Large specific absorption rates in the magnetic hyperthermia properties of metallic iron nanocubes

Article abstract of DOI:10.1016/j.jmmm.2010.05.012

We report on the magnetic hyperthermia properties of chemically synthesized ferromagnetic 11 and 16 nm Fe(0) nanoparticles of cubic shape displaying the saturation magnetization of bulk iron. The specific absorption rate measured on 16 nm nanocubes is 1690±160 W/g at 300 kHz and 66 mT. This corresponds to specific losses-per-cycle of 5.6 mJ/g, largely exceeding the ones reported in other systems. A way to quantify the degree of optimization of any system with respect to hyperthermia applications is proposed. Applied here, this method shows that our nanoparticles are not fully optimized, probably due to the strong influence of magnetic interactions on their magnetic response. Once protected from oxidation and further optimized, such nano-objects could constitute efficient magnetic cores for biomedical applications requiring very large heating power.

Full text of DOI:10.1016/j.jmmm.2010.05.012

ARTICLE IN PRESS
Journal of Magnetism and Magnetic Materials 322 (2010) L49–L52
Contents lists available at ScienceDirect
Journal of Magnetism and Magnetic Materials
journal homepage: www.elsevier.com/locate/jmmm
Letter to the Editor
Large specific absorption rates in the magnetic hyperthermia properties
of metallic iron nanocubes
B. Mehdaoui a,b, A. Meffre a,b, L.-M. Lacroix a,b, J. Carrey a,b,n, S. Lachaize a,b, M. Gougeon c,
M. Respaud a,b, B. Chaudret d
a Universite´ de Toulouse, INSA, UPS, LPCNO (Laboratoire de Physique et Chimie des Nano-Objets), 135 avenue de Rangueil, F-31077 Toulouse, France
b CNRS, UMR 5215, LPCNO, F-31077 Toulouse, France
c
ˆ
Institut CARNOT-CIRIMAT-UMR 5085, Batiment 2R1, 118 route de Narbonne, F-31062 Toulouse, France
d Laboratoire de Chimie de Coordination-CNRS, 205 rte de Narbonne, 31077 Toulouse cedex 4, France
a r t i c l e i n f o
a b s t r a c t
Article history:
We report on the magnetic hyperthermia properties of chemically synthesized ferromagnetic 11 and
16 nm Fe(0) nanoparticles of cubic shape displaying the saturation magnetization of bulk iron. The
specific absorption rate measured on 16 nm nanocubes is 16907160 W/g at 300 kHz and 66 mT. This
corresponds to specific losses-per-cycle of 5.6 mJ/g, largely exceeding the ones reported in other
systems. A way to quantify the degree of optimization of any system with respect to hyperthermia
applications is proposed. Applied here, this method shows that our nanoparticles are not fully
optimized, probably due to the strong influence of magnetic interactions on their magnetic response.
Once protected from oxidation and further optimized, such nano-objects could constitute efficient
magnetic cores for biomedical applications requiring very large heating power.
Received 7 January 2010
Received in revised form
3 May 2010
Available online 8 May 2010
Keywords:
Hyperthermia
Magnetic nanoparticle
Iron
Organometallic synthesis
& 2010 Elsevier B.V. All rights reserved.
1. Introduction
orientations of easy axis, to magnetic interactions and to thermal
activation, hysteresis loops of MNPs assemblies are far from being
square. Thus, more generally, one can write
Magnetic hyperthermia is a way to improve the efficiency of
chemotherapy or radiotherapy by raising the temperature of a
tumour to 41–45 1C during a few hours using magnetic nano-
SAR ¼ 4am0Happ
sSfexc
ð1Þ
´
particles (MNPs). In the clinical treatment set at the Charite
where is a dimensionless parameter, which characterizes the
relative area of the hysteresis loops with respect to the ideal
square. As a consequence of Eq. (1) and of the limitation of the
a
Hospital, Berlin, MNPs are first localised inside the tumour and
then excited by an alternating magnetic field of moderate
amplitude
m0Happ (12–25 mT) at a frequency fexc of 100 kHz [1].
Happ exc
displaying a high sS and hysteresis loops as square as possible
(large ).
f
product, the maximisation of the SAR requires MNPs
The power released by the NPs is assessed by their specific
absorption rate (SAR) or their specific losses-per-cycle (SLPC)
noted A, linked by the equation SAR¼Afexc. Increase in SARs above
1 kW/g could be beneficial for several aspects of the hyperthermia
applications but it implies a real challenge [2]. Indeed, due to
physiological issues, human body cannot be exposed to alternat-
a
High magnetization materials are essentially metallic Ni, Fe, Co
and their alloys. Among them, iron is the most promising one
since it displays a very large sS and is a priori the most
biocompatible. However, iron oxides have been so far the most
widely studied materials for magnetic hyperthermia, due to
their full biocompatibility and the relative simplicity of their
synthesis and handling. Optimized chemically synthesized Fe3O4
NPs have shown SLPC up to 1.5 mJ/g at 400 kHz [3]. Among high
magnetization materials, higher SLPC values up to 3.2 mJ/g at
400 kHz have been reported for Co MNPs [4]. The scarce results
published so far on Fe MNPs are disappointing as a consequence
of the lack of control of the surface oxidation and/or of a non-
optimal nanoparticle size [5–7]. The best results published so far
were performed on Fe/FexOy core–shell MNPs, which displayed
ing magnetic field of large Happ exc
optimization of the treatment must rely on the MNPs only. The
most favourable case, leading to SARmax ¼ 4m0Happ Sfexc occurs
when (i) the magnetization loop, characterized by a coercive field
HC and a saturation magnetization per unit mass S, is a perfect
f
product [2]. Therefore, the
s
s
square and (ii) HCEHapp. Unfortunately, due to the random
n Corresponding author at: Universite´ de Toulouse, INSA, UPS, LPCNO
(Laboratoire de Physique et Chimie des Nano-Objets), 135 avenue de Rangueil,
F-31077 Toulouse, France.
E-mail address: julian.carrey@insa-toulouse.fr (J. Carrey).
low sS and SLPC comparable to those measured on iron oxides [5].
0304-8853/$ - see front matter & 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmmm.2010.05.012

ARTICLE IN PRESS
L50
B. Mehdaoui et al. / Journal of Magnetism and Magnetic Materials 322 (2010) L49–L52
Our group has developed an organometallic approach allowing
2. Synthesis and characterzations
the controlled synthesis of pure metallic iron NPs displaying the
bulk magnetization [8–10]. We report in this article on the
magnetic and hyperthermia measurements of ferromagnetic 11
and 16 nm nanocubes exhibiting very large SLPC. Their efficiency
for the required application and the influence of magnetic
interactions on this efficiency is discussed.
The two samples of Fe nanocubes were synthesized by a two
steps organometallic route, which has been previously reported
[8]. A complete description of the size and shape control on iron
NPs synthesized by this route will be published elsewhere.
First, we prepared small Fe particles (ꢀ2 nm) through the
Fig. 1. TEM micrographs of the iron nanocubes measured in hyperthermia. (a) High-resolution TEM micrograph on an isolated nanocube in sample 1. (b) Sample 1:
nanocubes with a mean diameter of 16.3 nm. (c) Sample 2: nanocubes with a mean diameter of 11 nm. Insets: nanocube size distribution extracted from the analysis of
several TEM micrographs.

ARTICLE IN PRESS
B. Mehdaoui et al. / Journal of Magnetism and Magnetic Materials 322 (2010) L49–L52
L51
ampoule containing the colloidal solution was sealed under
vacuum to prevent any oxidation of the NPs. A typical ampoule
contains 9 mg of powder and 550 mg of mesitylene. The ampoule
was then placed into a calorimeter with 1.5 mL of deionised
water, the temperature of which was measured. The measure-
ment time was adjusted between 20 and 200 s depending on the
experiment, so that the temperature rise never exceeds 20 1C.
After the magnetic field stops, the water is shaken during roughly
20 s to ensure the ampoule thermalization and the homogeneity
of the water temperature, which is checked by putting two probes
at the top and the bottom of the calorimeter. The temperature rise
is measured after this process. For measurements longer than
20 s, the raw SAR values were corrected from the calorimeter
losses, which were previously calibrated. The SAR values at 66 mT
and their error bars were obtained by averaging three measure-
ments of 20 s on three different ampoules arising from the same
MNP synthesis (9 values). The complete magnetic-field depen-
dence of the SAR was measured on a single ampoule for each
synthesis and its SAR value renormalized accordingly.
Fig. 2. SQUID measurement at 300 K on sample 1 (16 nm) and sample 2 (11 nm).
The inset shows an enlarged view of the magnetization normalized by its value
at 5 T.
During hyperthermia experiments, the spatial organisation of
the MNPs in the ampoule was completely redistributed by the
application of the magnetic field. Indeed, while the MNPs
aggregates fall down by gravity at the bottom of the ampoule in
the absence of any applied magnetic field, the MNPs formed small
spikes along the field direction when a small magnetic field was
applied. Moreover, for larger magnetic fields, the MNPs self-
organized into regularly spaced levitating needles, which disin-
tegrated as soon as the magnetic field was stopped. This
phenomenon is shown in Fig. 3(a) for sample 2. The magnetic
field for which these needles formed ranged between 20 and 30
mT; it was higher for sample 1 than for sample 2. The formation of
columns in a magnetic field is a classical behaviour of ferrofluids
and is due to magnetic interactions between the MNPs [12].
Fig. 3(b) displays the SAR values at 300 kHz as a function of the
applied magnetic field for the two samples. In both cases, SAR
increases strongly in the range 10–30 mT and then follows a
roughly linear increase at higher magnetic field without complete
saturation. For sample 1, the increase is sharper and occurs at a
higher magnetic field than for sample 2. Such an abrupt increase
followed by a plateau is a typical feature of ferromagnetic samples
and was previously reported on ferromagnetic FeCo MNPs [13].
decomposition of the iron dimer {Fe[N(SiMe3)2]2}2, (376 mg,
0.5 mmol) in mesitylene (20 mL) at 150 1C under 3 bars of H2 in
a Fisher–Porter bottle (170 mL) [11]. In a second step, a mixture of
hexadecylammonium chloride (HDA ꢁ HCl) and hexadecylamine
(HDA) was added to the colloidal solution. For sample 1, the ratio
ammonium/amine was 1:2 (277 mg, 1 mmol HDA ꢁ HCl; 483 mg,
2 mmol HDA). For sample 2, this ratio was 1.1:2 (304.7 mg,
1.1 mmol HDA ꢁ HCl; 483 mg, 2 mmol HDA). The solution was
stirred for 20 min at 90 1C then pressurised under 3 bars of H2 and
heated at 150 1C. After 48 h, a black precipitate was formed at the
bottom of the Fisher–Porter bottle. The solvent was filtered off
and the precipitate was washed three times with 15 mL of toluene
to remove the surfactants in excess and other remaining
molecular species. The final iron content was 71% and 88% for
samples 1 and 2, respectively, as determined by chemical analysis.
Samples for transmission electron microscopy (TEM) were
prepared by the deposition of a drop of solution onto a carbon-
coated copper grid. High-resolution TEM revealed that the
nanocubes were single-crystalline and exhibited a bcc crystal
structure, with facets of (1 0 0) planes (see Fig. 1(a)). As revealed
by TEM micrographs, most of the nanocubes were embedded in
organic mesophases, which prevent them to form a true colloidal
solution (see Figs. 1(b) and 1(c)). Similar nanocube-filled
The sharp rise of the SAR occurs when
m0Happ reaches the coercive
field of the MNPs. SARs up to 16907160 and 13207140 W/g are
measured at 66 mT for samples 1 and 2, respectively. For sample
1, this value corresponds to SLPC of 5.670.5 mJ/g, which exceeds
mesophases have already been characterized for
a similar
system in which carboxylic acid was used instead of ammonium
chloride [9]. Size distributions measured from several TEM
micrographs showed that samples 1 and 2 were composed of
iron nanocubes of mean side lengths 16.371.5 nm (sample 1)
and 11.371.3 nm (sample 2), respectively, (see insets of Fig. 1(b)
and (c)).
Samples for magnetic measurements were prepared and sealed
under an argon atmosphere in order to preserve the metallic
character of Fe. SQUID measurements on powders of samples 1
and 2 are shown in Fig. 2. Their saturation magnetizations per unit
mass
s
S were 200710 and 17879 Am2/kg at 300 K, respectively,
just below the bulk value. Their coercive fields HC at 300 K are 16
and 5 mT, respectively.
3. Hyperthermia measurements
Fig. 3. (a) An hyperthermia experiment on sample 2 at a magnetic field of 66 mT
and 300 kHz. The needles formed by the nanoparticles under the influence of the
magnetic field are visible. (b) Magnetic field dependence of SAR at 300 kHz for the
two samples.
Hyperthermia experiments were performed on an induction
oven working at a frequency of 300 kHz and a maximum
magnetic field of 66 mT. For hyperthermia measurements, an

ARTICLE IN PRESS
L52
B. Mehdaoui et al. / Journal of Magnetism and Magnetic Materials 322 (2010) L49–L52
by a factor 3 the values reported on optimized chemically
synthesized iron oxide NPs [3].
since they should be protected from oxidation and their toxicity
must be tested. To protect the metallic core from oxidation,
a silica layer can be used, as recently patented by our group [17].
These non-trivial adjustments and post-treatments will constitute
the future developments of this work.
4. Discussion on efficiency
The efficiency of our system is however limited for two
reasons. First, the calculation of
with the experimental SAR value, and the experimental sS value
a at m0Happ¼66 mT using Eq. (1)
Acknowledgements
at 300 K leads to
respectively. Above
a¼0.11 and a¼0.09 for samples 1 and 2,
´
We acknowledge InNaBioSante foundation, AO3 program from
m
0Happ¼32 mT, values are approximately
a
´
´
Universite Paul Sabatier (Toulouse) and Conseil Regional de Midi-
Pyre´ne´es for financial support, V. Colliere (TEMSCAN) for HRTEM,
C. Crouzet for technical assistance and A. Mari for magnetic
measurements.
constant with the magnetic field but decreases below. These
values are rather low since a value up to
¼0.3 can be deduced
from the measurements on randomly oriented optimized iron
oxide NPs [3]. In our case, this low value is compensated by a
larger S. We argue that the presence of magnetic interactions is
the main reason for these low values. Indeed, in the case of
ꢀ
a
a
s
a
References
the Stoner–Wohlfarth (SW) model, valid for randomly oriented
magnetically independent nanoparticles, aSW¼0.25 [14]. When
the nanoparticles interact magnetically, it is well admitted that
both HC and MR decrease [15], thus reducing
clearly in the presence of system dominated by strong
[1] P. Wust, U. Gneveckow, M. Johannsen, D. Bo¨hmer, T. Henkel, F. Kahmann,
J. Sehouli, R. Felix, J. Ricke, A. Jordan, Int. J. Hyperthermia 22 (2006) 673.
[2] R. Hergt, S. Dutz, J. Magn. Magn. Mater. 311 (2007) 187.
[3] R. Hergt, R. Hiergeist, I. Hilger, W.A. Kaiser, Y. Lapatnikov, S. Margel,
U. Richter, J. Magn. Magn. Mater. 270 (2004) 345.
a below aSW. We are
a
interactions since (i) the nanoparticles form dense needles when
the magnetic field is on, (ii) the MR/MS ratio (MR is the remnant
magnetization and MS the saturation magnetization) in SQUID
measurements is 0.2 and 0.06 for samples 1 and 2, respectively,
whereas 0.5 is expected in the absence of interactions, (iii) the
saturation field is well above 2HC. The second limitation of our
system is that we measured huge SLPCs at a magnetic field of
66 mT while, so far, only magnetic fields close to 20 mT were
applied during medical hyperthermia treatments [16]. A fine
tuning of their size and/or of their dispersion must be considered
to decrease their coercive field so that high SLPC values could be
obtained at reduced magnetic field.
[4] M. Zeisberger, S. Dutz, R. Mu¨ller, R. Hergt, N. Matoussevitch, H. Bo¨nneman,
J. Magn. Magn. Mater. 311 (2005) 224.
[5] I. Baker, Q. Zeng, W. Li, C.R. Sullivan, J. Appl. Phys. 99 (2006) 08H106.
[6] Y. Krupskaya, C. Mahn, A. Parameswaran, A. Taylor, K. Kra¨mer, S. Hampel,
A. Leonhardt, M. Ritschel, B. Bu¨chner, R. Klingeler, J. Magn. Magn. Mater. 321
(2009) 4067.
[7] B. Mehdaoui, A. Meffre, L.-M. Lacroix, J. Carrey, S. Lachaize, M. Respaud,
M. Gougeon, B. Chaudret, J. Appl. Phys. 107 (2010) 09A324.
[8] F. Dumestre, B. Chaudret, C. Amiens, P. Renaud, P. Fejes, Science 303 (2004)
821.
[9] L.-M. Lacroix, S. Lachaize, A. Falqui, M. Respaud, B. Chaudret, J. Am. Chem. Soc.
131 (2009) 549.
[10] E. Snoeck, C. Gatel, L.M. Lacroix, T. Blon, S. Lachaize, J. Carrey, M. Respaud,
B. Chaudret, Nano Lett. 8 (2008) 4293.
[11] L.-M. Lacroix, S. Lachaize, A. Falqui, T. Blon, J. Carrey, M. Respaud,
F. Dumestre, C. Amiens, O. Margeat, B. Chaudret, P. Lecante, E. Snoeck,
J. Appl. Phys. 103 (2008) 07D521.
[12] M. Klokkenburg, B.H. Erne´, J.D. Meeldijk, A. Wiedenmann, A.V. Petukhov,
R.P.A Dullens, A.P. Philipse, Phys. Rev. Lett. 97 (2006) 185702.
[13] L.-M. Lacroix, R. Bel Malaki, J. Carrey, S. Lachaize, M. Respaud, G.F. Goya,
B. Chaudret, J. Appl. Phys. 105 (2009) 023911.
[14] R. Hergt, W. Andra¨, C.G. d’Ambly, I. Hilger, W.A. Kaiser, U. Richter,
H.-G. Schmidt, IEEE Trans. Magn. 34 (1998) 3745.
[15] D. Kechrakos, K.N. Trohidou, Phys. Rev. B 58 (1998) 12169.
[16] A. Jordan, R. Scholz, K. Maier-Hauff, M. Johannsen, P. Wust, J. Nadobny,
H. Schirra, H. Schmidt, S. Deger, S. Loening, W. Lanksch, R. Felix, J. Magn.
Magn. Mater. 225 (2001) 118.
5. Conclusion
Our measurements of very large SARs on metallic iron NPs
above 1 kW/g confirm the potential of high magnetization MNPs
for future hyperthermia applications when large SAR would be
required. However, progress in the control of their coercive field
and of their magnetic interactions should be done. Furthermore,
these nano-objects are not ready yet for biomedical applications
[17] F. Delpech, C. Nayral, N. El Hawi, WO/2009/071794.