Current Directions in Biomedical Engineering 2017; 3(2): 457–460

Ulrich Engelmann*, Eva Miriam Buhl, Martin Baumann, Thomas Schmitz-Rode and Ioana
Slabu*
Agglomeration of magnetic nanoparticles and
its effects on magnetic hyperthermia
Abstract: Magnetic fluid hyperthermia (MFH) is a
1 Introduction
promising approach for organ-confined tumor treatment. In
MFH, magnetic nanoparticles (MNP) are magnetically
Magnetic fluid hyperthermia (MFH) is one of the most
targeted at the tumor site and heated in an alternating
promising approaches for innovative cancer therapy, which
magnetic field. The heat produced by the MNP is used to
increasingly gained importance due to significant advances in
cause tumor cell death. At the tumor site, MNP bind to the
clinical application in the last decades [1, 2]. In MFH
cell membrane and form agglomerates before they are
biocompatible magnetic nanoparticles (MNP) are
internalized into the intracellular compartments. Intracellular
accumulated at the tumor site, e. g. by magnetic targeting,
immobilization and the formation of agglomerates influence
and exposed to an alternating magnetic field (AMF) in order
heating properties of MNP making it difficult to control the
to generate heat inside the tumor. For temperatures between
local heating inside the tumor. In this study, we investigated
42°C and 45°C tumor cell death is induced [3].
MNP agglomerated samples for their heating efficiency. We
Upon accumulation at the tumor site, MNP are bound
found an increase in heating of 22 % upon agglomeration. If
and grouped in agglomerates at the cell membrane before
MNP are additionally immobilized, however, the heating
they are internalized into cells via endocytosis [4].
decreases by 30 %. Consequently, due to the binding of
Internalized MNP are partly immobilized and tightly packed
bigger MNP agglomerates at cellular level, heating efficiency
inside vesicles [5]. Recently, it was shown that intracellular
inside tumors is assumed to decrease.
MNP generate less heating in MFH processes than free MNP
Keywords: magnetic iron-oxide nanoparticles, magnetic in suspension [6]. This finding is contradicted, however, by
fluid hyperthermia, nanoparticle agglomeration. the evidence that MNP agglomerations cause an amplified
heating [7].
https://doi.org/10.1515/cdbme-2017-0096 Since agglomeration of MNP occurs inevitably in
cellular environments, we aimed at investigating
agglomeration of iron-oxide MNP under controlled
conditions concerning their heating properties. In this study,
MNP agglomeration was achieved by adding of specific
amounts of sodium chloride to MNP solution, forcing the
MNP fatty acid shell to crystalize. The heating of these
______
samples is examined in order to discuss how MNP
*Corresponding authors: Ulrich Engelmann, Ioana Slabu:
Institute of Applied Medical Engineering RWTH Aachen University agglomeration affects heating in cellular environments.
and University Hospital, Pauwelsstr. 20 D-52074 Aachen,
Germany, e-mail: engelmann@ame.rwth-aachen.de (UE),
slabu@ame.rwth-aachen.de (IS)
Martin Baumann, Thomas Schmitz-Rode: Institute of Applied 2 Methods
Medical Engineering RWTH Aachen University and University
Hospital, Pauwelsstr. 20 D-52074 Aachen, Germany, e-mail: Iron-oxide magnetite (Fe3O4) MNP coated with lauric
baumann@ame.rwth-aachen.de (MB), smiro@ame.rwth-
(dodecanoic) acid (LA) were synthesized by co-precipitation
aachen.de (TSR)
as previously described in [8]. The resulting ferrofluid was
Eva Miriam Buhl: Electron Microscopic Facility, RWTH Aachen
University Hospital, Pauwelsstr. 30, D-52074 Aachen, e-mail: centrifuged twice and only the supernatant was retrieved to
ebuhl@ukaachen.de remove large aggregates of MNP and ensure monodispersity.

Unauthenticated
Download Date | 9/10/17 1:48 AM
458

By adding sodium chloride (NaCl; analytical grade,

Merck KGaA, Darmstadt, Germany) to the ferrofluid, the
ionic strength of the solution was increased leading to a
decreased solubility for LA. This process resulted in
crystallization and precipitation of LA [9] and, in this way, a
partial LA removal from the MNP surface was caused
inducing MNP agglomeration. Four different samples were
prepared by adding NaCl dissolved in DI-H2O with
concentrations of 0 M, 0.1 M, 0.25 M and 0.5 M NaCl to the
ferrofluid. Reference samples for each NaCl concentration
containing no MNP were prepared as well.
MNP samples were investigated via transmission
electron microscopy (TEM) at 60 kV using a Zeiss LEO 906
device (Carl Zeiss GmbH, Oberkochen, Germany). Each
sample was diluted to 300 µg(Fe)/mL and vortexed after one
hour rest. Then, 1 µL sample was pipetted carefully on 3 mm
diameter Cu/Rh-150 maxtaform-grids (Electron Microscopy
Sciences, Hatfield, PA, USA) and dried under ambient
conditions for one hour.
Furthermore, the hydrodynamic sizes of the
agglomerates in the MNP samples were characterized with a
dynamic light scattering (DLS) Zetasizer Nano S (Malvern
Instruments Ltd., Worcestershire, UK). For that, MNP
samples were diluted to 300 µg(Fe)/mL and vortexed after
one hour rest. Then, 2 µL from each sample was further
diluted in 500 µL DI-H2O and measured 3 times at 20 °C at a
backscattering angle of 173°. The average hydrodynamic
diameter (Z-Average) was determined from triplicate
measurements of three fractions of the same sample.
Heating of MNP was assessed using a custom-build
hyperthermia setup (Trumpf Hüttinger, Freiburg, Germany)
consisting of a DC generator, an AC-resonant oscillator and a
water-cooled copper coil (inner / outer diameter:
(20 / 30) mm; 8 turns). The cooling water (30 L) was kept in
a sealed circulatory system constantly held at (17 ± 1) °C by
a separate water cooling system. MNP samples of 1 mL were
prepared in 4 mL glass vials and heated to an initial
Figure 1: TEM images of either pure lauric acid (left row) or lauric
temperature of 35 °C on a hot plate. Following, each sample acid coated MNP (right row) dispersed in solutions with increasing
was inserted into the center of the coil and exposed to an NaCl concentrations: A,B: no NaCl; C,D: 0.1 M NaCl; E,F: 0.25 M
NaCl. G,H: 0.5 M NaCl. Images E to H show lauric acid as gray
AMF of frequency (265 ± 3) kHz and field amplitude
filaments.
(36.8 ±0.6) kA/m for 30 minutes. During AMF exposure, the
. The initial temperature rise is dependent on the fit
sample temperature was recorded with a fibre-optic
parameters Trise and b of the Box-Lucas function
thermometer Luxtron 812 (LumaSense Inc., Santa Barbara,
fitted to the acquired
CA, USA). A background subtraction was performed using
temperature data [10].
the data from the references measurements.
Three different iron concentrations of (150, 225,
The specific loss power (SLP), which is a measure of the
300) µg(Fe)/mL were measured. The lowest concentration
heating efficiency was calculated using the expression:
corresponds approx. to the iron concentration in the whole
, with the specific heat capacity of the
extracellular space after an injection of 42 ml Resovist to the
carrier liquid (= 4.187 J/g/°C for water), the weight fraction
human body [11]. Each sample was prepared and measured
of MNP and the initial temperature rise

Unauthenticated
Download Date | 9/10/17 1:48 AM
in triplicate, from which an average SLP value was several hundreds of nanometers in diameter formed upon
calculated. adding 0.1 M NaCl (Figure 1D). Figure 1F&H indicate MNP
agglomerates aligned along chains which was more
pronounced for samples with higher NaCl concentrations.
These chains consist of MNP agglomerates that form along
3 Results
filament-like structures (see inset in Figure 1F) causing an
immobilization of MNP in a LA matrix for the sample with
0.5 M NaCl (Figure 1H). Such structures were also observed
3.1 Characterization of MNP samples
for the pure LA samples containing no MNP having high
The results from the TEM measurements are depicted in NaCl concentration (Figure 1E&G).
Figure 1: Left row displays pure LA, while the right row Symmetrically shaped crystals visible in the reference
shows with LA coated MNP. When no NaCl was added, with 0.1 M NaCl (Figure 1C) and white “tree-like” branching
MNP were monodispersely distributed showing no structures observed in the reference with 0.25 M NaCl
agglomerates (Figure 1B). However, large agglomerates of (Figure 1E) were attributed to formation of NaCl crystals
upon drying (TEM reference images for pure NaCl not
shown).
0.0 M NaCl DLS measurements displayed in Figure 2 yielded an
0.1 M NaCl average hydrodynamic diameter of (91 ± 7) nm for untreated
0.25 M NaCl
100 MNP, which decreased with increasing NaCl concentration.
Hydrodynamic Diameter [nm]

0.5 M NaCl

90
3.2 Heating of MNP samples
The results of SLP determination are summarized in
Figure 3. Upon increasing NaCl concentration the SLP value
80 increased. This observation is valid independent of iron
concentration. The SLP increase for samples treated with up
to 0.25 M NaCl in comparison to untreated MNP (0.0 M) was
22.0 % for 150 µg(Fe)/mL, 20.0 % for 225 µg(Fe)/mL and
70
0.0 0.1 0.2 0.3 0.4 0.5 21.5 % for 300 µg(Fe)/mL.
NaCl concentration [M] However, this trend is not continuted for the samples treated
Figure 2: Mean hydrodynamic diameter (Z-Average) versus with the highest NaCl fraction of 0.5 M. For this case, the
NaCl concentration. SLP value decreased significantly compared to untreated
MNP by 22.8 % for 150 µg(Fe)/mL, 26.2 % for
225 µg(Fe)/mL and 30.0 % for 300 µg(Fe)/mL.
600 0.0 M NaCl
0.1 M NaCl
0.25 M NaCl
500 0.5 M NaCl 4 Discussion and conclusion
mean SLP [W/g(Fe)]

400 MNP agglomeration was reached by adding NaCl to the

MNP solution promoting crystallization of their fatty acid
300 shells. Qualitative assessment of MNP agglomeration size
and morphology via TEM imaging shows clear differences
200 among the samples in the MNP arrangement as well as in the
crystallization of LA upon adding NaCl (Figure 1). These
100
results are in line with literature findings which describe an
increase in the hydrodynamic diameter upon agglomeration
0
150 225 300 [12]. However, the DLS measurements show a decrease for
the MNP size with increasing NaCl concentration (Figure 2).
Concentration [µg(Fe)/mL]
We assume that bigger MNP agglomerates could not be
Figure 3: Mean SLP values versus MNP sample iron
concentration for non-agglomerated particles (0.0 M NaCl) up to
fully agglomerated MNP samples.

Unauthenticated
Download Date | 9/10/17 1:48 AM
460
assessed by DLS for two possible reasons: First, the
temperature-equilibrating time before each DLS run is
approx. 2 min, allowing bigger agglomerates to sediment and
which then do not contribute to the measured MNP. Second,
a filtering effect occurred at the pipette nozzle, when
extracting 2 µL during sample preparation. Consequently,
only remaining non-agglomerated monodispersed MNP (cf.
error bars in Figure 2) were measured with DLS.
Our observations of a clear increase in the SLP upon
agglomeration in the 0.1 M NaCl sample is in line with the
findings in [7]. With increasing agglomeration the particle-
particle distance is much shorter, which strongly increases
magnetic dipole interparticle interactions. This in turn leads
to a more pronounced ferromagnetic behaviour with
hysteresis losses that contribute to the heating process [13].
We observe an even stronger increase in SLP upon
further agglomeration combined with particle chain
formation along filaments for sample 0.25 M NaCl. Saville et
al. propose to treat MNP chains as a single magnetic rod with
ferromagnetic properties along the direction of particle
alignment [14]. This gives again rise to (ferromagnetic)
hysteresis losses, increasing the heating as described above.
The SLP results of sample 0.5 M NaCl (Figure 3)
exemplarily demonstrates the interplay of agglomeration and
immobilization effects concerning the heating efficiency: a
significant decrease in SLP of up to 30 % was observed -
despite agglomeration but for simultaneous immobilization
of MNP (Figure 1H). We attribute this decrease in SLP to a
partial blocking of Brownian relaxation caused by
immobilization of the MNP agglomerates. This phenomenon
was also reported in literature, where a decrease in SLP of
approx. 20 % was found upon immobilization [15].
Investigation with other measurement techniques, such as AC
magnetic susceptibility, confirm a reduction of Brownian
relaxation for MNP immobilized inside cells [16].
In conclusion, when considering agglomeration and
immobilization of MNP upon intracellular uptake as
competing effects on MNP heating, the SLP value seems to
be dominated by immobilization. This causes a decrease in
the heating efficiency under the condition that the
immobilization is sufficiently strong to block the Brownian
relaxation of the agglomerates.
Author’s Statement
Research funding: The author UE greatly appreciates funding
by the Cusanuswerk e.V. Conflict of interest: Authors state
no conflict of interest. Informed consent: Informed consent is
not applicable. Ethical approval: The conducted research is
not related to either human or animals use.
References
[1] Johannsen M, Gneveckow U, et al. Thermotherapy of
prostate cancer using magnetic nanoparticles: Feasibility,
imaging, and three-dimensional temperature distribution:
European Urology. 2007; 52:1653–1662.
[2] Wust P, Gneveckow U, et al. Magnetic nanoparticles for
interstitial thermotherapy - feasibility, tolerance and achieved
temperatures: International Journal of Hyperthermia. Int J
Hyperther. 2006; 22:673–685.
[3] Latorre M, Rinaldi C. Applications of Magnetic Nanoparticles
in Medicine: Magnetic Fluid Hyperthermia. Puerto Rico
Health Sci. J. 2009; 28:227–238.
[4] Wilhelm C, Gazeau F. Universal cell labelling with anionic
magnetic nanoparticles. Biomaterials. 2008; 29:3161–3174.
[5] Levy M, Wilhelm C, et al. Nanomagnetism reveals the
intracellular clustering of iron oxide nanoparticles in the
organism. Nanoscale. 2011; 3:4402-4410.
[6] Di Corato R, Espinosa A, et al. Magnetic hyperthermia
efficiency in the cellular environment for different
nanoparticle designs. Biomaterials. 2014; 35:6400-6411.
[7] Piñeiro-Redondo Y, Banobre-Lopez M, et al. The influence
of colloidal parameters on the specific power absorption of
PAA-coated magnetite nanoparticles: Nanoscale Research
Letters. 2011; 6.
[8] Slabu I, Guntherodt G, et al. Investigation of
Superparamagnetic Iron Oxide Nanoparticles for MR-
Visualization of Surgical Implants. Curr Pharm Biotechno.
2012; 13:545-551.
[9] Genck M. Make The Most of Antisolvent Crystallization.
Chemical Processing. 2010.
[10] Wildeboer RR, Southern P, et al. On the reliable
measurement of specific absorption rates and intrinsic loss
parameters in magnetic hyperthermia materials: Journal of
Physics D-Applied Physics. 2014; 47.
[11] Slabu I. Synthesis, Characterization and Application of
Superparamagnetic Iron Oxide Nanoparticles in Medical
Diagnostics and Therapy: MR-Visible Implants for Hernia
Repair and Novel Drug Targeting Models [Doctoral Thesis].
Shaker Verlag Aachen. 2015.
[12] Hondow N, Brydson R, et al. Quantitative characterization of
nanoparticle agglomeration within biological media. J
Nanopart Res. 2012; 14.
[13] Dennis CL, Jackson AJ, et al. Nearly complete regression of
tumors via collective behavior of magnetic nanoparticles in
hyperthermia: Nanotechnology. Nanotechnology. 2009; 20.
[14] Saville SL, Qi B, et al. The formation of linear aggregates in
magnetic hyperthermia: Implications on specific absorption
rate and magnetic anisotropy: Journal of Colloid and
Interface Science. J Colloid Interface Sci. 2014; 424:141–
151.
[15] Ludwig R, Stapf M, et al. Structural properties of magnetic
nanoparticles determine their heating behavior - an
estimation of the in vivo heating potential: Nanoscale
Research Letters. 2014; 9.
[16] Soukup D, Moise S, et al. In situ measurement of
magnetization relaxation of internalized nanoparticles in live
cells: ACS nano. 2015; 9:231–240
Unauthenticated
Download Date | 9/10/17 1:48 AM