A tale of two metals: New cerium iron borocarbide intermetallics grown from rare-earth/transition metal eutectic fluxes

Article abstract of DOI:10.1021/ja303370j

R₃₃Fe_14-xAl_x+yB_25-yC ₃₄ (R = La or Ce; x ≥ 0.9; y ≥ 0.2) and R₃₃Fe _13-xAl_xB₁₈C₃₄ (R = Ce or Pr; x < 0.1) were synthesized from reactions of iron with boron, carbon, and aluminum in R-T eutectic fluxes (T = Fe, Co, or Ni). These phases crystallize in the cubic space group Im3m (a = 14.617(1) A, Z = 2, R₁ = 0.0155 for Ce₃₃Fe_13.1Al_1.1B_24.8C₃₄, and a = 14.246(8) A, Z = 2, R₁ = 0.0142 for Ce ₃₃Fe₁₃B₁₈C₃₄). Their structures can be described as body-centered cubic arrays of large Fe₁₃ or Fe ₁₄ clusters which are capped by borocarbide chains and surrounded by rare earth cations. The magnetic behavior of the cerium-containing analogs is complicated by the possibility of two valence states for cerium and possible presence of magnetic moments on the iron sites. Temperature-dependent magnetic susceptibility measurements and Moessbauer data show that the boron-centered Fe₁₄ clusters in Ce₃₃Fe_14-xAl _x+yB_25-yC₃₄ are not magnetic. X-ray photoelectron spectroscopy data indicate that the cerium is trivalent at room temperature, but the temperature dependence of the resistivity and the magnetic susceptibility data suggest Ce^3+/4+ valence fluctuation beginning at 120 K. Bond length analysis and XPS studies of Ce₃₃Fe ₁₃B₁₈C₃₄ indicate the cerium in this phase is tetravalent, and the observed magnetic ordering at T_C = 180 K is due to magnetic moments on the Fe₁₃ clusters.

Full text of DOI:10.1021/ja303370j

Article
pubs.acs.org/JACS
A Tale of Two Metals: New Cerium Iron Borocarbide Intermetallics
Grown from Rare-Earth/Transition Metal Eutectic Fluxes
Patricia C. Tucker,† Jason Nyffeler,† Banghao Chen,† Andrew Ozarowski,‡ Ryan Stillwell,‡
and Susan E. Latturner*,†
†Department of Chemistry and Biochemistry, Florida State University, Tallahassee, Florida 32306, United States
‡National High Magnetic Field Laboratory, Florida State University, Tallahassee, Florida 32310, United States
S
* Supporting Information
ABSTRACT: R33Fe14−xAlx+yB25−yC34 (R = La or Ce; x ≤ 0.9; y ≤ 0.2) and R33Fe13−xAlxB18C34 (R =
Ce or Pr; x < 0.1) were synthesized from reactions of iron with boron, carbon, and aluminum in R−
T eutectic fluxes (T = Fe, Co, or Ni). These phases crystallize in the cubic space group Im3m (a =
̅
14.617(1) Å, Z = 2, R1 = 0.0155 for Ce33Fe13.1Al1.1B24.8C34, and a = 14.246(8) Å, Z = 2, R1 = 0.0142
for Ce33Fe13B18C34). Their structures can be described as body-centered cubic arrays of large Fe13 or
Fe14 clusters which are capped by borocarbide chains and surrounded by rare earth cations. The
magnetic behavior of the cerium-containing analogs is complicated by the possibility of two valence
states for cerium and possible presence of magnetic moments on the iron sites. Temperature-
dependent magnetic susceptibility measurements and Mossbauer data show that the boron-centered
̈
Fe14 clusters in Ce33Fe14−xAlx+yB25−yC34 are not magnetic. X-ray photoelectron spectroscopy data
indicate that the cerium is trivalent at room temperature, but the temperature dependence of the resistivity and the magnetic
susceptibility data suggest Ce3+/4+ valence fluctuation beginning at 120 K. Bond length analysis and XPS studies of
Ce33Fe13B18C34 indicate the cerium in this phase is tetravalent, and the observed magnetic ordering at TC = 180 K is due to
magnetic moments on the Fe13 clusters.
particularly interesting phenomena due to the fact that the
energy of the Ce 4f electron is often close to the Fermi level,
leading to the possibility of two valence states (Ce3+ or Ce4+)
and variation in valence with temperature or pressure. This can
result in mixed valence structures (Ce23Ru7Cd4),10 valence
fluctuation (Ce2RuZn4),11 and heavy fermion behavior
(CeRhIn5).12
In this work, two structurally related cerium iron borocarbide
intermetallics have been grown from Ce/T eutectic fluxes. Both
structures feature several cerium sites surrounding clusters of
iron capped with borocarbide chains. Ce33Fe13.1Al1.1B24.8C34
exhibits cerium valence fluctuation; its boron-centered Fe14
clusters do not possess magnetic moments. Ce33Fe13B18C34 is
comprised of tetravalent Ce ions (possibly the result of
chemical pressure) and cuboctahedral Fe13 clusters which order
ferromagnetically.
The Fe13 and Fe14 clusters are highly unusual structural
building blocks. R33Fe14−xAlx+yB25−yC34 (R = La, Ce) and
R33Fe13−xAlxB18C34 (R = Ce, Pr) occupy the relatively
unexplored phase space between intermetallic compounds
with isolated (and usually nonmagnetic) iron sites such as
La3.67FeC6 and ErFe4Al9Si6, and compounds with extended iron
building blocks featuring strong magnetic moments, such as
R6Fe14−xAlx and Nd2Fe14B.13−16 The variation in magnetic
properties of the iron clusters in the title phases is due to
different Fe−Fe distances and number of nearest neighbor iron
INTRODUCTION
■
Metal flux synthesis has proven to be a vital tool for the
discovery and crystal growth of new solids. This synthesis
technique allows for the solvation of refractory elements in low
melting metal solvents; the low reaction temperatures afforded
by this method can promote the growth of new intermetallic
compounds with complex structures.1 Expanding from single
metal solvents to mixed fluxes comprised of two metals can
allow for lowered solvent melting points through formation of
eutectics; it also increases the number of reactants that are
soluble and reactive in the flux. The use of mixed metal eutectic
fluxes has aided the discovery of a broad range of new
intermetallic phases and Zintl phase compounds such as
Ca21Ni2Zn36 (grown in Ca−Zn eutectic), Co7Zn3Sn8 (grown in
Zn−Sn eutectic), and LiCa2C3H (grown in Ca−Li mixture).2−4
Our explorations of reactions in La−Ni eutectic flux (mp 517
°C) have yielded several new multinary intermetallics
comprised of lanthanum, a transition metal, and one or more
main group elements; examples include La21T8M7C12 (T = Mn,
Fe; M = Ge, Sn, Sb, Te, Bi), La2Ni2−xRuxAl, and La14Sn-
(MnC6)3.5−7
Expanding on the La−Ni eutectic work, Ce−T and Pr−T (T
= Fe, Co, Ni) eutectic fluxes are now being investigated as
synthesis media for magnetic intermetallic materials. Phases
containing both a paramagnetic rare earth and a transition
metal may feature complex magnetic behavior as the moments
on these atoms couple in various ways at different temper-
atures, as is seen for compounds such as SmCo5 and
PrCo2P2.8,9 Cerium-containing intermetallics can exhibit
Received: April 8, 2012
Published: June 25, 2012
© 2012 American Chemical Society
12138
dx.doi.org/10.1021/ja303370j | J. Am. Chem. Soc. 2012, 134, 12138−12148

Journal of the American Chemical Society
Article
was negligible (below 1 mol %). Small amounts of aluminum (1−2%)
were indicated in the R33Fe14−xAlx+yB25−yC34 and PR33Fe13−xAlxB18C34
crystals. No aluminum was indicated in the data for Ce33Fe13B18C34
samples. The carbon and boron content was not able to be determined
due to the limitation of the EDXS in detecting the characteristic X-rays
of light elements. Quantitative determination of the carbon content of
Ce33Fe14−xAlx+yB25−yC34 was carried out by combustion analysis
(Atlantic Microlabs). A 60 mg sample of Ce33Fe14−xAlx+yB25−yC34
crystals was sent for analysis to determine the amount of C present.
Two runs of 30 mg each were performed, yielding carbon mass
percents of 5.46% and 5.57%.
X-ray Photoelectron Spectroscopy. Analyses of cerium valence
and light elements were performed on Ce33Fe14−xAlx+yB25−yC34 and
Ce33Fe13B18C34 samples previously screened by EDS, using a Physical
Electronics PHI 5100 series XPS with a nonmonochromated dual
anode (Al and Mg) source having a single channel hemispherical
energy analyzer. Large single crystals were affixed to a carbon coated
sample puck using carbon tape. The Al X-ray source was used to study
higher binding energies. To remove surface oxides and flux residue,
samples were sputtered by Ar+ ions as the sample stage was rotated.
Spectra were taken after every 5−10 min of sputtering to monitor the
disappearance of surface oxide peaks and appearance of new species.
Once no more changes in spectra were observed, sputtering was
stopped.
atoms, as well as the extent of aluminum substitution on iron
sites. Tailoring magnetic properties by careful control of
substitution in phases such as Nd2Fe14−xMxB and SmCo5−xMx
is a topic of widespread interest;16 studying this process in the
smaller iron clusters available in these new Ce/Fe/B/C phases
may shed light on how magnetic moments develop in hard
magnetic materials.
EXPERIMENTAL PROCEDURE
■
Synthesis. Starting materials were stored and handled in an argon-
filled drybox. Powders of the following elements were used in
reactions: carbon black (Strem Chemicals 95−97%), boron (Strem
Chemicals 95−97%), iron (Alfa Aesar 99%), manganese (Alfa Aesar
99%), aluminum (Alfa Aesar 99%), nickel (Alfa Aesar 99%), and iron-
57 (Isoflex USA 96.63% enrichment). Iron-57 was used for the
synthesis of an enriched sample of Ce33Fe14−xAlx+yB25−yC34 for
Mossbauer spectroscopy. R−T eutectics were either purchased (La−
̈
Ni, Alfa-Aesar; 67% La with mp 517 °C) or made by arc-melting
appropriate amounts of rare earth and transition metals under argon.
For instance, Ce−Co eutectic was made from cerium ingot (Alfa
Aesar, 99.999%) and cobalt slug (Alfa Aesar 99.95%), by arc melting a
76:24% mole ratio of these elements (mp 424 °C)17 on a water-cooled
copper hearth into a button that was turned over and remelted several
times to ensure homogeneity of the flux material. Similarly, Ce−Fe
eutectic was made from cerium ingot and iron slug (Alfa Aesar 99%) in
a 83:17% mole ratio (mp 592 °C); Ce−Ni eutectic from cerium ingot
and nickel slug (Alfa Aesar 99%) in a 78:22% mole ratio (mp 477 °C);
Pr−Fe eutectic from praseodymium ingot (Alfa Aesar 99.9%) and Fe
slug using a 79:21% mole ratio (mp 620 °C); Pr−Ni eutectic from
praseodymium ingot and nickel slug in a 81:19% mole ratio (mp 460
°C).17 These brittle eutectic ingots were fragmented into approx-
imately 1 mm3 size pieces; 1.5 g of eutectic were used per reaction.
Reactions in the various R−T eutectics were prepared by
sandwiching C, B, Ni, Al, and either Mn, Fe, or Fe-57 (combined in
a 1:1:1:1:1 mmol ratio) between layers of R−T eutectic, with more
flux on the top layer than the bottom one. Reactants and flux were
loaded into either alumina or steel crucibles (ca. 8 mm inside diameter,
30 mm length) and a second alumina crucible filled with Fiberfrax was
inverted and placed on top of the reaction crucible to act as a filter
during centrifugation. These crucible configurations were placed into
silica tubes, which were flame-sealed under a vacuum of 10−2 Torr.
The ampules were then heated to 950 °C in 3 h, held at this
temperature for 12 h, and then cooled to 850 °C in 10 h. The reaction
mixtures were subsequently annealed for 48 h at 850 °C and then
cooled to 600 °C in 84 h. At 600 °C the ampules were removed from
the furnace, quickly inverted, and placed into a centrifuge to decant the
molten flux. (Reactions in the higher-melting Ce−Fe eutectic were
instead cooled to 700 °C and centrifuged.) Products from these
reactions were stored in an argon-filled drybox in order to prevent
oxidation.
X-ray Diffraction. Selected single crystals of each phase were
mounted on cryo loops using paratone oil. X-ray diffraction data were
collected at 150 K on a Bruker SMART APEX2 CCD diffractometer
equipped with a Mo-target X-ray tube (λ = 0.71073). The data sets
were recorded as ω scans at 0.3° step width and integrated with the
Bruker SAINT software. Data were corrected for absorption effects
using the multiscan method (SADABS).18 The structures were refined
in the centrosymmetric space group Im3m (space group number 229)
̅
by full matrix least-squares procedures on |F2| using the SHELX-97
software package.19 The rare earth and iron atoms were located using
direct methods, and the light atoms were indicated by residual electron
density peaks of 5−10 e−/Å3 in the difference Fourier maps. Light
atom sites were initially assigned as boron. Allowing their occupancy to
vary did not distinguish boron from carbon (the scattering factors of
these elements are too similar), so assignments of these sites were
based on comparison of observed bond lengths to those in other rare
earth borocarbides (see discussion). During final refinement cycles
occupancies of each site were allowed to vary to identify possible site
mixing. Unit cell parameters and crystallographic data collection
information are found in Tables 1 and 2; atomic coordinates and
displacement parameters are found in the Supporting Information.
Table 1. R33Fe14−xAlx+yB25−yC34 and R33Fe13−xAlxB18C34
Phases Grown from R/T Fluxes and Parameters from Single
Crystal Diffraction Data Refinements
compound
unit cell (Å)
r-factor (R1/wR2)
Stoichiometric syntheses of R33Fe14−xAlx+yB25−yC34 and
R33Fe13−xAlxB18C34 were attempted by combining elemental forms of
R (R = La, Ce, Pr, Nd, Sm), Fe, B, and C in appropriate ratios (for
instance, a 6.6:2.8:5.0:6.8 mmol ratio for R33Fe14B25C34) in an alumina
crucible. These reactions were prepared in an argon-filled drybox, and
the crucible was placed in a fused silica tube and sealed under a
vacuum of 10−2 Torr. The ampule was then heated to 950 °C in 3 h,
held at this temperature for 168 h, and then cooled to 25 °C in 3 h.
Products were stored in an argon-filled drybox to prevent oxidation of
the powders.
Elemental Analysis. Elemental analysis was performed on all
samples using a JEOL 5900 scanning electron microscope (SEM) with
energy-dispersive X-ray spectroscopy (EDXS) capabilities. Flux-grown
crystals from each reaction were affixed to an aluminum SEM puck
using carbon tape, fractured to expose inner regions, and positioned so
that flat faces were perpendicular to the electron beam. Samples were
analyzed using a 30 kV accelerating voltage and an accumulation time
of 60s. A rare earth:iron ratio of 7:3 was consistently found for
R33Fe14−xAlx+yB25−yC34 and R33Fe13−xAlxB18C34 phases; nickel content
La33Fe13.9(1)Al0.3(1)B24.8(1)C34
Ce33Fe13.1(1)Al1.1(1)B24.8(1)C34
Ce33Mn14Al0.1(1)B24.9(1)C34
Ce33Fe13B18C34
14.8560(8)
14.617(1)
14.615(1)
14.246(8)
14.3881(9)
0.0184/0.0326
0.0155/0.0350
0.0155/0.0345
0.0142/0.0285
0.0127/0.0271
Pr33Fe12.9(1)Al0.1(1)B18C34
Powder X-ray diffraction data were collected using a Rigaku Ultima
III diffractometer with a Cu radiation source and a CCD detector.
Samples were ground with a small amount of silicon to act as an
internal standard. Experimental powder patterns were compared to
calculated patterns based on single crystal data to confirm the presence
of the desired phase. Unit cell parameters were determined by using
the accompanying refinement software contained in the JADE
software (a = 14.850(9), 14.581(4), and 14.390(1) Å for
La3 3 Fe1 4 − x Alx + y B2 5 − y C3 4
PR33Fe13−xAlxB18C34, respectively). Byproducts were identified by
comparison with patterns in the JADE software database.
, Ce3 3 Fe1 4 − x Alx + y B2 5 − y C3 4 , and
12139
dx.doi.org/10.1021/ja303370j | J. Am. Chem. Soc. 2012, 134, 12138−12148

Journal of the American Chemical Society
Article
Table 2. Crystallographic Data Collection Parameters for Selected R/T/B/C Phases
Ce33Fe13.1Al1.1B24.8C34
6055.58
Ce33Fe13B18C34
5952.93
cubic
Pr33Fe12.9Al0.1B18C34
formula weight (g/mol)
crystal system
space group
a (Å)
5977.8
Im3m
̅
14.617(1)
2
14.246(8)
14.3881(9)
2
Z
2
volume
3123.2(4)
6.44
2891(3)
2978.6(3)
6.67
density, calc (g/cm3)
radiation
6.84
Mo Kα
temperature (K)
index ranges
150
−18 ≤ h ≤ 19
−19 ≤ k ≤ 19
−19 ≤ l ≤ 19
1.97−28.21
17 445
−18 ≤ h ≤ 18
−18 ≤ k ≤ 18
−18 ≤ l ≤ 18
2.02−28.25
15 157
−18 ≤ h ≤ 18
−19 ≤ k ≤ 19
−18 ≤ l ≤ 18
2.00−28.20
16 857
theta range (deg)
reflections collected
unique data/parameters
419/42
388/36
402/38
μ (mm−1
R1/wR2
)
26.41
28.51
29.44
0.0155/0.0350
0.0186/0.0354
1.015/−1.350
0.0142/0.0285
0.0143/0.0285
0.826/−0.679
0.0127/0.0271
0.0137/0.0273
0.834/−1.430
R1/wR2 (all data)
residual peak/hole (e−/Å3)
Magnetic Susceptibility. Magnetic measurements were carried
out with a Quantum Design MPMS SQUID magnetometer. For each
sample, a large single crystal was weighed and then placed at the 3 cm
mark of a 12 cm piece of kapton tape. The kapton tape was folded in
half to enclose the crystal. Then the piece of tape containing the crystal
was placed in a plastic straw and then loaded into the magnetometer.
Both field-cooled (FC) and zero-field cooled (ZFC) measurements
were collected at an applied field of 100 Oe with a temperature range
of 2−300 K, and field dependence data were collected from 0 to 70
000 Oe at several temperatures.
transition metals (T = Fe, Co, Ni, Cu) contain a low-melting
eutectic in the rare-earth rich region.17 These eutectic melts
have proven to be excellent solvents for iron, carbon, and most
other elements, leading to formation of many new R/Fe/C/X
phases. A recent review of known ternary R/T/C carbides
classified such compounds into two groups: carbometallates
(such as La3.67FeC6, featuring isolated transition metal atoms
surrounded by carbon), or transition metal-rich carbides (such
as Nd2Fe17C3−x, featuring extensive Fe−Fe bonding).20
However, several new phases have recently been discovered
which fit between these classifications. Compounds such as
La21Fe8Sn7C12 (grown in La/Ni flux),5 Dy15Fe8C25,21 and the
title phases R33Fe14−xAlx+yB25−yC34 (R = La, Ce) and
R33Fe13−xAlxB18C34 (R = Ce, Pr) feature iron clusters of 4−
14 atoms capped around the edges by carbon and/or boron.
The intermediate size of these iron building blocks can lead to
complex magnetic behavior. This complexity is furthered in the
cerium iron borocarbide title phases due to the variable valency
of cerium.
Synthesis. R33Fe14−xAlx+yB25−yC34 (R = La, Ce) and
R33Fe13−xAlxB18C34 (R = Ce, Pr) grow from various R−T
eutectic fluxes as large silver faceted spheroids up to 1−2 mm in
diameter; see Figure 1. A manganese analog of one of the
structures, Ce33Mn14Al0.1B24.9C34, can also be synthesized. The
compounds degrade slowly in air (over several days) and react
rapidly with water. While product can be obtained from all the
fluxes that were investigated, the highest yields were found for
reactions in Ce−Co eutectic for Ce33Fe14−xAlx+yB25−yC34 (70%
yield based on carbon) and Pr−Ni flux for Pr33Fe13−xAlxB18C34
(40% yield based on carbon; yield was significantly lower in
Pr−Co). Yield increased with the addition of C and B, up to 1.2
mmol of each element added. Beyond this amount, unreacted C
and B were observed in the centrifuged ampules, indicating a
possible solubility limit. The Ce−Co and Ce−Ni fluxes are
more convenient to use compared to the higher melting Ce−Fe
eutectic, although they present the risk of cobalt or nickel
incorporation into products. These elements were not observed
in the EDS analyses of Ce33Fe14−xAlx+yB25−yC34, but trace
amounts of cobalt were detected in the manganese analog
Ce33Mn14Al0.1B24.9C34 grown from reactions in Ce−Co flux.
̈
Mossbauer Spectroscopy. Mossbauer Spectroscopy measure-
̈
ments on Ce33Fe14−xAlx+yB25−yC34 grown using enriched 57Fe were
carried out using a SEE Co. Mossbauer spectrometer installed with a
̈
gamma source, 57Co embedded in Rh matrix, with an initial intensity
of 110 milliCurie manufactured by Isotope. Temperature was
controlled with a Janis helium cryostat. Measurements were performed
at 3 mm/s at 5.5, 78, 150, and 294 K, and at 10 mm/s at 5.5 and 78
K.
Solid State NMR Spectroscopy. 11B MAS NMR spectra of
Ce33Fe14−xAlx+yB25−yC34 were collected on a Varian Unity Inova 500
MHz wide-bore NMR spectrometer equipped with a three channel
MAS probe. The Larmor frequency of 11B was 160.538 MHz. Single
crystals of Ce33Fe14−xAlx+yB25−yC34 were ground with NaCl in a drybox
(2:1 sample to NaCl by mass) to facilitate spinning of the sample in
the magnetic field; the mixture was packed into a 4 mm zirconia rotor
sealed with airtight screw caps. A typical small angle pulse (less than
15°) was applied to ensure quantitative spectra with a recycle delay of
5 s. Data were collected at 25 °C, and a 14 kHz spin rate was used. A
total of 14 920 scans were collected for the NMR spectrum. The
spectra were referenced to solid NaBH4 at −42.16 ppm (with respect
to BF3OEt2 at 0 ppm). Attempts were made to obtain 13C MAS NMR
data, but no resonances were observed.
Resistivity. Resistivity measurements were conducted on
Ce33Fe14−xAlx+yB25−yC34 by a conventional four-point dc method on
a Physical Property Measurement System (PPMS) by Quantum
Design. A crystal (size: 1 mm × 1 mm × 1 mm) was mounted on the
sample holder of a 4He probe with a small amount of glue and
connected to the electrodes of the sample holder with 0.001 in.
diameter gold wires using silver paint. Measurements were carried out
from 1.8 to 300 K, using an applied excitation current of 3 mA.
RESULTS AND DISCUSSION
■
All of the binary alloy phase diagrams of early rare earth
elements (R = La, Ce, Pr, Nd) combined with late first row
12140
dx.doi.org/10.1021/ja303370j | J. Am. Chem. Soc. 2012, 134, 12138−12148

Journal of the American Chemical Society
Article
Pr3+ ions yielding the latter. This is confirmed in both flux and
stoichiometric synthesis. La33Fe14−xAlx+yB25−yC34 is formed
from either La/T flux reactions (as crystals) or as silver
powder from stoichiometric combination of the elements.
Pr33Fe13−xAlxB18C34 is likewise isolated as crystals from Pr/T
fluxes or as silver powder in stoichiometric reactions. Cerium
ions can form both structures. For Ce flux reactions, the
formation of either Ce33Fe14−xAlx+yB25−yC34 or Ce33Fe13B18C34
is determined by the crucible used for synthesis. For reactions
carried out in alumina crucibles, Ce33Fe14−xAlx+yB−yC34 is
formed, possibly induced by leaching of trace Al from the
crucible. Reactions carried out in steel crucibles produce
Ce33Fe13B18C34 instead. Trace elements in the steel may
promote the formation of this structure. Attempts at
stoichiometric synthesis of the cerium phases produce only
the Ce33Fe14−xAlx+yB25−yC34 phase. The powder pattern
matches that calculated from the single crystal data for this
structure (Figure S1, Supporting Information). The amount of
Al incorporation was unable to be determined because no
single crystals were obtained, although trace amounts of Al
were indicated in the EDS analysis of the powder. Flux
reactions and stoichiometric syntheses using Nd and Sm did
not form either R33Fe14−xAlx+yB25−yC34 or R33Fe13−xAlxB18C34
phases; these elements are too small to occupy the rare earth
sites in these structures.
Light elements such as boron and carbon cannot be analyzed
by SEM-EDXS; large crystals of Ce33Fe14−xAlx+yB25−yC34 were
therefore studied using XPS to determine whether or not these
elements were present. Ar+ ion sputtering was used to remove
surface species from samples before XPS measurement. No
boron peaks were observed before sputtering, possibly
indicating coating of the crystals by traces of flux or surface
enrichment by other elements. After sputtering, a peak appears
at 196 eV, which is in the expected B 1s binding energy
region.26 The presence of carbon was also confirmed by the
XPS data with the C 1s binding energy of 284 eV similar to that
of other carbides such as Fe3C (283.9 eV), SiC (283.8 eV), and
BaC2 (283.5 eV).26 Carbon content was quantified by
combustion analysis on samples of Ce33Fe14−xAlx+yB25−yC34
(Atlantic Microlabs), which indicated 5.5% carbon by mass.
This is lower than the 6.7% value expected from the
stoichiometry, which may be due to flux coating of the
samples, or oxidation during handling.
Figure 1. SEM image of crystals of Ce33Fe13.1Al1.1B24.8C34 grown from
Ce/Fe flux, displaying truncated octahedral growth habit.
This parallels what has been observed in previous reactions in
La/Ni flux. For instance, in La/Ni flux syntheses of La6T14−xAlx,
reactions with T = Fe yield La6Fe14−xAlx with no nickel
incorporated; if T = Mn is used instead, the La6Mn14−xAlx
products do show doping of small amounts of nickel on some
of the manganese sites.22 The reactivity of transition metals in
these R/T eutecticswhether as added reactants or as
components of the flux eutecticappears to be Fe > Mn ≫
Co, Ni.
The synthesis of the title phases is complicated by the fact
that additional elements (Al and Ni) must be present in the flux
to aid in their growth. A series of reactions were carried out to
determine which elements were necessary for product
formation. Reactions that did not contain both Ni and Al
yielded predominantly pseudobinary phases such as Ce-
Fe2−xCox. Incorporation of trace amounts of aluminum may
be required to stabilize the structures. This is also seen for
phases such as LaFe13−xAlx and La6Fe14−xAlx; the binaries
LaFe13 and La6Fe13 are not known.22,23 EDS analyses and
structural refinement of single crystal XRD data for several
“R33Fe14B25C34” crystals consistently indicates substitution of Al
on the Fe2 site and the B3 site, resulting in a stoichiometry of
R33Fe14−xAlx+yB25−yC34 (see Table 1). These phases were
initially grown in alumina crucibles without Al reactant; the
strongly reducing flux reacted with the crucible, adding
aluminum metal to the reaction. When aluminum was
deliberately added to subsequent reactions, the yield increased
drastically. A very small amount of aluminum substitution is
also seen in Pr33Fe13−xAlxB18C34 (x ≈ 0.1), but is not seen for
the Ce33Fe13B18C34 analog.
Nickel does not appear to be incorporated into the products
at all, so its role in the reaction may be to aid in solubilizing the
boron reactant or to catalyze or nucleate the crystal growth of
the phase. Similar behavior was observed for the gallium flux
syntheses of Yb3Ga7Ge3 and β-SiB3, which require the presence
of Pd and Cu, respectively.24,25 Reactions that did not include
Ni resulted in a very small yield of product and unreacted
boron. Attempts at using copper or other elements instead of
nickel resulted in unreacted boron and crystals of CeFe2−xCox
instead of the title phases. Despite the complexity of the
synthesis, several analogs of each structure can be obtained
reproducibly.
Structure of R33M14−xAlx+yB25−yC34 (R = La, Ce; M = Fe,
Mn). This structure crystallizes in the cubic space group Im3m
̅
with unit cell parameters decreasing with the size of the rare
earth (see Table 1). The structure can be viewed as a bcc
packing of borocarbide-capped Fe14B clusters as shown in
Figure 2. The reaction of iron with lightweight metals or
metalloids in R/T fluxes to form iron clusters or layers capped
by the light elements is a reoccurring theme. In addition to the
two types of iron clusters reported here, carbon-capped
tetrahedral Fe4 units are found in the La21Fe8M7C12 structure,5
and aluminum-capped iron layers in La6Fe14−xAlx, both grown
in La/Ni flux.22 The iron cluster in R33Fe14−xAlx+yB25−yC34 is a
face capped cube (or tetrakis hexahedron) of 14 iron atoms,
with six Fe2 atoms (12e sites) at the capping apexes and eight
Fe1 sites (16f sites) defining the cube. The atom that centers
this cluster (on the 2a site) is much closer to the Fe1 sites than
the Fe2 sites (2.231(1) vs 2.946(1) Å in the
Ce33Fe14−xAlx+yB25−yC34 analog). The small electron density at
this site indicated that it is occupied by a light atom, either B or
C. The 2.231(1) Å distance from this site to the Fe1 atoms is
The identity of the rare earth determines whether
R33Fe14−xAlx+yB25−yC34 or R33Fe13−xAlxB18C34 is formed, with
larger La3+ ions producing the former structure, and smaller
12141
dx.doi.org/10.1021/ja303370j | J. Am. Chem. Soc. 2012, 134, 12138−12148

Journal of the American Chemical Society
Article
such as Nd2Fe14B and FeB.23,28 Boron is therefore a more likely
option. In the Ce33Mn14B25C34 analog, the corresponding Mn−
B bond length of 2.203(1) Å is within the range observed in
MnB4 (2.049−2.218 Å), further supporting the assignment of
this site as boron.29 For all analogs, allowing the occupancy to
vary resulted in a higher than 100% occupancy. This, and the
slightly long Fe−B bond lengths, points to incorporation of a
larger, heavier element on this 2a site. It was therefore modeled
as a mixture of boron and aluminum; this improved the
refinement (yielding more stable thermal parameters) and
indicated incorporation of 10−20% Al on this site.
The Fe14B cluster in Ce33Fe14−xAlx+yB25−yC34 features Fe−Fe
bond lengths of 2.462(1) and 2.576(1) Å. These are similar to
the 2.55 Å bond length in the Fe4 clusters of La21Fe8Sn7C12 and
well within the range of 2.429−2.740 Å observed for
Ce2Fe17.5,30 In the manganese analog, the Mn−Mn bond
lengths (2.462(1) and 2.544(1) Å) in the Mn14B cluster
compare well with those found in β-Mn which range from
2.363 to 2.680 Å.31 A small amount of Al substitution (15%) is
observed for the Fe2 sites of Ce33Fe14−xAlx+yB25−yC34. No Al
substitution on the Mn sites was observed based on the
refinement of the occupancies of these atoms from the single
crystal data of Ce33Mn14B25C34, but trace amounts of Co
observed in the EDXS analysis suggest there may be
incorporation of cobalt from the flux on these sites.
The iron clusters are capped at the six vertices by carbon
atoms, and capped on 12 edges by 4-atom linear chains of
carbon and boron (Figure 2b). The title compounds thus add
to the small database of reported structures containing
borocarbide chains, which includes R10B9C12 (R = La, Ce, Pr,
Nd), R5B2C5 (R = Ce−Tm), R5B4C5 (R = Ce, Pr, Nd), RBC
(R = Ce, Pr, Nd), R15B6C20, CeB2C2, Lu3BC3, and
Gd4B3C4.32−39 Assignments of the carbon and boron sites in
these chains were made based on bond lengths within the chain
and distances to cerium atoms surrounding this chain, and
comparison to other metal borocarbide structures. The B1 site
caps an edge of the Fe14B cluster; it is bonded to 2 Fe atoms, 1
C atom, and 4 Ce atoms. The associated bond lengths are
shown in Figure 2b and listed in Table 3. These bond lengths
compare well with those reported for similar compounds such
as CeB2C2, with a Ce−B bond distance of 2.8402(2) Å and B−
C bond distances of 1.5312(9)−1.6220(6) Å.37 The rest of the
Ce33Fe14−xAlx+yB25−yC34 borocarbide chain extends from the B1
atom, with bond distances between the C1−B2−C2 atoms of
C1−B2 of 1.45(1) Å and B2−C2 of 1.47(1) Å. These are
similar to the 1.44−1.48 Å range of bond lengths observed in
other compounds containing linear C−B−C units such as
Gd4B3C4 and Lu3BC3.38,39 Termination of these chains by C
atoms is thought to be favored due to the surrounding
octahedral coordination environment, and on the basis of the
higher electronegativity of carbon and its affinity for the
electropositive rare earth ions.40 The distances from atoms in
the borocarbide chain of Ce33Fe14−xAlx+yB25−yC34 to the
surrounding cerium ions exhibit the expected trend of shorter
Ce−C bonds (all below 2.8005(6) Å) and longer Ce−B bonds
(all above 2.871(3) Å); see Figure 2f.
Figure 2. (a) Structure of Ce33Fe13.1Al1.1B24.8C34. Red spheres are iron,
green spheres are cerium, black spheres are carbon, blue spheres are
boron. The boron-centered Fe14 clusters are drawn as red polyhedra
and the isolated carbon anion in octahedral coordination as gray
polyhedra. (b) Borocarbide-capped Fe14 cluster and associated bond
lengths. (c) Ce1 coordination environment (48k site). (d) Ce2
coordination environment (12d site). (e) Ce3 coordination environ-
ment (6b site). (f) Coordination environment around the borocarbide
chain.
The R33Fe14−xAlx+yB25−yC34 structure also features an isolated
carbon site (C4) surrounded by six cerium cations in an
octahedral configuration. These carbide anions (presumably
C4−) and their coordination environment are depicted as gray
octahedra in Figure 2. The six R−C bond lengths are all equal
and are only slightly affected by replacing Fe with Mn in the
much longer than typical Fe−C bonds in intermetallics (which
are generally 2.0 Å or shorter).5,27 It is only slightly longer than
the Fe−B bond lengths of 2.118 and 2.180 Å found in phases
12142
dx.doi.org/10.1021/ja303370j | J. Am. Chem. Soc. 2012, 134, 12138−12148

Journal of the American Chemical Society
Article
Table 3. Bond Lengths (in Å) for Selected R/T/B/C Phases
Ce33Fe13.1Al1.1B24.8C34
Ce33Fe13B18C34
Pr33Fe12.9Al0.1B18C34
Ce1−C4
Ce1−C1
Ce1−B1
Ce1−Ce1
Ce1−Ce2
Ce2−Ce3
B1−C1
B1−Fe1
B1−Fe2
C1−B2
2.6576(4)
2.672(1)
2.871(3)
3.7274(6)
3.8089(4)
3.6543(3)
1.507(11)
2.035(6)
2.462(4)
1.445(11)
2.941(4)
3.100(4)
1.469(11)
2.519(8)
2.5848(3)
2.8007(5)
2.231(1)
2.462(1)
2.576(1)
1.917(17)
2.447(17)
R1−C4
R1−C1
R1−B1
R1−R1
2.630(1)
2.580(1)
3.053(5)
3.601(2)
2.6471(3)
2.5912(5)
3.058(4)
3.6356(4)
R2−R3
B1−C1
C2−B1
C3−B2
R1-B2
3.562(1)
1.501(11)
1.499(11)
1.579(19)
2.854(6)
3.5970(2)
1.469(10)
1.500(10)
1.667(21)
2.913(6)
B2−Ce2
B2−Ce1
C2−B2
R1−C3
R1−Fe1
R2−C2
R2−B1
2.563(2)
3.148(2)
2.519(1)
2.946(5)
2.588(2)
3.1581(6)
2.5450(3)
2.998(4)
C2−Ce3
C2−Ce2
C2−Ce1
Fe1−B3
Fe1−Fe2
Fe1−Fe1
Fe2−C3
C3−Ce3
Fe1−B2
Fe1−Fe1
Fe1−Fe2
Fe1−C1
2.236(8)
2.579(2)
2.579(2)
1.947(8)
2.200(8)
2.598(1)
2.598(1)
1.976(7)
structure (Ce−C distances 2.6576(4) and 2.657(2) Å for the
Fe and Mn analogs, respectively).
Structure of R33Fe13−xAlxB18C34 (R = Ce, Pr).
shorter than similar bonds in Ce33Fe14−xAlx+yB25−yC34, with
Ce−Ce bonds in particular shrinking from 3.6543(3) and
3.7274(6) Å in the latter to 3.562(1) and 3.601(2) Å in
Ce33Fe13B18C34. This indicates a high likelihood of the presence
of Ce4+ in Ce33Fe13B18C34. To confirm this, XPS data were
collected on crystals of Ce33Fe13B18C34. After sputtering to
remove surface oxide species, the spectrum features two broad
peaks at binding energies of 900 and 918 eV (see Figure S2 in
Supporting Information), with associated shake-down peaks at
895 and 913 eV. These correspond to a 3d5/2−3d3/2 spin−orbit
doublet which arises exclusively from Ce4+; the 918 eV peak in
particular does not overlap with any possible transitions
characteristic of Ce3+ and is viewed as strongly indicative of
tetravalent cerium.41,42
The formation of tetravalent cerium in this compound may
be induced by chemical pressure. The R33Fe13B18C34 structure
is stable for the smaller rare earth cation Pr3+; the packing of
the Fe13 clusters and associated borocarbide chains promotes
the conversion of Ce3+ cations to smaller Ce4+ ions to stabilize
the compound. Similar effects are seen for systems such as
CexY1−xAl3, where substitution of smaller Y3+ into the rare earth
sites causes the average surrounding coordination environment
to shrink and induces formation of Ce4+ ions.43
Transport Properties of Ce33Fe14−xAlx+yB25−yC34. Resis-
tivity measurements performed on a single crystal of
Ce33Fe14−xAlx+yB25−yC34 indicate that this material behaves as
a poor metal (see Figure 4). The room temperature resistivity is
5.39 Ω cm; this lies in the overlap of typical metallic (10−5−101
Ω cm) and semiconductor (10−2−105 Ω cm) resistivity
ranges.44 The resistivity rises as the temperature is increased,
as is expected for metals. Two kinks in the temperature
dependence are observed, one at low temperature (10 K) which
corresponds to a magnetic ordering transition, and one at 125
K which may be due to cerium fluctuating valence; both these
features are mirrored in the magnetic susceptibility data (vide
infra).
R33Fe13−xAlxB18C34 crystallizes in the centrosymmetric space
group Im3m, with unit cell parameters of a = 14.246(8) Å for
̅
the Ce analog and a = 14.3881(9) Å for the Pr analog. The Fe13
iron cluster in this structure is perfectly cuboctahedral and is
centered by another iron atom instead of a boron atom (see
Figure 3). The Fe−Fe bond lengths between neighboring iron
atoms and those to the center iron are all equal; this distance of
2.579(2) Å for the cerium analog is slightly longer than the
distances seen for Ce33Fe14−xAlx+yB25−yC34 and La21Fe8Sn7C12,
but is well within the ranges seen in Ce2Fe17 and α-Fe (2.482−
2.866 Å).5,30 A very small amount of aluminum substitution
(4%) is seen on the central Fe2 site for Pr33Fe12.9Al0.1B18C34,
but not for the cerium analogue. All 12 of the Fe1 atoms
defining the cuboctahedron are capped by a C−B−C chain.
The Fe1−C1 bond is 1.947(8) Å in length; subsequent bond
lengths in the chain are C1−B1 of 1.50(1) Å and B1−C2 of
1.50(1) Å, comparing well to those in other structures. The six
square faces of the cuboctahedron are each capped by a B−C
unit with a Fe1−B2 distance of 2.236(8) Å, very similar to the
distances between the Fe14B cluster to the capping BCBC
chains in Ce33Fe14−xAlx+yB25−yC34. The B2−C3 distance of
1.58(2) Å is somewhat longer than expected, but the distances
from the C3 site to surrounding rare earth ions support the
assignment of this site as carbon and not boron. All terminal C
atoms are bonded to Ce atoms at distances of 2.43(1)−
2.717(1) Å, forming a framework that surrounds the central Fe
cluster. As is seen in the R33Fe14−xAlx+yB25−yC34 structure,
R33Fe13−xAlxB18C34 also features isolated C4− anions (C4 sites)
octahedrally coordinated by R cations with six equivalent C−
Ce distances of 2.630(1) Å.
It is notable that Ce33Fe13B18C34 has a smaller unit cell than
its praseodymium analog. All bond lengths associated with rare
earth sites are shorter in the Ce phase compared to the Pr
analog (see Table 3), contrary to what is expected given the
relative sizes of Ce3+ and Pr3+. The Ce−X bonds are also
NMR studies were carried out on Ce33Fe14−xAlx+yB25−yC34 in
an attempt to glean information about boron and carbon siting
12143
dx.doi.org/10.1021/ja303370j | J. Am. Chem. Soc. 2012, 134, 12138−12148

Journal of the American Chemical Society
Article
Figure 4. Resistivity data for Ce33Fe14−xAlx+yB25−yC34, measured on a
single crystal.
as well as the interaction of these nuclei with conduction
electrons. While no 13C resonances could be observed (likely
due to low natural abundance and broadening from
surrounding paramagnetic atoms), the 11B MAS NMR
spectrum shows a single peak at 11.2 ppm (Figure S3,
Supporting Information). This is likely due to the B2 site,
which is bonded to two carbon atoms; the resonances of the
two other boron sites (both bonded directly to iron) may be
too broadened to be observed. This 11B shift is similar to those
reported for YB4, LaB4 and CaCxB4−x, all of which have
resonances in the 5−56 ppm range.45,46 In these phases, the
small shift (relative to the larger Knight shifts expected for
metals) is attributed to a small boron s-electron density of
states at the Fermi level. Polar intermetallic phases often feature
a pseudogap in the DOS at Ef,47 which can lead to poor
conductivity and a small Knight shift. The NMR data and
resistivity measurements suggest that Ce33Fe14−xAlx+yB25−yC34
may exhibit such a pseudogap, leading to poor metal behavior.
Magnetic Behavior of R33Fe14−xAlx+yB25−yC34 Phases.
The magnetic properties of R33Fe14−xAlx+yB25−yC34 and
R33Fe13−xAlxB18C34 phases are of particular interest, given the
presence of two potentially paramagnetic species: the iron
clusters and the rare earth cations. The magnetic susceptibility
of La33Fe14−xAlx+yB25−yC34 is small and roughly temperature
independent (see Figure 5), indicating Pauli paramagnetic
behavior and therefore no magnetic moment on the iron atoms.
The slight Curie tail at low temperature is likely due to traces of
oxidation or paramagnetic impurities on the surface of the
crystal sample. The magnitude of the Pauli paramagnetism
(0.06 emu/mol; taking the stoichiometry into account, this
corresponds to roughly 1 × 10−3 emu/mol of metal atoms) is
consistent with that seen for other metallic compounds.48 The
cerium analog Ce33Fe14−xAlx+yB25−yC34 was studied using
susceptibility measurements, XPS measurements to investigate
cerium valence, and Mossbauer studies on an 57-Fe enriched
̈
sample to explore possible magnetic transitions of iron sites.
The temperature dependence of the susceptibility (Figure 5)
shows paramagnetic behavior above 200 K which can be fit to a
modified Curie−Weiss law (χ = (C/T − θ) + χp, where C is the
Curie constant, θ is the Weiss temperature, and χp corresponds
to the Pauli paramagnetism of the conduction electrons). The
resulting parameters, listed in Table 4, indicate an effective
magnetic moment per cerium ion of 2.36 μB, which is only
Figure 3. (a) Structure of Ce33Fe13B18C34. Red spheres are iron, green
spheres are cerium, black spheres are carbon, blue spheres are boron.
The cuboctahedral Fe13 clusters are drawn as red polyhedra and the
isolated carbon anion in octahedral coordination as gray polyhedra.
(b) Fe13 cluster and associated bond lengths. (c) Ce1 coordination
environment (48k site). (d) Ce2 coordination environment (12d site).
(e) Ce3 coordination environment (6b site). (f) Coordination
environment around the borocarbide chain.
12144
dx.doi.org/10.1021/ja303370j | J. Am. Chem. Soc. 2012, 134, 12138−12148

Journal of the American Chemical Society
Article
Figure 5. Temperature dependence of magnetic susceptibility for
R33Fe14−xAlx+yB25−yC34 phases, with an applied field of 100 Oe. Data
for La analog are in black; data for Ce analog (filled circles χm and
open circles 1/χm) are in red.
Figure 6. 57Fe Mossbauer spectra at various temperatures for 57-Fe
enriched sample of Ce33Fe14−xAlx+yB25−yC34.
̈
Table 4. Magnetic Data for R/Fe/B/C Phases
moment
per R3+ ion
(μB)
χP
(emu/mol
of FU)
observed
transition
compound
θ (K)
−8.4
La33Fe14B25C34
Ce33Fe14B25C34
Ce33Fe13B18C34
Pr33Fe13B18C34
0.06
0.06
Pauli paramagnetic
AF; TN = 9 K
2.36
non-Curie−Weiss
−44
FM; TC = 190 K
AF; TN = 16 K
4.15
0.06
slightly lower than the theoretical value of 2.54 μB for Ce3+
ions.48 The negative Weiss constant of θ = −8.4 K indicates the
presence of antiferromagnetic coupling forces, in agreement
with the observed antiferromagnetic ordering at TN = 9 K. This
ordering is likely due to the cerium ions; as in the La analog,
the iron atoms in Ce33Fe14−xAlx+yB25−yC34 do not appear to
have a magnetic moment.
Figure 7. Magnetization data for Ce33Fe14−xAlx+yB25−yC34 at several
temperatures.
The lack of magnetic moment on the iron atoms in
R33Fe14−xAlx+yB25−yC34 phases is somewhat surprising consid-
ering the magnetic behavior exhibited by iron atoms in the
smaller Fe4 clusters of La21Fe8Sn7C12.5 The quenching of the
iron magnetism may be due to the Fe14B clusters of
R33Fe14−xAlx+yB25−yC34 being “diluted” by the central boron
atom and by the aluminum substitution on the Fe(2) site.
Similar dilution effects are observed for R2Fe14B phases and
their substituted analogs. In the Nd2Fe14B structure, the iron
sites with the largest number of iron nearest neighbors have the
largest magnetic moments, and the moments of all the iron
atoms are lowered when R2Fe14−xTxB analogs are made with
linear paramagnetic behavior. Therefore, this transition at
around 120 K does not appear to be a magnetic ordering
phenomenon and is not associated with the iron sites. The fact
that this feature is not seen in the data for the lanthanum analog
indicates that it may be due to fluctuating valence of the cerium
ions. The Curie−Weiss fit of the susceptibility temperature
dependence supports the presence of Ce3+ at high temperature.
This is also supported by room temperature XPS data on
Ce33Fe14−xAlx+yB25−yC34 (Figure S4, Supporting Information)
which show the Ce 3d5/2 and 3d3/2 peaks and their satellites in
positions typical for Ce3+; the spectrum is very similar to those
seen for CePO4 and CeFeAsO.42,50 However, as the temper-
ature is lowered below 150 K, thermal contraction of the lattice
may become sufficient to promote Ce3+/Ce4+ valence
fluctuation. One or more of the three cerium sites may remain
trivalent; these remaining paramagnetic ions order antiferro-
magnetically at 9 K. Similar behavior is seen in the mixed valent
phase Ce23Ru7Cd4, where 20 out of the 46 cerium ions in the
cell remain trivalent and their magnetic moments order at 3.6
iron diluted by a nonmagnetic element T.16 The La(AlxFe1−x 13
)
family of compounds exhibits a similar drop in the iron
magnetic moment as the amount of Al substitution increases.49
The other notable feature in the susceptibility data of
Ce33Fe14−xAlx+yB25−yC34 is a broad bump from 75 to 150 K.
This is reproducibly seen in the magnetic data for several
different samples of this compound. It also corresponds to the
kink in the resistivity temperature dependence shown in Figure
4. Mossbauer spectra were collected at several temperatures
̈
spanning this transition to determine if it is caused by the iron
sites. The data taken at 78, 150, and 294 K are shown in Figure
6; the spectra consist of a broad peak with a shoulder (in
agreement with the presence of two iron sites in the structure).
No hyperfine splitting or variation in shift is observed between
78 and 294 K. Also, field dependence magnetization data taken
at various temperatures in this range (Figure 7) show normal
K.10 The 57Fe Mossbauer spectra of Ce Fe14−xAlx+yB25−yC34
̈
33
(Figure 6) show very little change below the antiferromagnetic
transition at 9 K; instead of hyperfine splitting, only a slight
broadening is observed in the 5.5 K spectrum which indicates
weak coupling of the iron with the surrounding Ce3+ ions as
they undergo their ordering transition.
12145
dx.doi.org/10.1021/ja303370j | J. Am. Chem. Soc. 2012, 134, 12138−12148

Journal of the American Chemical Society
Article
Magnetic Behavior of R33Fe13−xAlxB18C34 Phases. The
magnetic susceptibility data for Ce33Fe13B18C34 are shown in
Figure 8. The dominant feature is a distinct rise in susceptibility
Figure 10. Temperature dependence of magnetic susceptibility for
Pr33Fe13−xAlxB18C34, with an applied field of 100 Oe. Filled black
circles χm and open blue circles 1/χm.
Figure 8. Temperature dependence of magnetic susceptibility for
Ce33Fe13B18C34 at 100 Oe applied field.
praseodymium ion, significantly higher than the 3.58 μB
expected for Pr3+.48 The higher moment indicates that the
iron atoms in this phase exhibit magnetic moments, as is seen
for Ce33Fe13B18C34. However, no ferromagnetic ordering is
observed that would correspond to the transition at 180 K seen
for the cerium compound. This lack of ordering may be due to
the small amount of aluminum substitution consistently
observed in the EDS and crystallographic data for this analog.
This, and the slightly longer Fe−Fe bonds in the iron clusters
(2.598 Å in Pr33Fe13−xAlxB18C34 vs 2.579 Å in the cerium
analog), may hinder ordering of the magnetic moments. The
low temperature antiferromagnetic ordering observed at 16 K is
likely due to the moments on the Pr3+ ions; antiferromagnetic
coupling is supported by the negative Weiss constant of −44 K.
at 180 K and subsequent field-cooled/zero-field-cooled splitting
at lower temperatures. Since the room temperature XPS data
(Figure S2, Supporting Information) indicate that this phase
contains diamagnetic Ce4+ ions, this rise in susceptibility is
likely due to magnetic ordering of the iron clusters. Field
dependence data (Figure 9) taken above and below the
CONCLUSION
■
The two Ce/Fe/B/C intermetallic compounds grown in Ce/T
flux have similar structures, but the cerium and iron atoms in
these phases exhibit very different behaviors. The Fe14 clusters
in Ce33Fe14−xAlx+yB25−yC34 are nonmagnetic due to dilution of
Fe−Fe bonding by boron and aluminum; the iron atoms in the
Fe13 clusters in Ce33Fe13B18C34 do exhibit magnetic moments
and ordering. These two borocarbide-capped iron clusters offer
a “missing link” between isolated iron sites in intermetallics and
more extended iron building blocks such as iron layers and
networks featuring extensive Fe−Fe bonding. Ce33Fe13B18C34 is
an ordered array of identical magnetic iron nanoclusters; it is
not clear if these relatively isolated Fe13 clusters are interacting.
Further studies on the magnetism of this phase are planned,
although the synthesis yield must be improved for 57Fe
Figure 9. Field dependence of magnetization for Ce33Fe13B18C34 at
several different temperatures.
transition at 150 K show an increase in magnetization below
this temperature, confirming that the rise in susceptibility is
magnetic in origin. The nature of the magnetic ordering is not
clear; the magnetization is not saturated, which may indicate
that the magnetic coupling and ferri- or ferromagnetic ordering
occurs within individual Fe13 clusters, but not between them
(each cluster is 12.3 Å away from eight neighboring clusters).
The data above the ordering temperature cannot be fit to the
Curie−Weiss law, so determination of the iron moment is not
possible. The fact that the iron atoms in this structure have
magnetic moments while the Fe1 4 clusters in
Ce33Fe14−xAlx+yB25−yC34 do not is likely due to the fact that
the Fe13 clusters are not diluted by a central boron atom or
small amounts of Al substitution on the iron sites.
enrichment (for Mossbauer studies) or 11B enrichment to allow
̈
neutron diffraction studies.
Both Ce/Fe/B/C structures feature three crystallographically
unique cerium sites. All of the cerium ions in
Ce33Fe14−xAlx+yB25−yC34 are trivalent at high temperature, but
some convert to Ce4+ below 100 K. Ce33Fe13B18C34 appears to
feature predominantly tetravalent cerium ions. Both these
behaviors can be explained by the presence of different forms of
chemical pressure. The normal thermal contraction of bonds
occurring as Ce33Fe14−xAlx+yB25−yC34 cools is evidently
sufficient to trigger conversion of some cerium sites from
Ce3+ to Ce4+. The squeezing of cerium ions into the
R33Fe13B18C34 structure (which is more stable for smaller Pr3+
cations) stabilizes the tetravalent state. Experiments to
The magnetic susceptibility data for the Pr33Fe13−xAlxB18C34
analog, shown in Figure 10, can be fit to the Curie−Weiss law
above 100 K. This yields a magnetic moment of 4.15 μB per
12146
dx.doi.org/10.1021/ja303370j | J. Am. Chem. Soc. 2012, 134, 12138−12148

Journal of the American Chemical Society
Article
substitute smaller R3+ cations into these structures to observe
the effects are underway.
(16) Szytula, A.; Leciejewicz, J. CRC Handbook of Crystal Structures
and Magnetic Properties of Rare Earth Intermetallics; CRC Press: Boca
Raton, FL, 1994.
(17) Massalski, T. B.; Okamoto, H. Binary Alloy Phase Diagrams, 2nd
ed.; ASM International: Materials Park, OH, 1990.
(18) SAINT, version 6.02a; Bruker AXS Inc.: Madison, WI, 2000.
(19) Sheldrick, G. M. SHELXTL NT/2000, version 6.1; Bruker AXS,
Inc.: Madison, WI, 2000.
(20) Dashjav, E.; Kreiner, G.; Schnelle, W.; Wagner, F. R.; Kniep, R.;
Jeitschko, W. J. Solid State Chem. 2007, 180, 636−653.
(21) Davaasuren, B.; Borrmann, H.; Dashjav, E.; Kreiner, G.; Widom,
M.; Schnelle, W.; Wagner, F. R.; Kniep, R. Angew. Chem., Int. Ed. 2010,
49, 5688−5692.
(22) Benbow, E. M.; Dalal, N. S.; Latturner, S. E. J. Solid State Chem.
2009, 182, 3055−3062.
(23) Villars, P.; Calvert, L. D. Pearson’s HandbookCrystallographic
Data for Intermetallic Phases; ASM International: Materials Park, OH,
1998.
ASSOCIATED CONTENT
■
S
* Supporting Information
Table of atomic positions and displacement factors for R/Fe/
B/C compounds. Powder diffraction data, XPS data for cerium
phases, 11B MAS NMR spectrum for Ce33Fe14−xAlx+yB25−yC34.
Additional crystallographic data for all phases (CIF). This
material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
latturne@chem.fsu.edu
(24) Peter, S. C.; Malliakis, C. D.; Nakotte, H.; Kothapilli, K.;
Rayaprol, S.; Schultz, A. J.; Kanatzidis, M. G. J. Solid State Chem. 2012,
187, 200−207.
Notes
The authors declare no competing financial interest.
(25) Salvador, J. R.; Bilc, D.; Mahanti, S. D.; Kanatzidis, M. G. Angew.
Chem., Int. Ed. 2003, 42, 1929−1932.
ACKNOWLEDGMENTS
■
(26) Wagner, C. D.; Naumkin, A. V.; Kraut-Vass, A.; Allison, J. W.;
Powell, C. J.; Rumble, J. R. NIST X-ray Photoelectron Spectroscopy
Database, Version 3.5; National Institute of Standards and
Technology: Gaithersburg, 2003; http://srdata.nist.gov/xps/.
(27) Gerss, M. H.; Jeitschko, W.; Boonk, L.; Nientiedt, J.; Grobe, J.;
Morsen, E.; Leson, A. J. Solid State Chem. 1987, 70, 19−28.
(28) Shoemaker, C. B.; Shoemaker, D. P.; Fruchart, R. Acta Cryst. C
1984, 40, 1665−1668.
This work made use of the NMR facilities of the Department of
Chemistry and Biochemistry, the SEM facilities of the
Department of Physics, and facilities at the National High
Magnetic Field Laboratory. Financial Support from the NSF
(grants DMR-05-47791 and DMR-11-06150) is gratefully
acknowledged. We thank Dr. Eric Lochner for his assistance
in obtaining the XPS data. Mossbauer spectra were taken at the
NHMFL which is funded by the NSF through Cooperative
Agreement No. DMR-06-54118, the State of Florida, and the
DOE. The Mossbauer instrument was purchased using the
User Collaboration Grant Program 5064 funds awarded to A.O.
̈
(29) Anderson, S.; Carlsson, J. O. Acta Chem. Scandin. 1970, 24,
1791−1799.
̈
(30) Teplykh, P. A.; Pirogov, A. N.; Kuchin, A. G.; Teplykh, A. E.
Phys. B 2004, 350, E99−E102.
(31) Shoemaker, C. B.; Shoemaker, D. P.; Hopkins, T. E.; Yindepit,
S. Acta Cryst. B 1978, 34, 3573−3576.
REFERENCES
■
(32) Babizhetskyy, V.; Hiebl, K.; Kremer, R. K.; Mattausch, H.;
Simon, A. Solid State Sci. 2007, 9, 1126−1134.
(1) Kanatzidis, M. G.; Pottgen, R.; Jeitschko, W. Angew. Chem., Int.
̈
Ed. 2005, 44, 6996−7023.
(33) Bidaud, E.; Hiebl, K.; Hoffmann, R. D.; Pottgen, R.; Jardin, C.;
̈
(2) Stojanovic, M.; Latturner, S. E. J. Solid State Chem. 2007, 180,
907−914.
Bauer, J.; Gautier, R.; Gougeon, P.; Saillard, J. Y.; Halet, J. F. J. Solid
State Chem. 2000, 154, 286−295.
(3) Reynolds, P. C.; Stojanovic, M.; Latturner, S. E. J. Solid State
Chem. 2011, 184, 1875−1881.
(34) Babizhetskyy, V.; Hiebl, K.; Mattausch, H.; Simon, A. J. Phys.
Chem. Solids 2009, 70, 561−566.
(4) Lang, D. A.; Zaikina, J. V.; Lovingood, D. D.; Gedris, T. E.;
Latturner, S. E. J. Am. Chem. Soc. 2010, 132, 17523−17530.
(5) (a) Benbow, E. M.; Dalal, N.; Latturner, S. E. J. Am. Chem. Soc.
2009, 131, 3349−3354. (b) Zaikina, J. V.; Schellenberg, I.; Benbow, E.
(35) Babizhetskyy, V.; Hiebl, K.; Mattausch, H.; Simon, A. Solid State
Sci. 2009, 11, 501−506.
(36) Babizhetskyy, V.; Mattausch, H.; Simon, A.; Hiebl, K.; Yahia, M.
B.; Gautier, R.; Halet, J. F. J. Solid State Chem. 2008, 181, 1882−1890.
(37) vanDuijn, J.; Suzuki, K.; Attfield, J. P. Angew. Chem. 2000, 39,
365−366.
M.; Pottgen, R.; Latturner, S. E. Chem. Mater. 2011, 23, 1768−1778.
̈
(6) Zaikina, J. V.; Jo, Y. J.; Latturner, S. E. Inorg. Chem. 2010, 49,
2773−2781.
(38) Jardin, C.; Oeckler, O.; Mattausch, H. J.; Simon, A.; Halet, J. F.;
Sallard, J. Y.; Bauer, J. Z. Anorg. Allg. Chem. 2001, 627, 1389−1394.
(39) Jardin, C.; Oeckler, O.; Mattausch, H. J.; Simon, A.; Halet, J. F.;
Sallard, J. Y.; Bauer, J. Inorg. Chem. 2000, 39, 5895−5900.
(40) Lassoued, S.; Gautier, R.; Boutarfaia, A.; Halet, J. F. J.
Organomet. Chem. 2010, 695, 987−993.
(7) Zaikina, J. V.; Zhou, H.; Latturner, S. E. J. Solid State Chem. 2010,
183, 2987−2994.
(8) Kirchmayr, H. R.; Poldy, C. A. J. Magn. Magn. Mater. 1978, 8, 1−
42.
(9) Reehuis, M.; Brown, P. J.; Jeitschko, W.; Moller, M. H.; Vomhof,
T. J. Phys. Chem. Solids 1993, 54, 469−475.
(10) Tappe, F.; Hermes, W.; Eul, M.; Pottgen, R. Intermetallics 2009,
(41) Renaudin, G.; Dieudonne, B.; Avignant, D.; Mapemba, E.; El-
Ghozzi, M.; Fleutot, S.; Martinez, H.; Cerny, R.; Dubois, M. Inorg.
Chem. 2010, 49, 686−694.
̈
17, 1035−1040.
(11) Mishra, R.; Hermes, W.; Rodewald, U. C.; Hoffmann, R. D.;
(42) Beche, E.; Charvin, P.; Perarnau, D.; Abanades, S.; Flamant, G.
Surf. Interface Analys. 2008, 40, 264−267.
Pottgen, R. Z. Anorg. Allg. Chem. 2008, 634, 470−474.
̈
(12) Hegger, H.; Petrovic, C.; Moshopoulou, E. G.; Hundley, M. F.;
Sarrao, J. L.; Fisk, Z.; Thompson, J. D. Phys. Rev. Lett. 2000, 84, 4986−
4989.
(43) Goremychkin, E. A.; Osborn, R.; Sashin, I. L.; Riseborough, P.;
Rainford, B. D.; Adroja, D. T.; Lawrence, J. M. Phys. Rev. Lett. 2010,
104, 176402.
(13) Witte, A. M.; Jeitschko, W. Z. Naturforsch. B 1996, 51, 249−256.
(14) Sieve, B.; Gray, D. L.; Henning, R.; Bakas, T.; Schultz, A. J.;
Kanatzidis, M. G. Chem. Mater. 2008, 20, 6107−6115.
(15) McCallum, R. W.; Fisher, I. R.; Anderson, N. E.; Canfield, P. C.;
Kramer, M. J.; Dennis, K. W. IEEE Trans. Magn. 2001, 37, 2147−2149.
(44) West, A. R. Basic Solid State Chemistry, 2nd ed.; John Wiley &
Sons, Ltd.: Chichester, West Sussex, England, 1999.
(45) Schmitt, R.; Blaschkowski, B.; Eichele, K.; Meyer, H. J. Inorg.
Chem. 2006, 45, 3067−3073.
12147
dx.doi.org/10.1021/ja303370j | J. Am. Chem. Soc. 2012, 134, 12138−12148

Journal of the American Chemical Society
Article
(46) Zogal, O. J.; Florian, P.; Massiot, D.; Paluch, S.; Shitsevalova, N.;
Borshchevsky, D. F. Solid State Commun. 2009, 149, 693−696.
(47) (a) Han, M. K.; Morosan, E.; Canfield, P. C.; Miller, G. J. Z.
Kristallogr. 2005, 220, 95−101. (b) Gout, D.; Benbow, E.; Gourdon,
O.; Miller, G. J. J. Solid State Chem. 2003, 176, 538−548.
(48) Kittel, C. Introduction to Solid State Physics, 6th ed.; John Wiley
& Sons, Inc.: New York, 1986.
(49) Palstra, T. T. M.; Nieuwenhuys, G. J.; Mydosh, J. A.; Buschow,
K. H. J. Phys. Rev. B 1985, 31, 4622−4632.
(50) Blanchard, P. E. R.; Cavell, R. G.; Mar, A. J. Solid State Chem.
2010, 183, 1477−1483.
12148
dx.doi.org/10.1021/ja303370j | J. Am. Chem. Soc. 2012, 134, 12138−12148