Magnetization study of the ultra-hard material MgAlB14

Article abstract of DOI:10.1016/S0304-8853(03)00220-8

Magnetic susceptibility χ versus temperature T, magnetization M versus T, and isothermal M versus magnetic field H studies of the ultra-hard material MgAlB₁₄ were carried out in search of superconductivity or ferromagnetism in this compound. Two types of samples were synthesized: (1) powder and (2) chemically substituted and unsubstituted hot pressed pellets prepared from mechanically alloyed powders. χ(T) measurements on a powder sample revealed temperature-independent diamagnetism with a Curie-Weiss impurity concentration equivalent to ～ 1mol% of spin-12 ions. In contrast, M(T) and M(H) data on the hot pressed samples, both substituted and unsubstituted, showed evidence of ferromagnetic transitions above ～ 330K. Scanning electron microscopy and Auger microprobe analysis of the hot pressed samples indicated that both substituted and unsubstituted samples contained significant concentrations of Fe impurities. We conclude that pure MgAlB ₁₄ is neither a superconductor nor a ferromagnet above 1.8K and exhibits temperature-independent diamagnetism from 1.8K up to room temperature. The ferromagnetism observed in the hot pressed samples is likely due to Fe impurities abraded from the stainless steel mills used to mix the starting materials prior to hot pressing the samples.

Full text of DOI:10.1016/S0304-8853(03)00220-8

ARTICLE IN PRESS
Journal of Magnetism and Magnetic Materials 265 (2003) 23–32
Magnetization study of the ultra-hard material MgAlB14
J.M. Hilla,*, D.C. Johnstona, B.A. Cookb, J.L. Harringab, A.M. Russellb
a Ames Laboratory and Department of Physics and Astronomy, Iowa State University, Ames, IA 50011, USA
b Ames Laboratory, Iowa State University, Ames, IA 50011, USA
Received 5 December 2002
Abstract
Magnetic susceptibility w versus temperature T; magnetization M versus T; and isothermal M versus magnetic field
H studies of the ultra-hard material MgAlB14 were carried out in search of superconductivity or ferromagnetism in this
compound. Two types of samples were synthesized: (1) powder and (2) chemically substituted and unsubstituted hot
pressed pellets prepared from mechanically alloyed powders. wðTÞ measurements on a powder sample revealed
temperature-independent diamagnetism with a Curie–Weiss impurity concentration equivalent to B1mol % of spin- 21
ions. In contrast, MðTÞ and MðHÞ data on the hot pressed samples, both substituted and unsubstituted, showed
evidence of ferromagnetic transitions above B330 K: Scanning electron microscopy and Auger microprobe analysis of
the hot pressed samples indicated that both substituted and unsubstituted samples contained significant concentrations
of Fe impurities. We conclude that pure MgAlB14 is neither a superconductor nor a ferromagnet above 1.8 K and
exhibits temperature-independent diamagnetism from 1.8 K up to room temperature. The ferromagnetism observed in
the hot pressed samples is likely due to Fe impurities abraded from the stainless steel mills used to mix the starting
materials prior to hot pressing the samples.
r 2003 Elsevier Science B.V. All rights reserved.
PACS: 75.20.Àg; 75.30.Cr; 81.05.Je; 81.40.Rs
Keywords: Ultra-hard boride MgAlB14; Magnetization measurements; Ferromagnetic impurities
1. Introduction
superconductor. More recently, experiments have
shown that high frequency phonons are likely
involved in the electron–phonon pairing interac-
tion [4,5]. Materials with high phonon frequencies
tend to have high Debye temperatures yD and
more rigid elastic constants [6]. Masui et al. (Ref.
[4]) have recently estimated yD to be B400 K in
MgB2: High rigidity is linked to the hardness of a
material [7]. This suggests that other hard, non-
transition metal compounds may also exhibit
superconductivity.
The recent discovery by Akimitsu and co-
workers of superconductivity at 39 K in MgB2 [1]
indicated that the condensed matter physics
community may have overlooked interesting
physical properties in other known non-transition
metal compounds. Initial studies of MgB2 [2]
provided strong evidence that it was a BCS [3]
*Corresponding author. Tel.: +515-294-1364; fax: +515-
294-0689.
E-mail address: jmhill@ameslab.gov (J.M. Hill).
Young et al. (Ref. [8]) discovered recently
that single crystals of Ca1ÀxLaxB6 exhibited
0304-8853/03/$ - see front matter r 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0304-8853(03)00220-8

ARTICLE IN PRESS
24
J.M. Hill et al. / Journal of Magnetism and Magnetic Materials 265 (2003) 23–32
ferromagnetism with Curie temperatures up to
B600 K and very small saturation moments. This
discovery generated much theoretical and experi-
mental work on doped and undoped CaB6 and
SrB6: There are no transition metal atoms in either
of these compounds, suggesting that the observed
ferromagnetism arises from an as yet unknown
mechanism. This discovery suggests that other
boride compounds which do not contain transition
metal atoms may also exhibit this new ‘‘exotic’’
ferromagnetism. However, recent results [9,10]
provide strong evidence that the ferromagnetism
observed in Ca1ÀxLaxB6 is due to ferromagnetic
impurities present on the surface of the crystals.
Chemically modified MgAlB14 was discovered
by Ames Laboratory scientists a few years ago
to have a hardness ranging from 32–46 GPa,
making it one of the hardest materials known
[11]. For comparison, diamond, with a hardness
of 70–100 GPa (Vickers Hardness Number=
8400 kg=mm2 ¼ 82 GPa [12]) and a yD of 1860 K
[13], is the hardest bulk material known. MgAlB14
has an orthorhombic crystal structure containing
B12 icosahedra arranged in distorted, close-packed
layers linked together by the metal atoms and by
single B atoms [14–16]. The partially (B0:8)
occupied Al (Mg) sites form linear (zig-zag) chains
parallel to the a-axis. The interatomic distances
within the chains and between Mg and Al chains
are comparable to the interatomic distances in Mg
and Al metal [15,17]. Single crystal refinements of
MAlB14 (M=Tb, Dy, Ho, Er, Tm, Yb, Lu, Y, Li)
yielded similar results [17–19,16]. The compound
Mg2B14 was found by Guette et al. (Ref. [20]) to
have the same orthorhombic structure but with a
B/Mg ratio much closer to 7.
temperature. Mo.ssbauer spectra for Fe-doped
polycrystalline MgAlB14; Fe(MgAlB14Þ30; indi-
cated that the Fe exists in two oxidation states,
Feþ2 and Feþ3; in this material [23]. The tempera-
ture dependencies of the Seebeck coefficient and
electrical conductivity for Fe(MgAlB14Þ30 were
unchanged from those of the Fe-free materials
[23]. Infrared spectroscopy measurements on a
single crystal of MgAlB14 were also consistent with
non-metallic behavior [25]. Measurements on less
disordered MgAlB14 samples which contain no
metal atom vacancies would help to clarify the
conduction mechanism in this compound.
Magnetization measurements ranging from
5–280 K on polycrystalline M1ÀxAl1ÀyB14 (M =
Tb, Dy, Ho, Er) samples indicated local moment
magnetism consistent with the magnetism of
trivalent rare earth ions [17]. The magnetic
susceptibility followed the Curie–Weiss law, w ¼
C=ðT À yÞ; where the Curie constant is Nm2eff =3kB;
N is the number of magnetic atoms, meff is the
effective magnetic moment per rare earth atom,
and kB is Boltzmann’s constant. The effective
moments meff obtained ranged from 11 to 12 mB;
slightly larger than calculated values, and the
obtained Weiss temperatures were yE À 15 K for
M=Dy, Ho, and Er, and 10 K for M=Tb.
Measurements of the temperature dependence of
the electrical resistance of a single crystal of
Er1ÀxAl1ÀyB14 were consistent with variable-range
hopping [17]. This is in agreement with results
discussed above which indicate that the electrical
properties of MAlB14 are indicative of a non-
metallic, disordered system.
In a search for superconductivity and/or ferro-
magnetism in the ultra-hard material MgAlB14; we
have measured the magnetization of hot pressed
materials as well as a polycrystalline sample
prepared using conventional solid state synthesis.
The data for the powder sample are presented
directly following the Experimental Details section
below, followed by the data on hot pressed samples.
A few physical property studies have been
reported in the literature for this boride family.
Three studies reported that the temperature
dependence of the Seebeck coefficient and elec-
trical conductivity of polycrystalline MgAlB14
showed a simultaneous increase with temperature
above 300 K, indicative of hopping conduction in
a non-metallic, disordered system [21–23]. A recent
study on polycrystalline samples by Takeda et al.
(Ref. [24]) reported that the Seebeck coefficient
decreased with increasing temperature and that the
electrical conductivity increased with increasing
2. Experimental details
Powder samples of MgAlB14 were synthesized
by reacting 99.999% pure (metals basis) distilled

ARTICLE IN PRESS
J.M. Hill et al. / Journal of Magnetism and Magnetic Materials 265 (2003) 23–32
25
Mg metal (Ames Laboratory), 99.999% Al
spheres, and 99.99% amorphous B powder (Alfa
Aesar) in sealed Ta tubes at 1300ꢀC for approxi-
mately three weeks. Fast neutron activation
analysis (Elemental Analysis, Inc.) revealed that
the B powder, which was dried under vacuum at
600ꢀC and subsequently handled and stored in a
high-purity He atmosphere, was contaminated
with 6.0(3) wt% of oxygen. Powder X-ray diffrac-
tion (XRD) patterns were taken on a Rigaku X-
ray diffractometer with CuKa radiation. Fig. 1
shows the XRD pattern of a sample for which the
starting materials were briefly exposed to air
before sealing in the Ta tube. Lattice parameters
were obtained from a different XRD pattern using
Si powder as an internal standard. The XRD
pattern showed that the sample consists of
MgAlB14; together with a significant amount of
spinel, MgAl2O4; as the major impurity phase
arising from the O impurities in the B powder
starting material and possibly from exposing the
reactants to air. The MgAlB14 diffraction peaks
(ꢁ) were indexed in the orthorhombic space group
Imam (No. 74) with lattice parameters a ¼
technique such that the reactants were not exposed
to air before they were sealed into the Ta tube
resulted in samples which had larger amounts of
AlB12 impurities as shown in Fig. 2. We could not
identify all of the impurity phases present in the
samples. Changing the Mg:Al:B stoichiometry of
the starting materials from 1:1:14 to 1:1:10 or
0.8:0.8:14 did not affect the results. No single-
phase samples were obtained using the synthesis
method employed here.
Vekshina et al. (Ref. [26]) successfully removed
impurity phases from their samples of powder
MgAlB14 by dissolving their boron-poor impurity
phases and/or unreacted metals with boiling
aqueous HCl. Therefore, we treated some samples
with boiling 4 M HCl. However, the weight loss
due to the acid treatment was negligible. In
addition, the XRD patterns of a sample taken
before the acid treatment and after the acid
treatment were essentially identical, as shown in
Fig. 3, with only minor changes in the intensity of
some peaks. Treating the samples with NH4OH
also did not affect the phase composition of the
samples. We conclude that the above chemical
treatments of our samples had no significant effect
on their phase composition.
(
(
(
5:866ð2Þ A, b ¼ 8:114ð5Þ A, and c ¼ 10:312ð6Þ A,
similar to literature values [15]. The impurity
MgAl2O4 (2075 wt%) reflections are also indi-
cated in the figure (3). Altering the synthesis
Hot pressed samples of MgAlB14 and
MgAlB14:X (X=Si, TiB2) were prepared by
mechanical alloying with hardened chrome steel
1400
MgAlB14
1200
1000
MgAl2O4
MgAlB14
MgAl2O4
α-AlB12
1000
800
600
400
200
0
800
600
400
200
0
x
y
AlB12
β-AlB12
x
x
y
x
x
x
x
x
y
x
x
x
x
x
x
x
10
20
30
40
50
60
70
80
2θ (o)
0
10
20
30
2θ (o)
40
50
60
Fig. 1. X-ray powder diffraction pattern of MgAlB14 (CuKa
radiation). The solid curve is intensity I versus diffraction angle
2y: The MgAlB14 diffraction lines are indicated by ꢁ and are
indexed in the orthorhombic space group Imam (No. 74) with
Fig. 2. X-ray powder diffraction pattern of MgAlB14 (CuKa
radiation). The solid curve is intensity I versus diffraction angle
2y: The line positions for MgAlB14 (ꢁ) and MgAl2O4 (3) are
indicated. Reflections for a-AlB12 ðꢂÞ; AlB12 (&), and b-AlB12
(y) are also marked.
(
(
(
a ¼ 5:866ð2Þ A, b ¼ 8:114ð5Þ A, and c ¼ 10:312ð6Þ A. The Miller
indices of the strongest reflections are shown. Impurity
MgAl2O4 reflections are indicated by 3:

ARTICLE IN PRESS
26
J.M. Hill et al. / Journal of Magnetism and Magnetic Materials 265 (2003) 23–32
1500
1000
500
0
Before Acid
MgAlB14
MgAl2O4
After Acid
10
20
30
40
50
60
70
80
2θ (o)
Fig. 3. X-ray powder diffraction pattern of MgAlB14 showing
intensity I versus diffraction angle 2y (CuKa radiation). The
top curve was taken before treatment of the sample with boiling
4 M HCl and the bottom curve was taken after the HCl
treatment. MgAlB14 (MgAl2O4) line positions are indicated by
ꢁ (3). For clarity, the intensity of the bottom curve was shifted
downwards by 1000 counts/s.
Fig. 4. Scanning electron microscope image of the microstruc-
ture of hot pressed MgAlB14 containing Fe, O, and C impurities
(BSE mode). The brightest areas correspond to Fe enrichment.
Energy dispersive spectroscopy spectra corresponding to two
points (labeled ‘‘1’’ and ‘‘2’’) on the micrograph are shown in
Fig. 5.
ball mills (Spex Industries) and consolidated by
vacuum hot pressing at 1300–1500ꢀC in BN-lined
graphite dies as described in Ref. [27]. Some
samples were examined for elemental content by
two field emission electron microprobe instru-
ments: a scanning electron microscope (SEM)
(Amray) and a scanning Auger microprobe (Jeol
JAMP-7830F). Fig. 4 shows a backscattered SEM
picture of an unsubstituted hot pressed MgAlB14
sample, which is clearly inhomogeneous. Energy
dispersive spectroscopy (EDS) analyses of spot
‘‘1’’ and spot ‘‘2’’ are shown in Fig. 5. For
chemical analysis, the grain size of these materials
is less than the minimum electron beam spot size
(B1mm). Therefore various multi-grain regions of
overall homogeneous chemical composition over
the 1mm spot area (as determined from z-contrast
imaging) were analyzed. It is clear that the
elemental content, particularly Fe, varies greatly
with position in this material. The bright regions in
Fig. 4 (spot 1) have a high Fe content, whereas the
light gray areas (spot 2) have a much lower Fe
content. EDS spectra from the dark gray areas
gave identical results as from spot 2. Fe is
evidently introduced into the material from milling
wear debris. Light elements are not detected as
efficiently by EDS as are the heavier elements.
Therefore, B, C, and O peak heights are not a
reliable quantitative indication of their relative
concentrations in this material.
In addition to the above analyses, a hot pressed
sample prepared from powder precursor materials
and nominally substituted with 0.63 at% Si for Al
was etched with an Arþ ion beam before analysis
in the Auger microprobe. Iron-rich precipitates are
seen as bright phases against the dark matrix
(MgAlB14 material) because of the high atomic
number contrast, and were distributed across the
surface of the specimen (not shown). Analysis of
the Auger electron spectra obtained from several
matrix regions gave an Al:Fe ratio of 46.7:
9.54 at% (sensitivity to the B signal is poor and
results are calculated with respect to all heavier
elements). Since Al loss during the processing
has been found to be negligible, this ratio can be
directly scaled to the nominal Al concen-
tration, with the result that the average Fe
content within the matrix phase is 1.09 at%,
corresponding to the approximate composition
(MgAlB14ÞFe0:17
:
Two additional samples were examined in
the SEM. The first was another 0.63 at%
Si-substituted composition, similar to the one

ARTICLE IN PRESS
J.M. Hill et al. / Journal of Magnetism and Magnetic Materials 265 (2003) 23–32
27
Fig. 5. Energy dispersive spectroscopy (EDS) spectra from the locations labeled 1(a) and 2 (b) in Fig. 4. Fe and B contents vary with
position in this material.
discussed above. Semi-quantitative EDS analysis
of three matrix regions gave Fe concentrations of
0.88, 2.25, and 2.3 at%, for an average of
1.81 at%. Three regions within the second sample,
an unsubstituted hot pressed MgAlB14 sample,
gave results of 1.65, 1.62, and 1.91 at% Fe,
for an average of 1.73 at%. It is expected that
the majority of the Fe would be found as a
second phase along MgAlB14 grain bounda-
ries. Alternatively, these values place an upper
limit on the amount of Fe substitutionally
incorporated within the MgAlB14 phase to be
about 1.8 at%.
superconducting quantum interference device
(SQUID) magnetometer.
3. Magnetization of polycrystalline powder
MgAlB14
The magnetization M versus magnetic field H
isotherm at T ¼ 300 K for our powder sample of
MgAlB14 (see Fig. 1) is shown in Fig. 6. The
saturation moment Ms was obtained by fitting the
high field data with the fit function
MðH; TÞ ¼ MsðTÞ þ wðTÞH;
ð1Þ
Magnetic susceptibility w versus temperature T
and isothermal magnetization M versus magnetic
field H measurements from 1.8 to 400 K were
carried out using a Quantum Design MPMS
where w(T) is the magnetic susceptibility, yielding
Ms ¼ 4:05ð5Þ ꢂ 10À3 G cm3/g, which is equivalent
to the contribution of 18.2 wt ppm of Fe metal

ARTICLE IN PRESS
28
J.M. Hill et al. / Journal of Magnetism and Magnetic Materials 265 (2003) 23–32
0.005
0.000
5
MgAlB14
MgAlB14
4
H = 10 kOe
T = 300 K
-0.005
-0.010
-0.015
-0.020
-0.025
3
Fit: 10 - 250 K
χ = χ0 + Cimp/(T − θ)
2
χ0 = −5.8(1)x10−7 cm3/g
Fit: 1 - 5.5 T
M = Ms + χH
C
imp = 1.6(1)x10−5 cm3 K/g
1
0
M
s = 4.05(5)x10−3 G cm3/g
χ = −5.49(1)x10−7 cm3/g
θ = −2.5(1) K
-0.030
-1
0
1
2
3
4
5
6
0
50
100 150 200 250 300 350 400
T (K)
H (T)
Fig. 6. Magnetization M versus magnetic field H of a MgAlB14
powder sample at 300 K (}). The solid curve is a fit to the high
field data at HX1T by the function shown in the figure
[Eq. (1)], with parameters also listed in the figure.
Fig. 7. Magnetic susceptibility w versus temperature T of a
MgAlB14 powder sample which contained approximately
2075 wt% MgAl2O4 (}). The solid curve is a fit to the data
by the function shown in the figure [Eq. (3)], with parameters
also listed in the figure.
impurities, and w ¼ À5:49ð1Þ ꢂ 10À7 cm3/g. Simi-
lar data taken at 50, 100, and 200 K showed that
Ms increased only slightly to Ms ¼ 4:31ð9Þꢂ
10À3 G cm3/g, allowing a fit to MsðTÞ to be made.
Then we extracted the susceptibility wðTÞ;
is the sum of the contribution from the diamag-
netic cores of the atoms, wcore; and the paramag-
netic Van Vleck susceptibility, wVV; which is
expected to be small in this non-transition metal
compound. The fit yielded y ¼ À2:5ð1Þ K;
Cimp ¼ 1:6ð1Þ ꢂ 10À5 cm3 K=g; and w0 ¼ À5:8ð1Þꢂ
10À7 cm3/g.
MðH; TÞ À MsðTÞ
wðTÞ ¼
;
ð2Þ
H
from MðTÞ at constant H ¼ 1 T: MðH; TÞ at
H ¼ 1T is in the linear regime of the MðHÞ plots,
as shown in Fig. 6, which is necessary for the
derived wðTÞ to be valid.
To analyze the contribution of wcore to w0; we
assumed that the sample was made up of 20 wt%
MgAl2O4 and 80 wt% MgAlB14: Diamagnetic core
susceptibilities wcore tabulated in Ref. [28] were
used for Mgþ2 (À3 ꢂ 10À6 cm3/mol), Alþ3
(À2 ꢂ 10À6 cm3/mol), and OÀ2 (À12 ꢂ 10À6 cm3/
mol). Further analysis depends strongly on the
assumed wcore of B. Using the Pascal constant for B
tabulated in Ref. [29], À7:2 ꢂ 10À6 cm3/mol B,
yielded a calculated w0 value for MgAlB14 most
consistent with the value obtained from the fit. The
value of wcore for 1mol of MgAl 2O4 is
The magnetic susceptibility wðTÞ in an applied
magnetic field of H ¼ 10 kOe for a powder
MgAlB14 sample which contained 2075 wt%
MgAl2O4; described above, is shown in Fig. 7.
The ferromagnetic impurity contribution MsðTÞ to
MðH; TÞ; from MðHÞ isotherms taken from 50 to
300 K as described above, is corrected for in Fig. 7.
The compound exhibits temperature-independent
diamagnetism at high temperature with a small
Curie–Weiss paramagnetic impurity contribution.
We fitted the data by
w
coreðMgAl2O4Þ ¼ À 5:5 ꢂ 10À5 cm3=mol
¼ À 3:9 ꢂ 10À7cm3=g:
ð5Þ
ð6Þ
Cimp
w ¼ w0 þ
;
ð3Þ
T À y
The value of wcore for 1mol of MgAlB 14 is
where y is the Weiss temperature and Cimp is
the impurity Curie constant. The T-independent
w0 term,
w
coreðMgAlB14Þ ¼ À 1:06 ꢂ 10À4 cm3=mol
¼ À 5:2 ꢂ 10À7cm3=g:
w0 ¼ wcore þ wVV
ð4Þ

ARTICLE IN PRESS
J.M. Hill et al. / Journal of Magnetism and Magnetic Materials 265 (2003) 23–32
29
The calculated value of wcore for our sample is
core
0.30
Hot Pressed MgAlB14
w
¼ 0:2ðÀ3:9 ꢂ 10À7 cm3=gÞ
0.25
0.20
0.15
0.10
0.05
0.00
calc:
T = 300 K
þ 0:8ðÀ5:2 ꢂ 10À7 cm3=gÞ
¼ À 7:8 ꢂ 10À8 cm3=g þ À4:2 ꢂ 10À7 cm3=g
¼ À 4:9 ꢂ 10À7 cm3=g;
ð7Þ
Fit: 4 - 5.5 T
which is slightly more positive than the value
obtained from the fit, À5:8ð1Þ ꢂ 10À7 cm3/g. The
value of wVV from Eq. (4) is thus small as expected:
M = Ms + χH
Ms = 0.269(1) µB/F.U.
χ = 0.0046(1) µB/[T F.U.]
VVD8:6 ꢂ 10À8 cm3/g of sample.
0
1
2
3
4
5
6
w
H (T)
Fig. 9. Magnetization M versus magnetic field H at 300 K of
hot pressed MgAlB14 (}). The solid curve is a fit to the high
field data by the function shown in the figure [Eq. (1)], with
parameters also listed in the figure.
4. Magnetization of hot pressed MgAlB14
The magnetization M(T) in an applied magnetic
field of H ¼ 10 kOe for a typical hot pressed
MgAlB14 sample is shown in Fig. 8. The data are
qualitatively different from the data for powder
MgAlB14 shown in Fig. 7 and are indicative
of ferromagnetism with a saturation moment
MsC0:4mB/FU. The Curie temperature Tc was
estimated to be B460 K by drawing a line through
the region of maximum slope as shown in the
figure. A magnetization M versus magnetic field H
isotherm at T ¼ 300 K is shown in Fig. 9. The
magnetization nearly saturates at fields above 1T.
The saturation moment Ms was obtained by fitting
the 4–5.5 T data with Eq. (1), yielding Ms ¼
0:269ð1ÞmB/FU, equivalent to the contribution
of B12 mol% Fe metal impurities, and w ¼
0:0046ð1ÞmB/[T FU]. For comparison, the drasti-
cally different MðHÞ isotherm at T ¼ 300 K
for a powder sample of MgAlB14 was shown in
Fig. 6.
Since the magnetization shows a significant
saturation moment, isothermal M(H) hysteresis
loop data were obtained. A hysteresis loop taken
at T ¼ 300 K for the hot pressed material is shown
in Fig. 10. The remnant magnetization and coer-
cive field are small, 0:0485mB/FU and B75 Oe;
respectively.
The MðTÞ in an applied magnetic field of H ¼
10 kOe for a hot pressed sample of nominal
composition MgAl0:95Si0:05B14 is shown in
Fig. 11. The data are very similar to those for the
unsubstituted hot pressed sample and indicate a
ferromagnetic transition. The Curie temperature
as estimated from the dashed line shown in the
figure is TcB335 K: An MðHÞ isotherm at T ¼
300 K is shown in Fig. 12. A fit to the 4–5.5 T data
by Eq. (1) yielded Ms ¼ 0:086ð1ÞmB/FU, equivalent
to the contribution of B4 mol% Fe metal impu-
rities, and w ¼ 0:00297ð8Þ mB/[T FU].
0.40
Hot Pressed MgAlB14
0.35
H = 10 kOe
0.30
0.25
0.20
0.15
M vs H curves
0.10
Temperature scan
0.05
Tc 460 K
0.00
0
100
200
300
400
500
T (K)
An M(H) isotherm at T ¼ 300 K for a hot
pressed sample with nominal composition
(MgAlB14ÞðTiB2Þ1:3 is shown in Fig. 13. The data
for this TiB2 doped sample are very similar to
the other samples, suggesting a significant con-
tribution to the magnetization arises from
Fig. 8. Magnetization M versus temperature T of hot pressed
MgAlB14: The solid circles (ꢁ) were taken by scanning the
temperature in a fixed magnetic field of H ¼ 10 kOe: The open
circles (3) are 10 kOe data from M(H) curves taken at fixed
temperature. A dashed line drawn through the region of
maximum slope was used to estimate TcB 460 K; as shown.

ARTICLE IN PRESS
30
J.M. Hill et al. / Journal of Magnetism and Magnetic Materials 265 (2003) 23–32
0.3
Hot pressed MgAlB14
T = 300 K
0.2
0.1
0.0
-0.1
-0.2
-0.3
-600
-400
-200
0
200
400
600
H (Oe)
Fig. 10. An MðHÞ hysteresis loop at 300 K of hot pressed MgAlB14 (}) in the low-field regime. The inset shows the entire loop with a
maximum applied magnetic field of 5.5 T.
0.12
0.10
0.08
0.06
0.04
0.02
0.16
Hot Pressed MgAl0.95Si0.05B14
T = 300 K
Hot Pressed MgAl0.95Si0.05B14
0.14
0.12
0.10
0.08
Fit: 4 - 5.5 T
M = Ms + χH
s = 0.086(1) µB/F.U.
χ = 0.00297(8) µB/[T F.U.]
0.06
0.04 H = 10 kOe
M
Data from M vs H curves
0.02
0.00
Tc 335 K
0.00
0
1
2
3
4
5
6
0
50
100 150 200 250 300 350 400
T (K)
H (T)
Fig. 12. Magnetization M versus magnetic field H at 300 K of a
hot pressed sample of nominal composition MgAl0:95Si0:05B14
(}). The solid curve is a fit to the high field data by the
function shown in the figure [Eq. (1)], with parameters also
listed in the figure.
Fig. 11. Magnetization M versus temperature T of a hot
pressed sample of nominal composition MgAl0:95Si0:05B14 (}).
The 10 kOe data were taken from isothermal MðHÞ curves. A
dashed line drawn through the region of maximum slope was
used to estimate TcB335 K; as shown.
ferromagnetic impurities. A fit to the 4–5.5 T data
by Eq. (1) yielded Ms ¼ 0:2709ð4ÞmB/FU, equiva-
lent to the contribution of B12 mol% Fe metal
impurities, and w ¼ 0:00177ð3ÞmB/[T FU].
amounts of both MgAl2O4 and AlB12 were formed
in various samples as secondary phases according
to XRD analysis. The magnetic susceptibility w of
our highest purity powder sample showed tem-
perature-independent diamagnetism plus a para-
magnetic Curie–Weiss contribution equivalent to
that of approximately 1mol% of spin- 12 impurities.
The nominally pure hot pressed MgAlB14
samples and samples containing additional Si or
TiB2 were found to contain varying amounts of
Fe impurities. Based on transmission electron
5. Summary and conclusions
Single phase powder samples of MgAlB14 were
not obtained by the conventional solid state
synthesis method employed here. Significant

ARTICLE IN PRESS
J.M. Hill et al. / Journal of Magnetism and Magnetic Materials 265 (2003) 23–32
31
0.30
0.25
0.20
0.15
0.10
0.05
0.00
under Contract No. W-7405-Eng-82. This work
was supported by the Director for Energy Re-
search, Office of Basic Energy Sciences.
(MgAlB14)(TiB2)1.3
T = 300 K
References
Fit: 4 - 5.5 T
M = Ms + χH
[1] J. Nagamatsu, N. Nakagawa, T. Muranaka, Y. Zenitani,
J. Akimitsu, Nature 410 (2001) 63.
Ms = 0.2709(2) µB/F.U.
χ = 0.00177(3) µB/[T F.U.]
[2] C. Buzea, T. Yamashita, Supercond. Sci. Technol. 14
(2001) R115 and cited references.
0
1
2
3
4
5
6
[3] J. Bardeen, L.N. Cooper, J.R. Schrieffer, Phys. Rev. 108
(5) (1957) 1175.
H (T)
[4] T. Masui, K. Yoshida, S. Lee, A. Yamamoto, S. Tajima,
Phys. Rev. B 65 (2002) 214513.
Fig. 13. Magnetization M versus magnetic field H at 300 K of a
hot pressed sample of nominal composition (MgAlB14
)
[5] H. Kotegawa, K. Ishida, Y. Kitaoka, T. Muranaka,
N. Nakagawa, H. Takigawa, J. Akimitsu, Phys. Rev. B
66 (2002) 064516.
(TiB2)1:3 (}). The solid curve is a fit to the high field data by
the function shown in the figure [Eq. (1)], with parameters also
listed in the figure.
[6] N.W. Ashcroft, N.D. Mermin, Solid State Physics, W.B.
Saunders Company, Philadelphia, PA, 1976.
[7] C. Kittel, Introduction to Solid State Physics, 7th Edition,
Wiley, New York, 1996.
microscopy images of samples substituted with
TiB2; selected area diffraction with resolution of
order 200 nm was performed, yielding distinct
TiB2 areas separate from MgAlB14 areas [30].
We, therefore, conclude that TiB2 did not act as a
traditional dopant, but rather primarily collected
in boride phases in the grain boundaries. Based on
the results of the SEM and Auger microprobe
analyses, we conclude that the Fe impurity content
is sufficient to account for the observed ferromag-
netism in the magnetization data.
Lee and Harmon (Ref. [31]) have calculated the
elastic constants of MgAlB14: The measured
hardness for the hot pressed samples (Ref. [11])
is larger than they calculated. Therefore, the Fe
and other dopants acting in conjunction with an
ultrafine microstructure may increase the hardness
of this compound. It is therefore of interest to
synthesize single phase MgAlB14 so that the
intrinsic hardness and elastic constants can be
measured. Magnetization measurements on a
single-phase sample would also facilitate a better
determination of the intrinsic magnetic behavior.
[8] D.P. Young, D. Hall, M.E. Torelli, Z. Fisk, J.L. Sarrao,
J.D. Thompson, H.-R. Ott, S.B. Oseroff, R.G. Goodrich,
R. Zysler, Nature 397 (1999) 412.
[9] S. Otani, T. Mori, J. Phys. Soc. Jpn. 71 (7) (2002) 1791.
[10] K. Taniguchi, T. Katsufuji, F. Sakai, H. Ueda,
K. Kitazawa, H. Takagi, Phys. Rev. B 66 (2002) 064407.
[11] B.A. Cook, J.L. Harringa, T.L. Lewis, A.M. Russell, Scr.
Mater. 42 (2000) 597.
[12] J.C. Anderson, K.D. Leaver, R.D. Rawlings,
J.M. Alexander, Materials Science, 4th Edition, Chapman
& Hall, London, 1990, p. 212.
[13] J. de Launay, The theory of specific heats and lattice
vibrations, in: F. Seitz, D. Turnbull (Eds.), Solid State
Physics, Vol. 2, Academic Press Inc., New York, 1956,
p. 233.
[14] V.I. Matkovich, J. Economy, Acta Crystallogr. Sect. B 26
(1970) 616.
[15] I. Higashi, T. Ito, J. Less-Common Met. 92 (1983) 239.
[16] I. Higashi, M. Kobayashi, S. Okada, K. Hamano,
T. Lundstro.m, J. Cryst. Growth 128 (1993) 1113.
[17] M.M. Korsukova, V.N. Gurin, Y.B. Kuzma,
N.F. Chaban, S.I. Chykhrii, V.V. Moshchalkov,
N.B. Brandt, A.A. Gippius, K. Khyu Nyan, Phys. Stat.
Sol. A 114 (1989) 265.
[18] M.M. Korsukova, V.N. Gurin, Y. Yu, L.-E. Tergenius,
T. Lundstro.m, J. Alloys Compounds 190 (1993) 185.
[19] M.M. Korsukova, T. Lundstro.m, L.-E. Tergenius,
V.N. Gurin, J. Alloys Compounds 187 (1992) 39.
[20] A. Guette, M. Barret, R. Naslain, P. Hagenmuller,
J. Less-Common Met. 82 (1981) 325.
Acknowledgements
[21] I.A. Bairamashvili, L.I. Kekelidze, O.A. Golikova,
V.M. Orlov, J. Less-Common Met. 67 (1979) 461.
[22] O.A. Golikova, S. Samatov, Phys. Stat. Sol. A 77 (1983)
449.
Ames Laboratory is operated for the US
Department of Energy by Iowa State University

ARTICLE IN PRESS
32
J.M. Hill et al. / Journal of Magnetism and Magnetic Materials 265 (2003) 23–32
[23] M.M. Kazanin, V.V. Kutasov, Phys. Stat. Sol. A 113
(1989) 143.
[28] P.W. Selwood, Magnetochemistry, 2nd Edition, Inter-
science, New York, 1956, p. 78.
[24] M. Takeda, F. Domingo, T. Miura, T. Fukuda, Mater.
Res. Soc. Symp. Proc. 691(2002) 209.
[29] E. Ko.nig, G. Ko.nig, Magnetic properties of coordination
and organometallic transition metal compounds, in:
K.-H. Hellwege, A.M. Hellwege (Eds.), Landolt-Bo.rnstein:
Numerical Data and Functional Relationships in Science
and Technology, Vol. 8, Springer, Berlin, 1976.
[30] T.L. Lewis, A.M. Russell, B.A. Cook, J.L. Harringa,
Mater. Sci. Eng. A, in press.
[25] H. Werheit, U. Kuhlmann, G. Krach, I. Higashi,
T. Lundstro.m, Y. Yu, J. Alloys Compounds 202 (1993) 269.
[26] N.V. Vekshina, L.Y. Markovskii, Y.D. Kondrashev, T.K.
Voevodskaya, Zh. Prikl. Khim. 44 (5) (1971) 958;
J. Appl. Chem. USSR 44 (1971) 970.
[27] A.M. Russell, B.A. Cook, J.L. Harringa, T.L. Lewis, Scr.
Mater. 46 (2002) 629.
[31] Y. Lee, B.N. Harmon, J. Alloys Compounds 338 (2002)
242.