 10, 109 Powder Diffraction of the XV Congress of the IUCr
(1974). (Toulouse, France, 1990), p. 127.
5. K.H.J. Buschow, Rep. Prog. Phys. 40, 1179 (1977). 42. J. Rodriguez-Carvajal, M.T. Fernandez-Diaz, and J.L.
6. K.H.J. Bushow, Handbook of Magnetic Materials, vol. 4 Martinez, J. Phys. Condens. Matter 3, 32158 (1991).
(Amsterdam: Elsevier, 1988). 43. J. Rodriguez-Carvajal, Phys. B 192, 55 (1993).
7. E. Belorizky, M.A. Fremy, J.P. Gavigan, D. Givord, and H.S. 44. S. Khazzan, N. Mliki, L. Bessais, and C. Djega-Mariadassou,
Li, J. Appl. Phys. 61, 3971 (1987). J. Magn. Magn. Mater. 322, 224 (2010).
8. K.H.J. Buschow, Rep. Prog. Phys. 54, 1123 (1991). 45. L. Bessais, C. Djega-Mariadassou, D.K. Tung, V.V. Hong,
9. J.J.M. Franse, R.J. Radwanski, and K.H.J. Buschow, and N.X. Phuc, J. Alloys Compd. 455(1), 35 (2008).
Handb. Magn. Mater. 7, 307 (1993). 46. N. Bouchaala, M. Jemmali, T. Bartoli, K. Nouri, I. Hentech,
10. K.H.J. Buschow, Magnetism and Processing of Permanent S. Walha, L. Bessais, and A. BenSalah, J. Solid State Chem.
Magnet, Volume 10 of Handbook of Magnetic Materials 258, 501 (2018).
(Amsterdam: Elsevier, 1997). 47. Y.M. Hao, B. Fu, Y. Zhou, and M. Zhao, Chin. Phys. Lett. 26,
11. J.X. Zhang, L. Bessais, C. Djega-Mariadassou, E. Leroy, and 077501 (2009).
A. Percheron-Guegan, Appl. Phys. Lett. 80, 1960 (2002). 48. G.J. Long, O.A. Pringle, F. Grandjean, and K.H.J. Buschow,
12. L. Bessais, E. Dorolti, and C. Djega-Mariadassou, Appl. J. Appl. Phys. 72, 4845 (1992).
Phys. Lett. 87, 192503 (2005). 49. A.G. Kuchin, A.N. Pirogov, V.I. Khrabrov, A.E. Teplykh,
13. B. Barbara, D. Gignoux, D. Givord, F. Givord, and R. Le- A.S. Ermolenko, and E.V. Belozerov, J. Alloys Compd. 313, 7
maire, Int. J. Magn. 4, 77 (1973). (2000).
14. H.R. Kirchmayr and C.A. Poldy, J. Magn. Magn. Mater. 8, 1 50. G.J. Long and O.A. Pringle, J. Appl. Phys. 74, 504
(1978). (1993).
15. K.J. Strnat, Handbook of Magnetic Materials, vol. 4 (Am- 51. Y.M. Hao, F.W. Wang, P.L. Zhang, X.D. Sun, and Q.W. Yan,
sterdam: Elsevier, 1988). J. Phys. Condens. Matter 11, 6113 (1999).
16. Z. Wang and R.A. Dunlap, J. Phys. Condens. Matter 5, 2407 52. Y.V. Shcherbakova, G.V. Ivanova, N.V. Mushnikov, and I.V.
(1993). Gervasieva, J. Alloys Compd. 308, 15 (2000).
17. Z.W. Li, X.Z. Zhou, and A.H. Morrish, Phys. Rev. B 51, 2891 53. K. Nouri, T. Bartoli, A. Chrobak, J. Moscovici, and L. Bes-
(1995). sais, J. Electron. Mater. 47, 3836 (2018).
18. M. Katter, J. Wecker, and L. Schultz, J. Appl. Phys. 70, 3188 54. G. Pokharel, K.S. Syed Ali, and S.R. Mishra, J. Magn.
(1991). Magn. Mater. 382, 31 (2015).
19. G. Calestani, N. Magnani, A. Paoluzi, L. Pareti, and C. 55. X. Yan, J. Liang, and S. Xie, Phys. Status Solidi (a) 134, 77
Rizzoli, Phys. Rev. B 68, 054424 (2003). (1992).
20. Y. Janssen, S. Chang, A. Kreyssig, A. Kracher, Y. Mozhar- 56. L. Bessais, K. Younsi, S. Khazzan, and N. Mliki, Inter-
ivskyj, S. Misra, and P.C. Canfield, Phys. Rev. B 76, 054420 metallics 19, 997 (2011).
(2007). 57. L. Néel, J. Phys. Radium 9, 148 (1948).
21. L. Bessais, E. Dorolti, and C. Djega-Mariadassou, J. Appl. 58. A.M. Tishin and Y.I. Spichkin, The Magnetocaloric Effect
Phys. 97, 013902 (2005). and Its Applications (Bristol: CRC Press, 2003), p. 476.
22. X.C. Kou, F.R. de Boer, R. Grossinger, G. Wiesinger, H. 59. M. Foldeaki, R. Chahine, and T.K. Bose, J. Appl. Phys. 77,
Suzuki, H. Kitazawa, T. Takamasu, and G. Kido, J. Magn. 3528 (1995).
Magn. Mater. 177, 1002 (1998). 60. V.K. Pecharsky and K.A. Gschneidner, J. Appl. Phys. 86,
23. H. Chen, Y. Zhang, J. Han, H. Du, Ch. Wang, and Y. Yang, 565 (1999).
J. Magn. Magn. Mater. 320, 1382 (2008). 61. J. Inoue and M. Shimizu, J. Phys. F 12, 1811 (1982).
 Saidi, Nouri, Walha, Dhahri, Kabadou, Jemmali, and Bessais

62. P.E. Brommer, Phys. B 154, 197 (1989). 67. E. Burzo, Rep. Prog. Phys. 61, 1099 (1998).
63. H. Liu, D. Wang, S. Tang, Q. Cao, T. Tang, B. Gu, and Y. Du, 68. L. Bessais, C. Djega-Mariadassou, and E. Koch, J. Phys.
J. Alloys Compd. 346, 314 (2002). Condens. Matter 14, 8111 (2002).
64. X.B. Liu and Z. Altounian, J. Magn. Magn. Mater. 292, 83 (2005).
65. E. Koch and W. Fischer, Z. Kristallogr. 211, 251 (1996). Publisher’s Note Springer Nature remains neutral with
66. M. Forker, A. Julius, M. Schulte, and D. Best, Phys. Rev. B regard to jurisdictional claims in published maps and institu-
57, 11565 (1998). tional affiliations.