ARTICLE

Maximizing Hysteretic Losses in

Magnetic Ferrite Nanoparticles via
Model-Driven Synthesis and Materials
Optimization
Ritchie Chen,† Michael G. Christiansen,† and Polina Anikeeva†,‡,*
†
Department of Materials Science and Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, United States and
‡
Research Laboratory of Electronics, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, United States

ABSTRACT This article develops a set of design guidelines for maximizing heat
dissipation characteristics of magnetic ferrite MFe2O4 (M = Mn, Fe, Co) nano-
particles in alternating magnetic fields. Using magnetic and structural nanopar-
ticle characterization, we identify key synthetic parameters in the thermal
decomposition of organometallic precursors that yield optimized magnetic
nanoparticles over a wide range of sizes and compositions. The developed
synthetic procedures allow for gram-scale production of magnetic nanoparticles
stable in physiological buffer for several months. Our magnetic nanoparticles display some of the highest heat dissipation rates, which are in qualitative
agreement with the trends predicted by a dynamic hysteresis model of coherent magnetization reversal in single domain magnetic particles. By combining
physical simulations with robust scalable synthesis and materials characterization techniques, this work provides a pathway to a model-driven design of
magnetic nanoparticles tailored to a variety of biomedical applications ranging from cancer hyperthermia to remote control of gene expression.

KEYWORDS: magnetic nanoparticles . magnetic hyperthermia . anisotropy energy . organometallic decomposition .

water-soluble nanoparticles . large-scale synthesis

T
he ability to tune the materials intended to limit nonspecific heating of
properties of magnetic nanoparticles healthy tissue via eddy currents induced
(MNPs) composed of spinel ferrites by the applied AMF.9 As a result, the field
MFe2O4 (M = Mn, Fe, Co) enables a variety parameters are usually limited to ampli-
of biomedical applications.1
3 For example, tudes 5
20 kA m
1 and frequencies below
achieving a high magnetic moment is es- 1 MHz. To achieve the desired therapeutic
sential for effective magnetic resonance effect under the field frequency product
imaging (MRI) contrast agents, while max- constraint, the MNP power dissipation rate
imizing power dissipation in alternating per gram, or specific loss power (SLP), has to
magnetic fields (AMF) is desirable for heat- be optimized. Because the hysteresis losses
induced necrosis of tumor tissues.4,5 Recent in MNPs depend on their saturation magne-
work demonstrating fine control over the size, tization (Ms) and the effective anisotropy
shape, composition, and surface passivation energy barrier, these two parameters can
of superparamagnetic MNPs has propelled be varied for SLP optimization for an AMF of
the latter method, magnetic hyperthermia, a given amplitude and frequency. Guided
into clinical trials.6 AMF-induced heat dissipa- by the physical model of coherent magne-
tion in MNPs has also found new applications tization reversal in single domain MNPs, this * Address correspondence to
[email protected].
in remote control of cellular signaling and article correlates the structural and mag-
gene transcription in vivo.7,8 netic properties of ferrite MNPs (MFe2O4, Received for review July 10, 2013
For therapeutic purposes, the MNPs must M = Mn, Fe, Co) to key synthetic parameters and accepted September 9, 2013.
be administered at the lowest concentra- that determine the MNP performance as
Published online September 09, 2013
tions possible, and the product of AMF heat dissipation agents.10,11 The synthetic 10.1021/nn4035266
frequency f and amplitude Ho should be procedures developed within this study are
less than 5  109 A m
1 s
1, a figure of merit easily scalable to production of gram scale C 2013 American Chemical Society

CHEN ET AL. VOL. 7 ’ NO. 10 ’ 8990–9000 ’ 2013 8990

www.acsnano.org
 ARTICLE
quantities of monodisperse MNPs with anisotropy for therapeutic purposes may significantly perturb
energies varying over 3 orders of magnitude in a size the anisotropy barrier, particularly for materials with
range of 8
28 nm stable in physiological buffers for low anisotropy energy density values (i.e., MnFe2O4,
several months. We experimentally observe some of Keff = 3.0  103 J m
3 or Fe3O4, Keff = 1.4  104 J m
3).
the highest SLP values reported to date at a given AMF At sufficiently high applied fields, the material ap-
amplitude and frequency, which are in qualitative proaches saturation and should no longer be expected
agreement with a generalized physical model of hys- to respond linearly. As a result, SLP values cannot be
teretic power dissipation that can be used for the made arbitrarily large by increasing the field amplitude
simulation-driven design of MNPs tailored for a specific and frequency, as predicted by the functional form of
biomedical application. LRT. Thus at therapeutically relevant field amplitudes,
well-motivated use of LRT-based calculations is limited
THEORETICAL BASIS to materials with comparatively high anisotropy en-
A MNP ensemble dissipates heat when magnetic ergy, such as CoFe2O4 (Keff = 2.0  105 J m
3). Assum-
moments of individual MNPs overcome an anisotropy ing that MNPs with coercive fields much higher than
energy barrier to realign with an applied field to reduce the applied field are able to rotate freely, the majority
their configurational energy. While spinel ferrites have of the hysteretic loss is attributable to frictional
cubic anisotropy in their bulk form, this work assumes heat generated by the rotation of the particle in the
an effective uniaxial symmetry due to surface effects medium.15
that introduce multiple sublattices on the surface of In the limit of field amplitudes larger than the
MNPs at the nanoscale.6,12 As an estimate, the effective coercive field, the shape and area of the resulting
uniaxial anisotropy energy constant (Keff) was taken to hysteresis loops approaches, but does not exceed
be approximately equal to the absolute value of the the theoretical limit for uniaxial single-domain MNPs
first-order cubic magnetocrystalline anisotropy con- described by the Stoner
Wohlfarth model at 0 K.16 To
stant KI. The transition of moments over the barrier is take into account thermal activation and perturbative
thermally activated, so that the expected power dis- time scale, an effective HC can be considered to vary
sipation depends not only on the amplitude of the with AMF frequency as well as the temperature of the
applied field's perturbation to the anisotropy energy environment.17
landscape, but also on the ambient temperature and In the intermediate regime, when the AMF ampli-
cyclic time scale of the perturbation. At a field magni- tude is less than the coercive field but still perturbs the
tude determined ideally by the materials' saturation barrier significantly, a numerical method is required to
magnetization and the effective anisotropy energy, the effectively model the SLP. We adopted the model by
barrier to coherent reversal vanishes for uniaxial MNPs. Carrey et al. to calculate the hysteresis loops for a set of
This critical field magnitude might be called the zero ferrite materials at these conditions for MNPs with
temperature coercive field, though it should be under- diameters between 5 and 30 nm.10 This model employs
stood that this terminology neglects the temperature a “macrospin” approximation that assumes moments
dependence of KI, and thus will not reflect the coercive are confined to local energy minima. Frictional losses
field actually observed near 0 K.13 When we consider from Brownian rotation are not considered because
the relative magnitude of the applied field to the zero they do not contribute significantly to heat dissipation
temperature coercive field, the hysteretic power dis- for the range of MNP sizes explored in this paper.
sipation of MNPs can be conveniently divided into To illustrate the origin of heat dissipation in MNPs,
several regimes. we first calculate hysteresis loops for magnetite
At field amplitudes well below the coercive field, the Fe3O4 with varying diameters with applied AMFs at
barrier is not significantly perturbed and the magneti- amplitude H0 = 15 kA m
1 and frequency f = 500 kHz
zation of the MNP ensemble scales linearly with the (Figure 1A). We assume that the MNPs are effectively
applied field. In this regime, hysteretic losses can be uniaxial, their easy-axes aligned with the applied
appropriately modeled with linear response theory field, and that the attempt rate is constant at 1010 Hz.
(LRT), such as the frequently cited treatment by Because of surface effects, an effective uniaxial anisot-
Rosensweig.14 In that work, coherent magnetization ropy is assumed for ferrite MNPs as long as the
reversal (“Neél relaxation”) was considered in the con- remanent to saturation magnetization ratio is less than
text of an unperturbed and stochastic magnetization 0.5, which can be experimentally determined.18 These
reversal, with the implication that physical rotation of assumptions are made primarily for convenience and
the particle must dominate for maximal heat dissipa- are expected to overestimate SLP values while making
tion. However, it has been experimentally demon- reasonable qualitative predictions. Neither perfect
strated that increasing the viscosity of the fluid alignment nor random alignment are well motivated
surrounding MNPs did not dramatically alter their SLPs, assumptions for MNPs that can freely rotate, and the
indicating that magnetization reversal can lead to actual behavior would likely involve intermediate cor-
significant heating.15 Field amplitudes commonly used relation and vary with the effective anisotropy energy

CHEN ET AL. VOL. 7 ’ NO. 10 ’ 8990–9000 ’ 2013 8991

www.acsnano.org
 ARTICLE
Keff value. LRT was used to determine the SLP values
for heat dissipation, which arises primarily from the
frictional heating generated by the physical rotation of
the particles when the applied field amplitude is much
smaller than the coercive field. On the other hand, the
Keff value of MnFe2O4 is an order of magnitude lower
than that of Fe3O4, which results in this transition from
a superparamagnetic to ferromagnetic hysteretic loss
to occur at larger MNP diameters. In fact, we find that
MnFe2O4 MNPs exhibit significant hysteretic losses
only at diameters greater than 22 nm.

RESULTS AND DISCUSSION

A comprehensive palette of ferrite MNPs of varying
size and composition is essential for an experimental
evaluation of the predictive ability of the model illu-
strated in Figure 1. While thermal decomposition of
organometallic precursors in the presence of coordi-
nating ligands is an effective method to prepare
monodisperse and uniformly shaped MNPs, achieving
bulk values of magnetic properties in a variety of MNP
sizes remains a synthetic challenge.19
21 To produce a
wide range of ferrite MNPs, we have built upon and
contrasted two major organometallic syntheses based
on the thermal decomposition of metal
oleate and
metal-acetylacetonate (acac) precursors.20,21 While
Figure 1. SLP calculations based on a physical model of
hysteretic power loss. (A) Field-dependent magnetization thermal decomposition of metal
oleate precursors
curves from numerical simulations for Fe3O4 MNPs of offers fine control over MNPs size and composition,
diameters varying between 10 and 24 nm. (B) SLP deter- the material's Ms is known to be poor as compared to
mined by integrating the area of hysteresis loops for Fe3O4,
MnFe2O4, and CoFe2O4 as a function of MNP diameter. Field bulk values, which leads to low hysteretic losses.22 This
parameters used were H0 = 15 kA m
1 and f = 500 kHz. is attributed to the formation of wüstite (Fe1‑xO, where
x = 0.05
0.17),23 an antiferromagnetic phase that is
of the MNP relative to the ambient thermal energy.10 paramagnetic at room temperature and hence does
For MNPs with diameters less than 15 nm (at f = not contribute to hysteretic power dissipation in these
500 kHz), the hysteresis loop appears almost reversible MNPs.24 In contrast, a combination of reducing and
due to the negligible anisotropy barrier for the given oxidizing ligands present during the thermal decom-
temperature and frequency. This superparamagnetic position of metal-acac precursors yields MNPs com-
behavior results in low power loss (Figure 1B). As the prising mixed Fe2þ/Fe3þ necessary for the magnetite/
energy barrier scales with MNP volume, Ea ∼ Keff 3 d3, maghemite spinel ferrite structure.20 The resulting
the area of the hysteresis loop becomes significantly MNPs exhibit high saturation magnetization values
larger and displays a ferromagnetic shape at diameters approaching that of bulk. However, acac-based syn-
above 18 nm. SLP reaches its maximum value for iron thesis produces MNPs below 10 nm in diameter and
oxide MNPs with 20 nm diameters, which corresponds consequently requires a multistep seed-mediated
to a hysteresis loop with the largest area at the chosen approach to grow shells around the MNPs in 1 nm
field parameters. As the MNP diameter increases be- increments. In this study, we overcome the challenges
yond 20 nm, the anisotropy energy increases such that of these two synthetic routes to produce high-quality
the field amplitude no longer exceeds the coercive magnetic materials desirable for biomedical applica-
field and only minor hysteresis loops can be accessed, tions as well as essential to experimental validation of
leading to a decrease in overall heat dissipation. the physical model.
Figure 1B summarizes the dependence of the SLP Thermal Decomposition of metal
oleate Precursors. Two
on the diameter as well as the Keff value of the MNPs. key modifications are introduced into the previously
Because the Keff value of CoFe2O4 is an order of reported synthesis based on the thermal decomposition
magnitude greater than that of Fe3O4, only minor of metal
oleate precursors12,14 to produce MFe2O4
hysteresis loops are accessed at the chosen AMF (where M = Mn, Fe, or Co) with narrow size distributions
amplitude. Our calculations do not account for cubic and diameters tunable from 11 to 24 nm (Figure 2).
anisotropy of the ferrite material and hence does not First, the heating rate is reduced by three times, as
hold for materials like CoFe2O4 with especially large the previously reported heating 3.3 °C min
1 yielded

CHEN ET AL. VOL. 7 ’ NO. 10 ’ 8990–9000 ’ 2013 8992

www.acsnano.org
 ARTICLE
synthesis of MnFe2O4 and CoFe2O4 MNPs (Figure 2E
J).
While it has been hypothesized that nucleation and
growth occur separately at 240 °C and ∼300 °C,
respectively,21 recent reports show that the transition
occurs within 10 °C of each other,23 and that the bulk of
the homogeneous nucleation occurs above 300 °C.25,26
This suggests that there is overlap between the two stages
and hence reducing the heating rate allows for sufficient
time for nanocrystals to nucleate and grow isotropically as
the precursor fully decomposes above 300 °C.
The second modification is the use of the same
solvent, 1-octadecene, for the synthesis of all MNPs.
In this case, the MNP diameter is simply determined
by setting the final annealing temperature between a
range of 300
325 °C rather than choosing solvents
with different boiling points. Iron oxide MNPs varying
from 11 to 24 nm in diameter are synthesized when the
annealing temperature is increased from 305 to 325 °C
(Figure 2A
D). However, under identical conditions,
MnFe2O4 nanoparticles are cubic (Figure S2A), so the
solvent amount needs to be decreased from 25 to
10 mL to promote supersaturated isotropic growth. In
10 mL 1-octadecene, MnFe2O4 MNPs with diameters
10
18 nm are synthesized when the final temperature
is varied between 305 and 325 °C (Figure 2E
H).
Decreasing the solvent to 5 mL leads to polydisperse
MNP samples. Size tuning of CoFe2O4 MNPs using this
approach is limited to j20 nm (Figure 2I,J) because of
preferential facet growth above 310 °C.25 Our simpli-
fied size tuning procedure allows for straightforward
scaling of the MNP production up to gram-scale quan-
tities (Figure S3). An ability to consistently produce
multiple grams of monodisperse MNPs is essential for
standardized biomedical experimentation as well as
for future clinical applications of these materials.
Thermal Decomposition of Metal Acetylacetonate Precursors.
In addition to the metal
oleate based chemistry,
we have also employed metal-acac precursors to pro-
duce tertiary ferrite MNPs (Figure 3). A synthetic route
adapted from Sun et al. produces CoFe2O4 and
Figure 2. Transmission electron microscopy (TEM) images
of MNPs synthesized via the thermal decomposition of MnFe2O4 MNPs of 9 and 7 nm in diameter, respectively
metal
oleate precursors. The diameter, the standard devia- (Figure 3A,D).20 This synthetic procedure requires
tion (nm) and the synthesis temperature are indicated on maintaining the reaction temperature at 200 °C for 2
the images. All reactions were heated from 200 °C to the
indicated temperature at 1 °C/min unless otherwise noted. h to promote nuclei formation necessary for mono-
(A
D) Iron oxide MNPs synthesized at 305, 310, 325 °C in dispersity prior to raising the temperature to reflux.
1-octadecene yielding 11, 16, and 22 nm diameter MNPs, However, the prolonged nucleation period reduces the
respectively. The 24 nm MNPs were synthesized at 330 °C in
1-eicosene. (E
H) Manganese ferrite MNPs synthesized at final MNP size during growth. To produce MNPs great-
305, 315, 325 °C in octadecene yielding 11, 16, and 19 nm er than 10 nm in diameter, we replace Mn(acac)2 and
diameter MNPs, respectively. The 28 nm MNPs were synthe- Co(acac)2 with the chloride salts MnCl2 and CoCl2 and
sized at 330 °C in eicosane. (I and J) The 13 nm cobalt ferrite
MNPs synthesized at 305 °C. For 20 nm particles, the heating directly heat the reaction solution to reflux at a rate of
rate was increased to 3.3 °C/min to 310 °C. Scale bar = 20 nm. 3.3 °C/min. Because Fe(acac)3 decomposes at a tem-
perature different from Mn(acac)2 and Co(acac)2,27
slightly faceted iron oxide nanoparticles in our reaction replacing the Mn2þ or Co2þ source with chloride salts
conditions (Figure S1). The reduced heating rate of results in direct incorporation as Fe(acac)3 decom-
1 °C min
1 yields monodisperse and spherical iron oxide poses. These modifications yield CoFe2O4 and
MNPs with a diameter distribution of less than 5% MnFe2O4 MNPs with diameters 12 and 11 nm, respec-
(Figure 2A
D). Similar trends are also observed in the tively (Figure 3B,E).

CHEN ET AL. VOL. 7 ’ NO. 10 ’ 8990–9000 ’ 2013 8993

www.acsnano.org
 ARTICLE
TABLE 1. Summary of Magnetic Properties of As-Synthesized
and Water-Soluble MNPs at 300 Ka

d dmag d*mag Ms (300 K) M*s (300 K)

sample (nm) (nm) (nm) (emu/g) (emu/g) Φferrimagnetic

Iron Oxide Oleate 11 9.2 9.1 28 70 0.58 (0.56)

16 8.5 13.1 18 69 0.15 (0.55)
18 9.6 16.6 39 64 0.15 (0.78)
22 9.8 18.8 41 65 0.09 (0.62)
24 10.5 16.3 22 67 0.08 (0.31)
MnFe2O4 Oleate 11 7.7 7.6 8 47 0.34 (0.33)
16 10.8 10.2 13 54 0.31 (0.26)
19 9.8 9.6 5 25 0.14 (0.13)
28 9.0 11.4 3 31 0.03 (0.07)
CoFe2O4 Oleate 13 15.2 30.8
20 3 7
MnFe2O4 Acac 7 6.6 6.5 53 51 0.84 (0.80)
11 9.2 10.6 75 74 0.58 (0.89)
26 10.9 25.3 86 92 0.07 (0.92)
CoFe2O4 Acac 9 39 37
12 60 62
14 58 59
a
Average diameters (d) were extracted from TEM images. Magnetic diameters
(dmag) were obtained from linear fits of root temperature hysteresis curves in the
low field range. *Indicates the sample was measured from water-soluble MNP
solutions. The Φferrimagnetic indicates the volume fraction that is ferromagnetic. The
Figure 3. TEM images of MFe2O4 MNPs synthesized from bracketed values indicate the volume fraction after phase transfer into water.
the thermal decomposition of metal-acetylacetonate (acac)
precursors. The size and the standard deviation (nm) are indicate that the MNPs synthesized from metal
oleate
indicated on the images. (A
C) Cobalt ferrite MNPs with
diameters of 9, 12, and 14 nm. (D
F) Manganese ferrite precursors have low saturation magnetization Ms
MNPs with diameters 7, 11, and 26 nm. Scale bar = 20 nm. values comparable to the bulk material (Table 1 and
Figure 4A). Comparatively low Ms values for small MNPs
As is evident from Figure 1B, the low magnetocrys- can be attributed to the formation of a magnetically
talline anisotropy constant Keff of MnFe2O4 necessitates frustrated surface layer due to incomplete coordina-
MNPs with diameters greater than 20 nm for appreci- tion of metal ions. This effect typically becomes less
able SLP values to be observed. Replacing benzyl ether significant leading to increased Ms with increasing
with a higher boiling point temperature solvent like MNP diameter.30 Surprisingly, MNPs synthesized
1-octadecene yields polydisperse samples (Figure S4A).28 via the metal
oleate based route exhibit further de-
Monodisperse 16 nm MNPs are synthesized when a crease in Ms with increasing particle size. Saturation
higher molar ratio of ligand to solvent is used in dioctyl magnetization values for iron oxide MNPs fall from
ether (Figure S4B).29 However, increasing the tempera- ∼35 emu g
1 to less than 20 emu g
1 for MNP
ture in different solvents is found to be insufficient to diameters increasing from 5 to 22 nm.21 To further
increase the nanoparticle size above 20 nm. investigate the decrease of the Ms with MNP size, we
While the previous seed-mediated approach can calculate the magnetic volume, a proxy for the magnetic
only increase the MNP diameter in 1 nm increments, moment that assumes bulk saturation magnetization of
resulting in cumbersome multistep procedures for the magnetized material, from room temperature mag-
MNPs of e14 nm in diameter,20 here we developed netization curves using a linear fit for static magnetic
a straightforward process allowing for 5
7 nm shell susceptibility in the low field limit. The behavior of
growth in a single reaction step. By eliminating the 2 h randomly oriented particles converges on the Langevin
nucleation period and directly heating the reaction function in the limit of low fields regardless of their
solution to reflux, we have synthesized 14 nm CoFe2O4 anisotropy energy.10 We find that the as-synthesized
and 26 nm MnFe2O4 MNPs using seeds with diameters magnetic diameter of these MNPs never increases be-
of 7 and 11 nm, respectively (Figure 3C,F). yond 10 nm despite a larger measured physical diameter
Magnetic Properties of MNPs. We first investigated the (Table 1 and Figure 4B). Similarly, for the ternary metal
magnetic behavior of the as-synthesized materials oxides, the measured saturation magnetization also de-
dispersed in toluene by measuring the field depen- creases with increasing MNP size from 54 to 31 emu g
1
dence of the magnetization at room temperature for MnFe2O4 and 15 to 3 emu g
1 for CoFe2O4 MNPs.
and at 5 K. Room temperature magnetization curves Furthermore, the as-synthesized MNPs do not
measured by vibrating sample magnetometer (VSM) exhibit saturation even at high fields, suggesting the

CHEN ET AL. VOL. 7 ’ NO. 10 ’ 8990–9000 ’ 2013 8994

www.acsnano.org
 ARTICLE
MNPs, the observed exchange bias, paramagnetic-like
susceptibility, and small magnetic cores of our MNPs
suggest instead the presence of a wüstite phase.
Powder X-ray diffraction (XRD) studies clearly demon-
strate the presence of both wüstite and magnetite/
maghemite phases, with the (111), (200), (220), (311),
and (222) peaks from the wüstite fcc structure clearly
identified (Figure 6A). Wüstite has been observed as an
intermediary species when the synthesis environment
is not sufficiently oxidizing to form maghemite or
magnetite phases, which is characteristic of thermal
decomposition of metal
oleate precursors.23,35,36 At
the nanoscale, the magnetization of wüstite phase
does not saturate even at high fields at 5 K.24 Com-
bined with evidence from literature, our data suggests
that the MNP samples may be comprised of a spinel-
like phase coherently embedded in a wüstite matrix.
Because wüstite is a metastable phase, conversion to
magnetite can be readily achieved under certain
conditions.22,35 The diameters of MNPs with composi-
tion MFe2O4 (M = Mn, Fe, Co) synthesized by the
thermal decomposition of metal
oleate precursors
can be easily tuned while maintaining narrow size
distribution; however, the initial as-synthesized state
does not make the material particularly suitable for
remote heating applications due to its low Ms.
Figure 4. Magnetic properties of iron oxide MNPs in the as- In contrast, thermal decomposition of metal-acac
synthesized (open circles) and water-soluble state (filled-in precursors produces monodisperse MFe2O4 MNPs with
circles). (A) Ms increases after high-temperature annealing Ms comparable to bulk values in the as-synthesized
in phase transfer step. (B) Increase in magnetic diameter
trending linearly with increasing MNP size. Solid black line state (Table 1). As expected, the saturation magnetiza-
represents 1:1 correspondence, gray dashed lines are from tion of CoFe2O4 and MnFe2O4 MNPs increases with
linear fits to the data with intercept set at 0. increasing MNP diameter reaching 63 and 95 emu g
1
presence of a paramagnetic-like phase (Figure 5A). for CoFe2O4 and MnFe2O4, respectively, above 10 nm.
The inset in Figure 5A details the field-dependent Hysteresis loops of the MnFe2O4 MNPs, 11 nm in
magnetization at low fields and presents evidence of diameter, measured at 5 K indicate a low coercive field
exchange-bias, which arises not only in composition- of 250 Oe with no evidence of exchange bias
ally hybrid structures but also at order
disorder inter- (Figure 5B). CoFe2O4 MNPs display a markedly higher
phases.31 The example hysteresis curves of as-synthe- coercivity of 20 kOe with no exchange bias. A dip
sized 18 nm iron oxide and 11 nm MnFe2O4 MNPs in Ms to 28 emu g
1 is measured for the 16 nm
display a shift toward negative fields by 1300 and 850 MnFe2O4 MNPs. This may be attributed to the strong
Oe, respectively. Because the characteristic length for reducing environment of excess oleylamine, which
exchange bias is on the order of 1
2 nm nano- was employed to increase the MNP diameter beyond
meters,32 the observed shifts cannot be attributed 11 nm.35 MNPs of MnFe2O4 with diameters of 26 nm
to dipole
dipole interactions between particles with synthesized via our one-step heterogeneous nuclea-
different magnetic phases. As the size of the MNP tion exhibits the highest measured Ms (95 emu g
1) in
increases, the decreasing magnetic diameter indicates the synthesized MNP set due to its increase in volume
that less than 20% of the volume fraction of the MNP is to surface area ratio.
ferromagnetic, suggesting heterogeneity within the as- The tertiary ferrite MNPs synthesized from metal
synthesized MNPs (Table 1). Levy et al. have shown that acetylacetonate precursors exhibit saturation behavior
regions of structural disorder in their iron oxide nano- at high fields, Ms values comparable to those of bulk,
particles contribute to this effect since the moments of and no evidence of exchange bias. Because the syn-
the magnetically frustrated phases do not fully align thesis involves a mixture of oleic acid, oleylamine,
even at high field.33 The observed exchange bias in and 1,2-hexadecanediol with different reducing and
these samples indicates that a ferromagnetic and an oxidizing capabilities,35,37 the reaction conditions may
antiferromagnetic phase are in intimate contact.34 favor mixed valence states required for optimal mag-
While Levy et al. attribute the magnetically fru- netic properties of the ferrites. On the other hand, the
strated phases to strain in the crystal structure of their thermal decomposition of metal
oleate precursors

CHEN ET AL. VOL. 7 ’ NO. 10 ’ 8990–9000 ’ 2013 8995

www.acsnano.org
 ARTICLE
Figure 5. Field-dependent magnetization loops for tertiary ferrites measured at 5 K. (A) As-synthesized MNPs from thermal
decomposition of metal
oleate precursors. (Inset) Low-field region shows negative field shift characteristic of exchange bias.
(B) As-synthesized MNPs from thermal decomposition of metal-acac precursors. (C) MNPs from metal
oleate synthesis in (A)
after phase transfer into water. (D) MNPs from metal-acac synthesis in (B) after phase transfer into water.

only requires the addition of oleic acid and the reaction buffered saline (PBS) which has been attributed to the
conditions may not be sufficiently oxidizing to form formation of salt bridges (Figure S5).39
ferrimagnetic spinel phases. Interestingly, we find that the magnetic properties
High Temperature Phase Transfer of MNPs into Aqueous of the MNPs prepared by the thermal decomposition
Solutions. While high temperature thermolysis of orga- of metal
oleate precurors are significantly improved
nometallic precursors produces monodisperse MNPs during the phase transfer process. For iron oxide MNPs,
with controllable shape and size, the surface is gen- Ms values increase from 15 to 40 emu g
1 to an average
erally coated with hydrophobic ligands, which make of 67 emu g
1, which is comparable to bulk values
them unusable for biological applications in aqueous of magnetite (92 emu g
1, Figure 4A). Furthermore, the
environments. To render the nanocrystals hydrophilic, coercive field decreases from 1350 to 350 Oe and there
we employ high-temperature phase transfer in diethy- is no observable exchange bias (Figure 5C). The mag-
lene glycol, a high boiling point polar solvent miscible netic diameters are also found to correlate closely
with organic solvents such as toluene, to drive the with the physical MNP diameters determined by TEM,
coordination of poly(acrylic acid) (PAA) onto the surface and the ferromagnetic volume fraction exceeds 60%
of the MNPs at temperatures greater than 200 °C.38 This (Table 1). These improvements in the magnetic proper-
strategy is advantageous because the high tempera- ties can be attributed to the conversion of the wüstite
ture promotes the exchange of the original surfactant phase into a ferrimagnetic spinel phase. The high
with the polyelectrolyte through mass action. Further- temperature phase transfer step provides an oxidizing
more, the multiple binding sites prevent PAA desorp- environment that enables the transformation of
tion, while allowing its extension into water to make the metastable wüstite into maghemite and magnetite
nanocrystals highly soluble in aqueous media. Ligand phases, with no evidence of FeO peaks apparent in
exchange with PAA resulted in the MNPs to be highly the powder diffraction pattern (Figure 6B). Our data
stable in Good's buffers such as tris(hydroxymethyl)- are in agreement with prior reports that observe the
aminomethane (Tris) and (4-(2-hydroxyethyl)-1- conversion of wüstite MNPs into Fe3O4 and γ-Fe2O4
piperazineethanesulfonic acid) (HEPES), although time- during annealing in air at temperatures between
dependent aggregation was observed in phosphate 140
200 °C.35

CHEN ET AL. VOL. 7 ’ NO. 10 ’ 8990–9000 ’ 2013 8996

www.acsnano.org
 ARTICLE
Figure 7. SLP measurements as a function of MNP size
(measured by TEM) obtained at H0 = 15.5 ( 1.4 kA m
1
and f = 500 kHz. The calculations based on theoretical
models are normalized to the maximum SLP value calcu-
lated for 20 nm iron oxide MNPs, and the experimental SLP
are normalized to the metal content determined from
elemental analysis.

(g/mL of the ferrofluid), and (dT/dt) is the experimentally

measured slope of the temperature increase as a func-
tion of time inside the AMF. The resulting SLP values are
Figure 6. Powder X-ray diffraction patterns of (A) as-synthe-
sized iron oxide nanoparticles with the wüstite FeO phase plotted in Figure 7 as a function of MNP size.
marked with a red asterisk, and (B) after phase-transfer into The trends simulated in Figure 1B are qualitatively
water. (C) Reference pattern of bulk Fe3O4 (black) and FeO reproduced in the experimental measurements illu-
(red).
strated in Figure 7. CoFe2O4 MNPs from both oleate
The Ms values of the MnFe2O4 and CoFe2O4 MNPs and acac-based syntheses do not yield significant
synthesized from metal
oleate precursors improves hysteretic losses under the given AMF parameters,
dramatically by 4
10 times; however, their extremely and the observed heat dissipation originates from
low starting values still impede the application of these their Brownian rotation.15 In the case of MnFe2O4 with
MNPs for efficient heat dissipation (Table 1). The comparatively low Keff values, the small magnitude of
magnetic properties of the MNPs synthesized from the anisotropy energy barrier of the MNPs smaller than
the metal-acac precursors are preserved during the 22 nm in diameter relative to the thermal energy at
phase transfer (Figure 5D). Furthermore, the ferromag- room temperature results in superparamagnetic like
netic volume fraction for MnFe2O4 MNPs is found to behavior, with little or no observable hysteretic loss.
exceed 80% (Table 1). These observations also support Only MNPs with diameters of 26 nm exhibit significant
our initial treatment of iron oxide and MnFe2O4 MNPs heating. For materials with intermediate anisotropy
as having uniaxial anisotropy because the remanent to energies such as iron oxide, an optimal size range
saturation magnetization ratio is calculated to be less between 18 and 22 nm is identified for the applied
than 0.5 from the SQUID hysteresis curves, while the AMF parameters. The SLP value of iron oxide MNPs
value calculated for CoFe2O4 MNPs is 0.84 and suggests 22 nm in diameter was measured to be 716 ( 31 W g
1,
cubic anisotropy.18 which is among the highest recorded values for syn-
Measurements of the Specific Loss Power. To experimen- thetic ferrite materials at the given frequency and field
tally verify the predictive capabilities of the numerical strength.40 This behavior is consistent with the calcu-
calculations, heat dissipation of the MNPs dispersed in lated areas of the hysteresis loops in Figure 1A, which
water is measured during exposure to an AMF produced increase progressively due to a transition from rever-
by a homemade coil with a soft ferromagnetic toroid sible superparamagnetic to ferromagnetic hysteresis
core driven by a RLC circuit. MNP solutions (2 mg/mL) in loss regimes.
water are placed into an AMF of amplitude H0 = 15.5 ( Also of note is the difference in SLP between tertiary
1.4 kA m
1 and frequency f = 500 kHz. The temperature ferrite MNPs prepared from the two different synthetic
increase is recorded as a function of time, and the SLP is routes. MNPs prepared from the metal
oleate precur-
calculated from the slope using the expression: sors have lower measured SLP values compared to
similar sized MNPs prepared from metal-acac precur-
C dT sors consistent with their lower magnetization at a
SLP ¼
m dt given AMF amplitude. For example, 11 nm MnFe2O4
where C is the specific heat capacity of water per unit MNPs prepared from acac precursors have SLP values
volume (C = 4.184 J K
1 mL
1), m is the concentration of 35 W g
1 while MNPs prepared from oleate

CHEN ET AL. VOL. 7 ’ NO. 10 ’ 8990–9000 ’ 2013 8997

www.acsnano.org
 ARTICLE
precursors have SLP values of 13 W g
1. Furthermore, correlation with simulation was observed. Using our
their low ferromagnetic volume fraction leads to de- synthetic procedures we achieved some of the highest
viations from numerical simulations as seen in the heating rates measured with 22 nm iron oxide MNPs at
measured SLP value of 44 W g
1 for 28 nm MNPs the specified magnetic field conditions.
prepared from oleate precursors. In contrast, MnFe2O4
MNPs prepared from acetylacetonate chemistries CONCLUSION
trend with predictions due to their optimized magnetic In this article, the domain of applicability of coherent
properties, with a measured SLP of 304 W g
1. magnetization reversal model in single-domain MNPs
In summary, we find that tuning of the MNP di- is generalized to field amplitudes and frequencies
ameter using thermal decomposition of metal
oleate relevant to magnetic hyperthermia, and heat dissipa-
precursors can be accomplished by simply changing tion is evidently correlated to the anisotropy energy of
the final reaction temperature, while monodispersity the material. Within a framework of this model a suit of
can be achieved with slow heating rates. Furthermore, synthetic procedures is developed to produce water-
the magnetic properties of the MNPs produced by this soluble MNPs with magnetocrystalline anisotropy en-
method can be significantly improved by annealing at ergy varying over 3 orders of magnitude in gram-scale
high temperatures during the ligand transfer step, quantities (Figure S3).
which results in the conversion of the antiferromag- While magnetic hyperthermia for cancer applica-
netic wüstite phase present in the as-synthesized tions has benefitted from decades of research, recent
MNPs into a ferrimagnetic mixed spinel phase neces- studies have demonstrated the use of hysteretic heat
sary for efficient hysteretic heat dissipation. In addition dissipation by MNPs for biomedical treatments beyond
to metal
oleate based synthesis, we have explored tumor necrosis. These innovative applications, such as the
metal-acac based synthesis, which requires the use of remote control of action potential firing in neurons,
ligands with different reducing capabilities, to produce demand MNPs with optimized magnetic properties to
MNPs with saturation magnetization values approach- achieve therapeutic effects at biologically relevant time
ing bulk. While previous reports of metal-acac synthe- scales. Thorough understanding and tailoring of the
sis have employed incremental coating of ∼1 nm magnetocrystalline anisotropy energy and the saturation
layers to produce MNP diameters above 10 nm, our magnetization of the MNPs at a given field amplitude
synthetic route allows for a one-step growth of 5
7 nm and frequency allows for control over the hysteretic heat
thick shells, which produces high-quality MFe2O4 dissipation critical for these biomedical applications.
MNPs with diameters of up to 26 nm. All our MNPs Equipped with a large-scale reproducible synthetic tool-
are made soluble in aqueous environments and are box and a clearer understanding of MNP heat dissipation
stable in physiological buffers for several months. in the presence of an AMF, we can now begin to
These stable aqueous dispersions were subjected to standardize hyperthermia treatments by selecting materi-
an applied alternating magnetic field and their hys- als with the proper anisotropy energy relevant for a
teretic power loss properties were compared to the particular therapeutic application. Tailoring hysteretic
simulated trends illustrating the predictive ability power loss will enable precise control over local tempera-
of the coherent magnetization reversal model. When ture changes at cell surfaces decorated with MNPs, and
the properties of the MNPs were not optimal, poor may allow for fine temporal control over cellular functions.

MATERIALS AND METHODS complex and 2.5 mmol of oleic acid were dissolved in 25 mL of
1-octadecene in a 250 mL 3-neck flask and evacuated for 30 min.
Sodium oleate was purchased from TCI America. All other
Then the solution was heated to 200 °C under nitrogen flow,
solvents and reagents were purchased from Sigma-Aldrich and
used without further purification. further heated to 310 °C at a rate of 1 °C/min, and held at the
specified temperature for 1 h. After the heating mantle was
Preparation of Metal
Oleate Complex. The metal
oleate
MFe2(C18H33O2)8 precursor (where M = Fe, Co, Mn) was pre- removed and the solution was cooled to room temperature, an
pared by reacting sodium oleate and the respective metal entire reaction solution was transferred into a 50 mL conical tube
chloride salt.19,21 We scaled our preparation such that the total along with a 1:1 mixture of ethanol and hexane. The sample was
metal content was 60 mmol per reaction. For example, an iron
centrifuged at 6000 rpm for 10 min to collect the synthesized
cobalt metal complex CoFe2(C18H33O2)8 was prepared by dis- nanocrystals. The pelleted nanocrystals were redispersed in
solving 40 mmol of FeCl3, 20 mmol of CoCl2, and 160 mmol of 10 mL of hexane, flocculated with 5 mL of ethanol, and centri-
sodium oleate in 100 mL of ethanol, 100 mL of Milli-Q water, and fuged at 6000 rpm for 10 min twice to remove excess ligand
200 mL of hexane and heated to reflux at 60 °C for 4 h. After and solvent. The nanocrystals were then dispersed in toluene.
removal of the aqueous phase, the organic phase was heated to MNPs (10, 18, and 23 nm) were synthesized by setting the final
70 °C for 2 h and then placed under vacuum at 110 °C for an temperature to 305, 320, and 325 °C, respectively. Twenty-six
additional 2 h to remove residual solvent, leaving behind a nanometer MNPs were prepared by setting the final temperature
viscous metal
oleate product. to 330 °C in 25 mL of 1-eicosene.
Synthesis of Monodisperse and Spherical MFe2O4 Nanocrystals of To synthesize MnFe2O4 15 nm MNPs, 5 mmol of the
Different Sizes from Metal
Oleate Precursors. To synthesize 15 nm metal
oleate complex and 2.5 mmol of oleic acid were dis-
in diameter iron oxide nanoparticles, 5 mmol of the metal
oleate solved in 10 mL of 1-octadecene in a 250 mL 3-neck flask and

CHEN ET AL. VOL. 7 ’ NO. 10 ’ 8990–9000 ’ 2013 8998

www.acsnano.org
 ARTICLE
evacuated for 30 min. Then the solution was heated to concentration of the transition metal ions on a Horiba ACTIVA-S.
200 °C under nitrogen flow, further heated to 315 °C at a rate To remove residual ligands that may interfere with the degra-
of 1 °C/min, and held at the specified temperature for 1 h. MNPs dation of MNPs in aqua regia, we annealed the samples in air
(10 and 18 nm) were synthesized by setting the final tempera- at 400 °C for 4 h. Organic ligands are expected to be fully
ture to 305 and 325 °C, respectively. Twenty-eight nanometer decomposed above 350 °C. The amount of magnetic material in
MNPs were prepared by setting the final temperature to 330 °C a given volume of 100 μL was determined by ICP-AES after
in 10 mL of 1-eicosene. annealin,g then digesting in a solution of aqua regia (1:3 v/v
To synthesize CoFe2O4 13 nm MNPs, 5 mmol of the metal
37 wt % HCl to 70 wt % HNO3) overnight at 60 °C. The quantity of
oleate complex and 2.5 mmol of oleic acid were dissolved in metal determined was used to normalize all magnetization and
25 mL of 1-octadecene in a 250 mL 3-neck flask and evacuated SLP values that were experimentally obtained.
for 30 min. Then the solution was heated to 200 °C under Magnetic Characterization. MNPs in the as-synthesized state
nitrogen flow, further heated to 305 °C at a rate of 1 °C/min, and was dispersed in toluene and sealed in quartz tubes using
held at the specified temperature for 1 h. Twenty nanometer rubber stoppers. Similarly for water-soluble MNPs, measure-
MNPs were prepared by setting the final temperature to 310 °C ments were conducted with the sample dispersed in water.
with a change in heating rate from 1 to 3 °C/min. Hysteresis curves at 5 K were measured using a superconduct-
Synthesis of Monodisperse MnFe2O4 Nanocrystals of Different Sizes ing quantum interference device (SQUID, MPMS-XL, Quantum
from Metal Acetylacetonate (Acac) Precursors. Seven nanometer Design). Room temperature hysteresis curves were generated
MnFe2O4 MNPs were synthesized by previously reported meth- on a vibrating sample magnetometer (VSM, Digital Measure-
ods in a 250 mL 3-neck flask.20 To increase the size of the MNPs, ment Systems Model 880A).
MnCl2 instead of Mn(acac)2 was used. Eleven nanometer MNPs Powder X-ray Diffraction. Powder XRD diffraction patterns were
were synthesized by mixing Fe(acac)3 (2 mmol), MnCl2 (1 mmol), collected with a PANalytical Multipurpose Diffractometer
oleic acid (6 mmol), oleylamine (6 mmol), 1,2-hexadecanediol equipped with Cu KR radiation in the 2θ range of 20
80°.
(10 mmol), and 20 mL of benzyl ether and evacuated for 30 min. Samples were prepared from evaporating successive casting of
Then the solution was heated to reflux for 1 h at a rate of dispersed MNPs onto glass.
3.3 °C/min under nitrogen flow. MNPs 16 nm in diameter were Stability Test. A total of 0.1 mg of MNPs as determined by [Fe]
synthesized by mixing Fe(acac)3 (2 mmol), MnCl2 (1 mmol), oleic content was dispersed in either 1 Tris base, acetic acid and
acid (6.31 mmol), oleylamine (12.16 mmol), and 2 mL of dioctyl EDTA (TAE), HEPES, or 1 phosphate buffer saline (PBS) for one
ether and heated to 200 °C under nitrogen flow for 2 h. The month and its hydrodynamic diameter was compared to the
reaction was further heated to 330 °C at a rate of 3.3 °C/min. To same MNPs dispersed in Milli-Q water using dynamic light
grow MnFe2O4 MNPs greater than 20 nm, 50 mg of 11 nm MNP scattering (DLS, Malvern Nano ZS90).
seeds dispersed in hexane was added to Fe(acac)3 (2 mmol), Specific Loss Power Measurements. Calorimetry measurements
MnCl2 (1 mmol), oleic acid (6 mmol), oleylamine (6 mmol), 1,2- in the presence of an AMF were conducted with a custom built
hexadecanediol (10 mmol), and 20 mL of benzyl ether and setup. A toroid composed of a soft ferromagnetic material
evacuated for 30 min at 60 °C. Then the solution was heated to optimized for high frequency power transformers was specially
reflux for 1 h at a rate of 3.3 °C/min under nitrogen flow. machined to include a gap large enough to accommodate a
Nine nanometer CoFe2O4 MNPs were synthesized by pre- sample vial and used as a coil core. A transformer circuit with a
viously reported methods in a 250 mL 3-neck flask.20 To increase resistive ballast in the primary circuit was used to generate high,
the size of the nanoparticles, MnCl2 instead of Mn(acac)2 reasonably stable currents in the secondary while simulta-
was used. Eleven nanometer MNPs were synthesized by neously matching the impedance of the variable frequency
mixing Fe(acac)3 (2 mmol), CoCl2 (1 mmol), oleic acid (6 mmol), 200 W amplifier (1020L, Electronics & Innovation). In the sec-
oleylamine (6 mmol), 1,2-hexadecanediol (10 mmol), and 20 mL ondary, the coil acted as the resistive and inductive elements of
of benzyl ether and evacuated for 30 min. Then the solution was an RLC resonance circuit, with a high voltage series capacitor
heated to reflux for 1 h at a rate of 3.3 °C/min under nitrogen setting the resonant frequency. The field magnitude was mea-
flow. sured by a custom built probe employing a pickup loop and
Phase Transfer. The MNPs were transferred into water using an oscilloscope (TDS2022C, Tektronix). Error bars placed on
a protocol adopted from Zhang et al.38 A total of 0.5 g of the field value result primarily from the moderate heating of
poly(acrylic acid) (Mw ∼ 1800) was dissolved in 10 mL diethylene the core over the course of 30 s, which was mitigated by a
glycol and heated to 110 °C under nitrogen. One milliliter of simple cooling system circulating ice water to the coil via
MNPs dispersed in hexane (∼50 mg/mL) was injected into the silicone tubing.
mixture and then heated to reflux at 240 °C for 3 h. The sample Temperature measurements made by an AMF insensitive
was precipitated with 10 mL of 1 M HCl then washed twice with fiber optic temperature probe were recorded as an AMF was
water. A 50 mM NaOH solution was added to disperse the MNPs, applied for 30 s. Each measurement was repeated 8 times and
and the mixture was sonicated for 30 min. control samples with only water were measured after every
Previously reported studies from the thermal decomposi- 4 trials to determine the background heating rate. All samples
tion of iron oleate indicate a yield of greater than 95%, although were 1 mL with a MNP concentration of approximately 2 mg/mL.
the method to quantify this was not reported. On the basis of The SLP value measured was normalized to the metal content
the total MNP weight after removing organic matter by anneal- determined by elemental analysis.
ing at 400 °C for 4 h, we obtain a yield of ∼70% per synthesis Conflict of Interest: The authors declare no competing
based on the molarity of iron from the nanoparticles to the financial interest.
amount contained in the precursor (∼300 mg of as-synthesized
MNPs per reaction). Similarly, the yield of MNPs from the Acknowledgment. This work was supported in part by the
thermal decomposition of iron acac precursors is ∼60% and is Sanofi Biomedical Innovation Award and the MRSEC Program of
typical of that reported (∼80 mg of as synthesized MNPs per the National Science Foundation under award number DMR-
reaction). Approximately 40% of the MNPs is lost upon phase 0819762. R.C. is supported by the National Science Foundation
transfer into water. To obtain ∼1 g of water-soluble MNPs, we Graduate Research Fellowship. Authors thank X. Jia for her
scaled up the synthesis by 7 times and dispersed the MNPs after advice on transmission electron microscopy, D. C. Bono on his
phase transfer scaled by 10 times into phosphate buffered assistance with vibrating sample magnetometry, M. F. Rubner
saline (PBS) at a final concentration of 1 mg/mL. and C. A. Ross for their insightful comments on the manuscript.
Transmission Electron Microscopy (TEM). As-synthesized MNPs Supporting Information Available: TEM images of iron oxide
dispersed in hexane were drop casted onto carbon-coated MNPs synthesized from metal-oleate precursors with a heating
copper grids (Ted Pella, Inc.) and washed three times with rate of 3.3 °C/min to 320 °C; TEM images of MnFe2O4 MNPs
methanol. TEM images were taken using a JEOL 2010F electron synthesized from metal-oleate precursors in different amounts
microscope at 200 kV. of solvent at 320 °C; an image of iron oxide MNPs dispersed in
Elemental Analysis. Inductively coupled plasma atomic PBS at a concentration of 1 mg/mL produced via thermal
emission spectroscopy (ICP-AES) was used to determine the decomposition of iron oleate precursor scaled up by 7-times;

CHEN ET AL. VOL. 7 ’ NO. 10 ’ 8990–9000 ’ 2013 8999

www.acsnano.org
 ARTICLE
TEM images of MnFe2O4 MNPs synthesized from the thermal 21. Park, J.; An, K.; Hwang, Y.; Park, J.; Noh, H.; Kim, J.; Park, J.;
decomposition of metal acetylacetonate (acac) precursors in Hwang, N.; Hyeon, T. Ultra-Large-Scale Syntheses of Mono-
1-octadecene and dioctyl ether; DLS plots of intensity % vs size disperse Nanocrystals. Nat. Mater. 2004, 3, 891–895.
of iron oxide nanoparticles. This material is available free of 22. Salas, G.; Casado, C.; Teran, F. J.; Miranda, R.; Serna, C. J.;
charge via the Internet at http://pubs.acs.org. Morales, M. P. Controlled Synthesis of Uniform Magnetite
Nanocrystals with High-Quality Properties for Biomedical
REFERENCES AND NOTES Applications. J. Mater. Chem. 2012, 22, 21065–21075.
23. Bronstein, L. M.; Huang, X.; Retrum, J.; Schmucker, A.; Pink,
1. Lu, A. H.; Salabas, E. e. L.; Schüth, F. Magnetic Nanoparticles: M.; Stein, B. D.; Dragnea, B. Influence of Iron Oleate
Synthesis, Protection, Functionalization, and Application. Complex Structure on Iron Oxide Nanoparticle Formation.
Angew. Chem., Int. Ed. 2007, 46, 1222–1244. Chem. Mater. 2007, 19, 3624–3632.
2. Mornet, S.; Vasseur, S.; Grasset, F.; Duguet, E. Magnetic 24. Gheisari, M.; Mozaffari, M.; Acet, M.; Amighian, J. Prepara-
Nanoparticle Design for Medical Diagnosis and Therapy. tion and Investigation of Magnetic Properties of Wüstite
J. Mater. Chem. 2004, 14, 2161–2175. Nanoparticles. J. Magn. Magn. Mater. 2008, 320, 2618–2621.
3. Cullity, B. D.; Graham, C. D. Introduction to Magnetic 25. Bao, N.; Shen, L.; An, W.; Padhan, P.; Heath Turner, C.; Gupta,
Materials; Wiley-IEEE Press: Hoboken, NJ, 2011. A. Formation Mechanism and Shape Control of Mono-
4. Jang, J.; Nah, H.; Lee, J.; Moon, S. H.; Kim, M. G.; Cheon, J. disperse Magnetic CoFe2O4 Nanocrystals. Chem. Mater.
Critical Enhancements of MRI Contrast and Hyperthermic 2009, 21, 3458–3468.
Effects by Dopant-Controlled Magnetic Nanoparticles. 26. Kwon, S. G.; Hyeon, T. Formation Mechanisms of Uniform
Angew. Chem., Int. Ed. 2009, 48, 1234–1238. Nanocrystals via Hot-Injection and Heat-Up Methods.
5. Lee, J.; Jang, J.; Choi, J.; Moon, S. H.; Noh, S.; Kim, J.; Kim, J.; Small 2011, 7, 2685–2702.
Kim, I.; Park, K. I.; Cheon, J. Exchange-Coupled Magnetic 27. Song, Q.; Ding, Y.; Wang, Z. L.; Zhang, Z. J. Tuning the
Nanoparticles for Efficient Heat Induction. Nat. Nanotechnol. Thermal Stability of Molecular Precursors for the Non-
2011, 6, 418–422. hydrolytic Synthesis of Magnetic MnFe2O4 Spinel Nano-
6. Maaz, K.; Mumtaz, A.; Hasanain, S. K.; Ceylan, A. Synthesis crystals. Chem. Mater. 2007, 19, 4633–4638.
and Magnetic Properties of Cobalt Ferrite (CoFe2O4) 28. Xie, J.; Peng, S.; Brower, N.; Pourmand, N.; Wang, S. X.; Sun,
Nanoparticles Prepared by Wet Chemical Route. J. Magn. S. One-Pot Synthesis of Monodisperse Iron Oxide Nano-
Magn. Mater. 2007, 308, 289–295. particles for Potential Biomedical Applications. Pure Appl.
7. Huang, H.; Delikanli, S.; Zeng, H.; Ferkey, D. M.; Pralle, A. Chem. 2006, 78, 1003–1014.
Remote Control of Ion Channels and Neurons through 29. Noh, S.; Na, W.; Jang, J.; Lee, J.; Lee, E. J.; Moon, S. H.; Lim, Y.;
Magnetic-Field Heating of Nanoparticles. Nat. Nanotechnol. Shin, J.; Cheon, J. Nanoscale Magnetism Control via Surface
2010, 5, 602–606. and Exchange Anisotropy for Optimized Ferrimagnetic
8. Stanley, S. A.; Gagner, J. E.; Damanpour, S.; Yoshida, M.; Hysteresis. Nano Lett. 2012, 12, 3716–3721.
Dordick, J. S.; Friedman, J. M. Radio-Wave Heating of Iron 30. Morales, M. P.; Veintemillas-Verdaguer, S.; Montero, M. I.;
Oxide Nanoparticles Can Regulate Plasma Glucose in Mice. Serna, C. J.; Roig, A.; Casas, L.; Martínez, B.; Sandiumenge, F.
Science 2012, 336, 604–608. Surface and Internal Spin Canting in γ-Fe2O3 Nanoparticles.
9. Hergt, R.; Dutz, S. Magnetic Particle Hyperthermia; Chem. Mater. 1999, 11, 3058–3064.
Biophysical Limitations of a Visionary Tumour Therapy. 31. Nogués, J.; Schuller, I. K. Exchange Bias. J. Magn. Magn.
J. Magn. Magn. Mater. 2007, 311, 187–192. Mater. 1999, 192, 203–232.
10. Carrey, J.; Mehdaoui, B.; Respaud, M. Simple Models 32. Nogués, J.; Sort, J.; Langlais, V.; Skumryev, V.; Surinach, S.;
for Dynamic Hysteresis Loop Calculations of Magnetic Munoz, J. S.; Baró, M. D. Exchange Bias in Nanostructures.
Single-Domain Nanoparticles: Application to Magnetic Phys. Rep. 2005, 422, 65–117.
Hyperthermia Optimization. J. Appl. Phys. 2011, 109, No. 33. Levy, M.; Quarta, A.; Espinosa, A.; Figuerola, A.; Wilhelm, C.;
083921. Garcia-Hernandez, M.; Genovese, A.; Falqui, A.; Alloyeau,
11. Mehdaoui, B.; Meffre, A.; Carrey, J.; Lachaize, S.; Lacroix, D.; Buonsanti, R.; et al. Correlating Magneto-Structural
L.-M.; Gougeon, M.; Chaudret, B.; Respaud, M. Optimal Size Properties to Hyperthermia Performance of Highly Mono-
of Nanoparticles for Magnetic Hyperthermia: A Combined disperse Iron Oxide Nanoparticles Prepared by a Seeded-
Theoretical and Experimental Study. Adv. Funct. Mater. Growth Route. Chem. Mater. 2011, 23, 4170–4180.
2011, 21, 4573–4581. 34. Kavich, D. W.; Dickerson, J. H.; Mahajan, S. V.; Hasan, S. A.;
12. Kodama, R. H. Magnetic Nanoparticles. J. Magn. Magn. Park, J. H. Exchange Bias of Singly Inverted FeO/Fe3O4
Mater. 1999, 200, 359–372. Core-Shell Nanocrystals. Phys. Rev. B 2008, 78, 174414.
13. Yoon, S.; Krishnan, K. M. Temperature Dependence of 35. Hou, Y.; Xu, Z.; Sun, S. Controlled Synthesis and Chemical
Magnetic Anisotropy Constant in Manganese Ferrite Conversions of FeO Nanoparticles. Angew. Chem., Int. Ed.
Nanoparticles at Low Temperature. J. Appl. Phys. 2011, 109, 2007, 46, 6329–6332.
No. 07B534. 36. Casula, M. F.; Jun, Y.-w.; Zaziski, D. J.; Chan, E. M.; Corrias, A.;
14. Rosensweig, R. E. Heating Magnetic Fluid with Alternating Alivisatos, A. P. The Concept of Delayed Nucleation in
Magnetic Field. J. Magn. Magn. Mater. 2002, 252, 370–374. Nanocrystal Growth Demonstrated for the Case of Iron
15. Fortin, J.; Wilhelm, C.; Servais, J.; Ménager, C.; Bacri, J.; Oxide Nanodisks. J. Am. Chem. Soc. 2006, 128, 1675–1682.
Gazeau, F. Size-Sorted Anionic Iron Oxide Nanomagnets 37. Xu, Z.; Shen, C.; Hou, Y.; Gao, H.; Sun, S. Oleylamine as Both
as Colloidal Mediators for Magnetic Hyperthermia. J. Am. Reducing Agent and Stabilizer in a Facile Synthesis of
Chem. Soc. 2007, 129, 2628–2635. Magnetite Nanoparticles. Chem. Mater. 2009, 21, 1778–
16. Stoner, E. C.; Wohlfarth, E. P. A Mechanism of Magnetic 1780.
Hysteresis in Heterogeneous Alloys. Philos. Trans. R. Soc., A 38. Zhang, T.; Ge, J.; Hu, Y.; Yin, Y. A General Approach for
1948, 599–642. Transferring Hydrophobic Nanocrystals into Water. Nano
17. Usov, N. A.; Grebenshchikov, Y. B. Hysteresis Loops of Lett. 2007, 7, 3203–3207.
an Assembly of Superparamagnetic Nanoparticles with 39. Xu, Y.; Qin, Y.; Palchoudhury, S.; Bao, Y. Water-Soluble Iron
Uniaxial Anisotropy. J. Appl. Phys. 2009, 106, No. 023917. Oxide Nanoparticles with High Stability and Selective
18. Ibusuki, T.; Kojima, S.; Kitakami, O.; Shimada, Y. Magnetic Surface Functionality. Langmuir 2011, 27, 8990–8997.
Anisotropy and Behaviors of Fe Nanoparticles. IEEE Trans. 40. Guardia, P.; Di Corato, R.; Lartigue, L.; Wilhelm, C.; Espinosa,
Magn. 2001, 37, 2223–2225. A.; Garcia-Hernandez, M.; Gazeau, F.; Manna, L.; Pellegrino,
19. Bao, N.; Shen, L.; Wang, Y.; Padhan, P.; Gupta, A. A Facile T. Water-Soluble Iron Oxide Nanocubes with High Values
Thermolysis Route to Monodisperse Ferrite Nanocrystals. of Specific Absorption Rate for Cancer Cell Hyperthermia
J. Am. Chem. Soc. 2007, 129, 12374–12375. Treatment. ACS Nano 2012, 6, 3080–3091.
20. Sun, S.; Zeng, H.; Robinson, D. B.; Raoux, S.; Rice, P. M.;
Wang, S. X.; Li, G. Monodisperse MFe2O4 (M = Fe, Co, Mn)
Nanoparticles. J. Am. Chem. Soc. 2003, 126, 273–279.

CHEN ET AL. VOL. 7 ’ NO. 10 ’ 8990–9000 ’ 2013 9000

www.acsnano.org