Adjusting the Néel relaxation time ofFe3O4/ZnxCo1−xFe2O4 core/shellnanoparticles for optimal heat generation inmagnetic hyperthermia

Fernando Fabris1, Javier Lohr1,2, Enio Lima Jr1,Adriele Aparecida de Almeida1, Horacio E Troiani3, Luis M Rodríguez1,Marcelo Vásquez Mansilla1, Myriam H Aguirre4,5 , Gerardo F Goya5,Daniele Rinaldi6, Alberto Ghirri7, Davide Peddis8, Dino Fiorani8,Roberto D Zysler1,9 , Emilio De Biasi1,9,∗ and Elin L Winkler1,9,∗

1 Instituto de Nanociencia y Nanotecnología CNEA-CONICET—Centro Atómico Bariloche, S. C. deBariloche, 8400, Argentina2 Laboratorio Argentino de Haces de Neutrones—CNEA—Argentina3 Laboratorio de Caracterización de Materiales y Óxidos No-Estequiométricos, Gerencia de InvestigaciónAplicada, Centro Atómico Bariloche, S. C. de Bariloche, 8400, Argentina4 Instituto de Ciencias de Materiales de Aragón & Laboratorio de Microscopías Avanzadas, Universidad deZaragoza, Mariano Esquillor s/n, Zaragoza, E-50018, Spain5 Instituto de Nanociencias y Materiales de Aragón & Dep. Física de la Materia Condensada, Universidadde Zaragoza, Mariano Esquillor s/n, Zaragoza, E-50018, Spain6Department of Materials, Environmental Sciences and Urban Planning (SIMAU), Universita ̀ Politecnicadelle Marche, I-60131 Ancona, Italy7 Istituto Nanoscienze, CNR, via Campi 213/a, I-41125 Modena, Italy8 Istituto di Struttura della Materia, CNR, Area della Ricerca di Roma 1, C.P. 10, I-00015 MonterotondoStazione, Rome, Italy9 Instituto Balseiro-Universidad Nacional de Cuyo, Argentina

E-mail: debiasi@cab.cnea.gov.ar and winkler@cab.cnea.gov.ar

Received 3 August 2020, revised 7 October 2020Accepted for publication 21 October 2020Published 18 November 2020

AbstractIn this work it is shown a precise way to optimize the heat generation in high viscosity magneticcolloids, by adjusting the Néel relaxation time in core/shell bimagnetic nanoparticles, for magnetic fluidhyperthermia (MFH) applications. To pursue this goal, Fe3O4/ZnxCo1−xFe2O4 core/shell nanoparticleswere synthesized with 8.5 nm mean core diameter, encapsulated in a shell of ∼1.1 nm of thickness,where the Zn atomic ratio (Zn/(Zn+Co) at%) changes from 33 to 68 at%. The magneticmeasurements are consistent with a rigid interface coupling between the core and shell phases, wherethe effective magnetic anisotropy systematically decreases when the Zn concentration increases, withouta significant change of the saturation magnetization. Experiments of MFH of 0.1 wt% of these particlesdispersed in water, in Dulbecco modified Eagles minimal essential medium, and a high viscosity butteroil, result in a large specific loss power (SLP), up to 150Wg−1, when the experiments are performed at571 kHz and 200 Oe. The SLP was optimized adjusting the shell composition, showing a maximum forintermediate Zn concentration. This study shows a way to maximize the heat generation in viscousmedia like cytosol, for those biomedical applications that require smaller particle sizes.

Supplementary material for this article is available online

Nanotechnology

Nanotechnology 32 (2021) 065703 (11pp) https://doi.org/10.1088/1361-6528/abc386

∗ Authors to whom any correspondence should be addressed.

0957-4484/21/065703+11$33.00 © 2020 IOP Publishing Ltd Printed in the UK1

Keywords: core/shell nanoparticles, magnetic fluid hyperthermia, Néel relaxation time

(Some figures may appear in colour only in the online journal)

1. Introduction

Magnetic fluid hyperthermia (MFH) is a promising techniquefor new cancer therapies based on the use of magneticnanomaterials. In its simplest form, magnetic nanoparticles(MNPs) provoke the death of tumor cells by increasing thelocal temperature, therefore improving the cytotoxic effects ofradiotherapy and chemotherapy [1–3]. The heating of targetedtumors in MFH is obtained through the magnetic losses ofMNPs exposed to alternating magnetic fields. The magneticlosses are originated from the phase shift between the nano-particle’s magnetic moment and the applied AC magneticfield, and the magnetic relaxation dynamics depends on therelaxation time of two concurrent mechanisms [4]. One ofthem, called mechanical or Brown mechanism, is the physicalrotation of the MNPs characterized by the time relaxation τB,that is given by:

th

=V

k T

3, 1B

hyd

B( )

where η is the viscosity of the medium, Vhyd the hydro-dynamic volume of the MNPs and kBT the thermal energy.The second mechanism involves the inversion of magneticmoment within the crystal lattice and is called magnetic orNéel mechanism. The Néel mechanism is well described fornon-interacting and monodomain MNPs by the Néel relaxa-tion time, τN, given by:

t t= K V k Texp , 2N 0 eff mag B( ) ( )/

where Keff and Vmag are the effective anisotropy and magneticvolume of the single-domain NPs, respectively, and τ0 is theattempt relaxation time of the system (typically between 10−9

and 10−11 s) [5]. In any given situation, the dominatingrelaxation process will be the one with the shorter relaxationtime. For systems with large magnetic anisotropy energy and/or low viscosity, the Brown mechanism dominates, whereasNéel mechanism governs in the opposite situation [6].

Despite the promising results, MFH has not been yetincorporated into current oncological protocols mainly due tothree factors: (i) the low values of specific loss power (SLP)attained so far, (ii) the cytotoxicity effects of the MNPs and(iii) the large variability of SLP values even for differentbatches of colloids synthesized by the same methods. The lowSLP values result in the need of high local concentration ofMNPs, which increases the cytotoxicity and reduces thespecificity of the therapy due to undesired apoptosis of sur-rounding healthy tissues [7, 8]. Considerable efforts havebeen invested to develop more efficient nano-heaters, andthus current experimental research in the area of MFH isfocused on the improvement of the heating power through thecontrol of the morphological [9–13], magnetic [14–18], sur-face chemistry [19] and spatial-assembling [20–22] propertiesof the MNPs. Regarding the second factor, i.e. the lack ofreproducibility of MNPs heating properties, the challenge is

related to the variability of SLP when experiments are per-formed in different liquid media (solution, cell culture orin vivo), due to the influence of the different liquid viscositieson the SLP. The high viscosity of the cellular environment isexpected to hinder/block Brown relaxation and, as a con-sequence, the heat generation is usually reduced in systemswhere this mechanism is dominant [15, 23–29].

Therefore, it is crucial to tune individual parameters thatcontrol the magnetic relaxation mechanisms in order to opt-imize the SLP in a medium that simulates the viscosity of thecellular environment. A very effective way to obtain this finecontrol on the magnetic relaxation in MNPs is the design andfabrication of bi-magnetic rigidly coupled core/shell nano-particles composed of hard and soft magnetic materials. Themagnetic exchange coupling at the interface between the coreand shell results in a system with effective anisotropy dif-ferent from those of both phases [30–36]. Thus, it is possibleto adjust precisely the magnetic anisotropy of the soft-hardcore–shell NPs by changing the composition of one of the twophases and preserving the size, morphology and high valuesof magnetization [37].

In a previous work, we have shown that is possible toselect the dominant magnetic relaxation mechanism of heatgeneration between Brown or Néel ones by changing thecomposition of the shell in Fe3O4/ZnxCo1−xFe2O4 NPs [38].In this work, we selected the Néel mechanism in viscousmedia, due to its relevance in biological applications, and wemaximize the heat generation by tuning the composition ofthe shell of the MNPs. We showed that by changing the zincfraction of the ZnxCo1−xFe2O4, the magnetic anisotropy andtherefore the magnetic relaxation can be tuned, resulting inlarger values of SLP. We explored highly viscous mediaincluding butter oil, water and Dulbecco modified Eaglesminimal essential medium (DMEM). These results show thepotential of core/shell systems for the development ofnanoparticles with specific characteristics for cellular envir-onmental conditions.

2. Material and methods

2.1. Synthesis of Fe3O4 core

The magnetic core/shell nanoparticles were synthesized byseed-mediated thermal decomposition of organo-metallic pre-cursors at high temperature based in the literature [37, 38].Firstly, Fe3O4 core is synthesized from Fe(III) acetylacetonate(12 mMol) in presence of 1,2-octanediol (24 mMol), oleicacid (12 mMol), Oleylamine (30 mMol) and Benzyl ether(190 mMol) as solvent. This solution is heated at 473 K during20min under N2 flow (0.1 ml min−1) and intense mechanicalstirring. After that, the solution was heated until the refluxcondition (563 K) with a heating rate of 15 Kmin−1. Thesolution was kept in reflux during 60min.

2

Nanotechnology 32 (2021) 065703 F Fabris et al

2.2. Synthesis of Fe3O4/ZnxCo1−xFe2O4 MNPs

A fraction of 8.2 ml of the solution obtained from the coresynthesis was mixed in 220 mMol of Benzyl ether, 4.5 mMolof 1,2-octanediol, 3 mMol of oleic acid, 3 mMol of Oleyla-mine in the presence of 1.2 mMol of Fe(III) acetylacetonate,x×0.6 mMol of Zn(II) acetylacetonate and (1− x)×0.6 mMol of Co(II) acetylacetonate. The synthesis procedureused was the same used for the core. In total, 6 samples withthe values of x=0.40, 0.50, 0.60, 0.68, 0.75 and 0.82 wereproduced and labeled according: Zn40, Zn50, Zn60, Zn68,Zn75 and Zn82, respectively. After the synthesis, the nano-particles were precipitated by adding 8 times in volume of asolution containing ethanol and acetone (4:1) followed bycentrifugation (14 000 rpm during 30 min). Finally, the oleic-acid coated hydrophobic samples in powder form were dis-persed in chloroform or toluene.

2.3. Particle induced x-ray emission (PIXE)

The overall chemical composition of the samples was deter-mined by PIXE measurements. They were performed with a3 MeV H+ beam in a NEC 5SDH 1.7 MV tandem acceleratorwith a NEC RC43 end-station for material analysis. In thisexperiment the samples, in powder form, were attached to thesample holder using carbon tape. To obtain the final results,PIXE spectra were processed using GUPIX software.

2.4. Transmission electron microscopy (TEM)

TEM and high resolution TEM (HRTEM) images areobtained in a Philips CM200 transmission electron micro-scope equipped with an Ultra-Twin lens operating at 200 kVand by a FEI-Tecnai F30 at 300 kV field emission gun fittedwith a SuperTwin® lens allowing a point resolution of 1.9 Å.TEM samples were prepared by dropping a chloroformsolution containing the nanoparticles on a copper grid with anultrathin hollow carbon film. The diameter (D) of each par-ticle was obtained by measuring the area of transmissionimage and calculating a diameter supposing a circle. Then, thediameter distribution of each sample was obtained measuringabout 500 particles. The mean diameter ⟨D⟩ and size dis-persion σ were obtained by fitting the size distribution with alognormal distribution.

2.5. Magnetic measurements

Magnetization measurements as function of temperature infield-cooling and in zero-field-cooling modes MFC(T) andMZFC(T), respectively, with using a field of 50 Oe, as well asfunction of applied field, were performed in a SQUID mag-netometer (MPMS 5000 from Quantum Design). For this, theparticles in toluene solution were mixed with epoxy resin andwaited until completely solvent evaporation before the resincuring. The particle concentration in epoxy resin were about0.2 wt% after subtraction the organic mass fraction estimatedby the powder thermogravimetric measurements. AC magn-etic measurements were carried out by means of QuantumDesign PPMS P500 AC/DC magnetometry system based on

the mutual-inductance technique connected to a PPMS elec-tronics system. The real (in-phase’) and imaginary (out-of-phase χ″) components of the AC magnetic susceptibility weremeasured as a function of temperature. Real and imaginaryparts were measured with high accuracy by using a calibrationcoil array. Susceptibility was measured in the temperaturerange T=100–350 K, varying the frequency in the 10 Hz to10 kHz range with a driving field H=14 Oe.

2.6. Hydrophilic MNPs

For MFH experiments performed in water and DMEM, thehydrophobic character of the as-made nanoparticles (oleicacid-coated) is changed to hydrophilic one. For this, the NPsare washed with methanol during 6 h and subsequently in hotacetone (313 K) during 40 h. After that, MNPs were pre-cipitated by using a magnet and 10 mg of these washed NPsare added to 1.5 ml of 27% Ammonium Hydroxide solution(pH 11) containing anhydrous glucose (1000 mg). The NPsremain in this solution at 313 K during 20 h, exposed toultrasound for some intervals. Finally the nanoparticles werewashed with water and diluted in a concentration of 0.1 wt%in water or DMEM.

2.7. Magnetic fluid hyperthermia (MFH)

MFH experiments were performed in D5-F1 rf generator (nB—NanoScale Biomagnetics, Spain). The AC applied field hasamplitude of H0=200 Oe and frequency of f=571 kHz.Initial temperature was the room temperature (about 292–294 K). To perform the MFH the particle in chloroform weredispersed in butter oil by intense sonication at 60 °C until thecompletely chloroform evaporation, while for the MFHexperiments performed in water and DMEM, the hydrophilicones are used. MNPs concentration of about 0.1 wt% in allmatrixes were produced and confirmed by magnetic momentmeasurements. For this, the saturation magnetization of eachnanoparticle was measured by using about 10 mg of powder.The amount of organic surfactants and solvent residues ineach sample was taken into account from TGA measure-ments. After that, the concentration of the nanoparticles in thesamples used in the hyperthermia experiments was calculatedby comparing the saturation magnetization with the respectivemagnetization curve. The main factor of error in the SLPdetermination are the numerical analyzes of the experimentaldata and the power loss due to no-adiabatic condition of thehyperthermia experiment. In order to reduce these effects, adouble-wall vessel with a vacuum zone between them wasused. Moreover, the power loss is proportional to the differ-ence between the room temperature and the sample temper-ature (ΔT), with an exponential dependence with the time[39, 40]. In this way, the sample was thermalized in thehyperthermia equipment for several minutes before to applythe magnetic field, and then the experiment was conducteduntil an increment of temperature (ΔT) not higher than 4 K isreached. Finally, the slope of ΔT/t used to calculate the SLPwas obtained from the linear fit in the region of ΔT<1 K forall samples.

3

Nanotechnology 32 (2021) 065703 F Fabris et al

3. Results and discussion

Representatives image of TEM in bright field mode and sizeshistograms of the Fe3O4 core and Fe3O4/ZnxCo1−xFe2O4

core/shell samples are presented in the figure 1 and in figuresS1 and S2 is available online at stacks.iop.org/NANO/32/065703/mmedia of the supplementary material. A differentmorphology is observed when comparing the images of corewith core/shell samples. Although the lattice parameters,density and atomic weight of the Fe3O4 and ZnxCo1−xFe2O4

spinel phases are very similar, the shell overgrown poly-crystalline, as observed from HRTEM images of Zn50nanoparticles shown in figures 1 and S1. The different crystalstructure between Fe3O4 and ZnxCo1−xFe2O4 ferrites NPs can

be better analyzed by fast Fourier transform (FFT). Differentinterplanar distances of both compounds generate a mismatchstrain at the interface that can be seen by inverse FFT,selecting suitable diffraction poles as is showed in thefigures 1(g)–(i). In this example, the inverse FFT is taken of(022) and (113) diffraction planes which generates an imagewith information of these planes in the real space, reducingthe background and eliminating the non selected crystallineplanes. The procedure reveals a shell with similar crystallineorientation to the core but with a crystalline disorder layer atthe interface due to the mismatch between different ferrites.From the size histogram fitting with a lognormal distribution,a mean core diameter ⟨DCore⟩=8.5 nm and standard devia-tion of σcore=0.2 is obtained. After the shell overgrowth, the

Figure 1. (a) and (b) TEM images of the core nanoparticles and (d) and (e) of the core/shell Zn50 sample. (c) and (f) are the size histogramsof the core and Zn50, respectively. In (g) is presented HRTEM image where a FFT is performed of the red rectangle zone (h) followed by aninverse FFT with the red circle marks (i).

4

Nanotechnology 32 (2021) 065703 F Fabris et al

total size of particles increases, as summarized in the table 1.The average core/shell nanoparticles diameter grows up to10.7 nm, indicating a shell thickness of 1.1 nm.

The analysis of the overall composition of the nano-particles was performed by PIXE technique and the results areshown in table 1. The measured atomic ratio between Zn andCo is lower than the nominal, but present a systematic evolution

with it, changing from =+

0.33Zn

Co Zn⎡⎣ ⎤⎦ to =

+0.68,Zn

Co Zn⎡⎣ ⎤⎦

for Zn40 and Zn80 MNPs systems, respectively. From this data,and assuming a stoichiometric Fe3O4 core with mean size of8.5 nm and the mean total size of the core/shell particles, wecalculated the shell stoichiometry shown in the table 1.

Magnetization loops measured at 5 K are presented infigure 2. From these measurements it is noteworthy that thecoercive field (HC) of the core/shell samples decreasesmonotonously with the nominal zinc fraction in the shell, asshown in the inset of the figure and summarized in table 2.This result indicates a decrease of the effective magneticanisotropy Keff of the core/shell samples with the Zn fraction,as was already found in similar nanomaterials [32, 33]. Themagnetization inversion process of Fe3O4/ZnxCo1−xFe2O4

core/shell nanoparticles can be analyzed from the theoreticalapproach developed for bi-magnetic soft/hard exchange-coupled nanostructures, assuming a rigid magnetic couplingat the interface between the core and the shell; within this

Table 1. Average particle size ⟨D⟩ and distribution width σ from the fit of a lognormal size distribution of the TEM data. Chemicalcomposition obtained by PIXE for the core and the Fe3O4/ZnxCo1−xFe2O4 core/shell the samples. The shell stoichiometry was calculatedassuming the ⟨D⟩ values, and Fe3O4 cores of ⟨D⟩=8.5 nm for all samples. The last column shows the magnetic activation volume Vmag

obtained from relaxation magnetization measurements.

TEM PIXE

Sample ⟨D⟩ (nm) σ Fe (%) Co (%) Zn (%)+Zn

Co Zn(at%) Calculated shell stoichiometry Vmag (nm)

Core 8.5 0.20 — — — — — 9Zn40 10.4 0.17 90.02(48) 6.69(24) 3.28(24) 33(5) Zn0.19Co0.39Fe2.41O4 10.5Zn50 10.9 0.19 89.99(56) 5.96(27) 4.05(30) 40(8) Zn0.21Co0.31Fe2.48O4 10.4Zn60 10.5 0.17 88.76(39) 5.55(19) 5.74(27) 51(8) Zn0.33Co0.31Fe2.36O4 11.4Zn68 10.7 0.14 90.10(41) 4.19(19) 5.71(29) 58(9) Zn0.31Co0.23Fe2.46O4 12.0Zn75 10.9 0.19 92.54(50) 2.92(21) 4.53(29) 61(11) Zn0.24Co0.16Fe2.60O4 13.0Zn82 11.0 0.20 90.66(40) 2.97(16) 6.37(27) 68(10) Zn0.33Co0.15Fe2.52O4 10.3

Figure 2. Hysteresis cycles of all samples measured at 5 K. In the inset the HC (black dots) and MS (red triangles) as function of theZn/(Zn+Co) at% obtained by PIXE are shown. The HC and MS core values are also presented with empty stars.

5

Nanotechnology 32 (2021) 065703 F Fabris et al

approximation the coercive field results [30, 31, 33, 41–43]:

r rr r

=+

+H

K K

M M2 , 3C

cC

sc

C s

eff effSh

ShSh

Sh

( )

where MS is the magnetization saturation of each phase(c=core and sh=shell), ρ is the volume fraction and Keff isthe effective magnetic anisotropy of the core (c) and shell (sh).According to this model the HC can be estimated for a core/shell particles from the core parameters MS=90 emu g−1,⟨DCore⟩=8.5 nm, K c

eff=3×105 erg cm−3 [28], and shellparameters as thickness of 1 nm, Keff

sh =6×106 erg cm−3 or5×104 erg cm−3 for the CoFe2O4 [44] and ZnFe2O4 [45]phases, respectively; resulting in HC=6.4 kOe and HC=0.4 kOe, respectively. These estimated values are in goodagreement with the values experimentally obtained, supportingthe rigid coupling between the core and shell where the effec-tive magnetic anisotropy decreases with the Zn concentration ofthe shell.

It can be noticed in figure 2 (see also table 2) that the MS

of all samples does not change significantly, between 87 and99 emu g−1, when the shell stoichiometry changes, consistentwith reported in previous works [16, 38].

At room temperature all the samples present reversiblebehavior when the magnetization loops are measured with DCfield, as observed from figure S3, signaling that the nano-particles systems are superparamagnetic at this temperature. Inorder to obtain the blocking temperature (TB) of the systems, themagnetization curves were measured as function of temperaturein zero-field-cooling MZFC(T) and field-cooling MFC(T) proto-cols, and on samples with 0.2 wt% dispersed in epoxy resin, asshown in figure 3. For all core/shell samples, the maxima ofMZFC(T) curves (and the irreversible point) are located at highertemperatures than the corresponding nude magnetic core,indicating higher blocking temperatures associated to theZnxCo1−xFe2O4 shell. Figure 3 also shows the distribution ofblocking temperature ( f (TB)) calculated from the magnetizationcurves for an assembly of non-interacting (or weakly interacting

particles) as [46–49]:

µ-

f TT

M T M T

T

1 d

d. 4B

ZFC FC( ) ( ( ) ( )) ( )

The mean blocking temperature (⟨TB⟩) is shown in table 2for all samples. It can be noticed that ⟨TB⟩ decreases as the Znfraction increases, indicating a reduction of the energy barrierfor the magnetic moment thermal fluctuation. In a system ofnon-interacting particles with uniaxial anisotropy, the energybarrier is defined as KeffVmag and the corresponding blocking

Table 2. Parameters obtained from the magnetic measurements. Themean blocking temperature (á ñTB ) was obtained from the arithmeticmean of the distribution as òá ñ =T T P T Td ,B ( ) and distribution width

(sTB) was defined as variance, given by, òs = á ñ -T T P T Td ,T2

B2

B ( ) ( )where =

òP T f T

f T TB dB( ) ( )

( )is the probability density.

TB

HC (5 K) MS á ñTB sTB Keff

Sample kOe emu g−1 K K ×105 erg cm−3

Core 0.4(0.1) 80(4) 28.5 29.5 3.0(1.7)Zn40 5.4(0.4) 97(4) 164.8 75.2 10.9(2.5)Zn50 4.4(0.3) 99(3) 127.1 62.0 8.7(3.0)Zn60 3.7(0.3) 97(4) 153.5 73.5 8.3(2.6)Zn68 3.3(0.3) 87(3) 130.3 57.9 5.8(2.4)Zn75 2.0(0.2) 89(3) 106.8 62.2 3.7(3.0)Zn82 1.8(0.2) 99(3) 99.3 52.9 6.9(2.5)

Figure 3. Temperature dependence of the MZFC(T) and MFC(T)curves of the core and Fe3O4/ZnxCo1−xFe2O4 NPs with 0.2 wt% inEpoxy resin and measured under applied field of 50 Oe. The redtriangles is the distributions of blocked temperature f (TB) for eachsample calculated from equation (4).

6

Nanotechnology 32 (2021) 065703 F Fabris et al

temperature is:

tt

á ñ =Tk

K V1

ln , 5BB

0

meff mag ( )

⎛⎝⎜

⎞⎠⎟

where Vmag is the magnetic volume, and tm is the measurementtime (e.g. tm∼100 s for DC magnetometry measurement andτm∼1/f for AC magnetometry and hyperthermia experiment).To extract reliable values of Keff from the above relation, weperformed further relaxation magnetization and ac-susceptibilitymeasurements experiments to accurately determine Vmag and τ0of the present core/shell MNPs systems.

The magnetic volume was determined from the depend-ence of magnetic viscosity with respect to an external field atlow temperature (T=5 K). In a typical experiment, thesample is brought to a negative saturation field; then a reversefield is applied and the time dependence of magnetization ismeasured. This experiment is repeated for several reversefields, around the coercivity value, in order to determine themaximum value of magnetic viscosity (Smax), as shown infigure S4. From Smax it is possible to obtain the so-calledfluctuation field, Hf, which represents an effective field withthe same effect on magnetization than thermal fluctuations[50]. With these parameters the magnetic activation volumeobtained from Vmag=kBT/MsHf represents the smallestvolume of material that reverses coherently in an event[50, 51]. The values of Vmag for all samples are reported intable 1. Details about the calculation of Vmag are provided insupplementary material. Remarkably, the obtained values ofVmag are in excellent agreement with the nanoparticle volumemeasured by TEM (table 1), as expected if the core and shellmagnetic phases were rigidly coupled and therefore themagnetic moment undergo a coherent reversion against theenergy barriers of the system [50, 52, 53].

For all the samples the temperature dependence of χ″

shows a broad maximum that shift toward lower temperaturewhen the frequency decreases. On the other hand, χ′ presentsa monotonous raising behavior in almost all samples exceptfor the system with largest Zn concentration, Zn82, wherealso a frequency dependent peak is noticed (figure S5 sup-plemental material). For this sample, the frequency depend-ence of the χ″ peaks is successfully fitted with thephenomenological Vogel–Fulcher law as shown in figure S5[54], resulted in τ0=2(1) 10−11 s. When the Zn concentra-tion decreases, the fitting of the frequency dependence peaksgives non-physical values of τ0, in the 10−5

–10−7 s range.This indicates that these exchange coupled core/shell systemswith larger magnetic anisotropy are outside of the applic-ability of the model.

From the experimental parameters ⟨TB⟩, Vmag, and usingτ0=2(1) 10−11 s, obtained from the ZFC and FC magneti-zation, relaxation and ac-susceptibility measurements,respectively we have calculated the effective magneticanisotropy for all the samples. The results are reported intable 2 and in figure 4. As noticed, the Keff decreases when theconcentration of Zn increases.

Figure 5 shows the results of MFH experiments of the as-made hydrophobic nanoparticles (oleic acid-coated) with

concentration of 0.1 wt% dispersed in butter oil (η=477 mPa s [55]), and 0.1 wt% hydrophilic nanoparticles(coated with glucose) dispersed in water (η=0.89 mPa s[56]) and DMEM (η=0.94 mPa s [57]). The measurementswere performed with a field amplitude (H0) of 200 Oe and afrequency of 571 kHz. The figure shows the variation of thetemperature ΔT with time, with respect to the initial roomtemperature ∼292–294 K. The SLP for all samples wasestimated from the slope of the initial part of the ΔT(t) curveby:

= C m m T tSLP d d , 7liq liq NPs( ) ( )/ /

where Cliq are the specific heat (Cbutter oil=2.2 J K−1 g−1,Cwater=4.18 J K−1 g−1 and CDMEM=4.18 J K−1 g−1), mliq

the mass of the liquid and mNPs is the mass of the nano-particles. The values of SLP for each sample dispersed inbutter oil, water and DMEM were estimated from the initialslope of the evolution of temperature with time and presentedon the right of figure 5. From these results it is observed thatthe samples have comparable heat absorption efficiencies,despite changes of viscosity by more than two orders ofmagnitude and the different hydrodynamic volumes. Theindependence of the SLP with the viscosity medium and theMNPs hydrodynamic volume indicates that the contributionof Brown mechanism in the heat generation is negligible andthe Néel relaxation dominates. Conversely, when the SLP ofdifferent MNPs systems dispersed in the same medium arecompared, a non-monotonous behavior of the SLP with thecomposition is observed with a clear maximum in the studiedZn composition range. When the MNPs are diluted in butteroil the SLP maximum is observed for the sample with 50% ofZn/Zn+Co shell atomic ratio, and near 60% when thesamples are dispersed in water and DMEM. These resultsevidence that, by changing the effective anisotropy with the

Figure 4. Magnetic anisotropy estimated from ⟨TB⟩, Vmag, andτ0=2(1) 10−11 s experimental parameters, (solid black dots) andcalculated from the SLP measurements (open red dot) as function ofZn/(Zn+Co) atomic shell ratio. The magnetic anisotropy of thecore, estimated from ⟨TB⟩, is also presented with empty star.

7

Nanotechnology 32 (2021) 065703 F Fabris et al

shell composition, the τN can be tuned to maximize the SLPvalue.

In order to confirm this picture, the SLP were numericallycalculated considering that only the magnetic mechanismcorresponding to the Néel relaxation contributes to the heatabsorption. The simulations were carried out using a modelbased on non-interacting single domain nanoparticles, asreported by De Biasi et al [58]. This model assumes that thenanoparticles are spatially fixed (i.e. the Brown’s relaxationmechanism is not enabled) and the thermal effects are mani-fested in two different ways. The first effect is to promote thechange of the magnetic moment orientation of the particles froman energy minimum to another, which is weighted by means ofthe probability to find a given particle in the superparamagneticregime: tº = t t-L L H T, , e ,m m *( ) / where τ* is the effectiveinversion time of the magnetization. This probability is the sameused in kinetic Monte Carlo calculations [59–61]. The secondeffect is the fluctuation of the magnetic moment within the sameenergy minimum, which results in the decrease of the effectivemagnetic moment, resulting in a statistical average of themagnetization vector projection in the direction of the appliedfield ( má ñ

). Taking these considerations in to account, the total

average magnetization of the system as a function of time isgiven by:

m m má ñ = á ñ + - á ñ + á ñM L L P P1 , 8SP 0 0 1 1( )[ ] ( )

where má ñ LSP corresponds to the superparamagnetic contrib-

ution and m m- á ñ + á ñL P P1 0 0 1 1( )[ ] to the blocked one. Thesubscript SP indicates the integration region corresponding tothe entire phase space, while the subscripts 0 and 1 indicate therespective energy minimum associated to the energy landscapedetermined by external field that corresponds to the integrationinterval, with respective normalized population given by P0 andP1 (P0+P1=1). The temporal evolution of magnetization isreflected in two aspects: firstly, the variation of the energylandscape, which results in changes of the statistical averages;secondly, the evolution of the populations P0 and P1 due to theinversion in the magnetization of some particles of the assemblygiven by P0(t+δt)=P0(t)+L[P0

∞− P0(t)], where δt is thetime interval within which the problem has been discretized(δH/δt is the sweep speed of the external field) and P0

∞ is theequilibrium population associated with the minimum energy 0.To perform the calculation the experimental data were taken asinput of the magnetic relaxation (see tables 1 and 2), namely the

Figure 5. Panels on the left present the magnetic fluid hyperthermia experiments measured at 571 kHz and 200 Oe for the nanoparticlessystems 0.1 wt% dispersed in butter oil (up), water (middle) and DMEM (down). Panels on the right give the SLP values as function of themeasured Zn/(Zn+Co) atomic shell ratio, in butter oil (up), water (middle) and DMEM (down). The stars symbols display the simulatedvalues of SLP for the core–shell nanoparticles dispersed in butter-oil and the red lines are the SLP of Fe3O4 core nanoparticles.

8

Nanotechnology 32 (2021) 065703 F Fabris et al

magnetic saturationMS, the magnetic activation volume, τ0, andthe frequency of the hyperthermia experiment ( f=571 kHz),and the applied field parameters (H0=200 Oe). We assumedthat the easy axes of the particles are randomly oriented. Thesimulated SLP data are shown in figure 5 for the highestviscosity medium, which are in good agreement with the valuesobtained from the experimental curve. In the simulations wehave allowed to vary only the value of the effective anisotropyenergy density Keff, forcing the rest of the parameters to befixed, according to the experimental conditions. In figure 4 wecompare the obtained density of magnetic anisotropy energyfrom the simulations with the experimental values, where a verygood agreement is obtained. These results also confirm that theNéel magnetic relaxation is the dominant mechanism.

Notice that although the shell grows epitaxial over thecore, a disordered interface was resolved and also multigrainshell structure was observed. However, dc-magnetic mea-surements are consistent with rigid magnetic couplingbetween the core and shell phases where single phase coer-cive loops were observed for all the temperature and com-position range. This picture is strongly supported by themagnetic relaxation measurements where the activationvolume, which corresponds to the smallest volume thatreverses coherently in an event, coincides with the measuredby TEM. Within the rigid coupling regime the magneticanisotropy can be extrapolated from the properties of the coreand shell phases which facilitate the design of core/shellsystem. In fact, the measured coercive fields for theFe3O4/ZnxCo1−xFe2O4 systems agree with the averagecoercivity expected for a powder of rigid coupled hard/softMNPs. The measurements also evidence the presence of weakmagnetic interactions, ascribed to dipolar interparticle inter-action that cannot be neglected despite the large nanoparticlesdispersions (0.2 wt% for the magnetic measurements).

The magnetic hyperthermia experiment with the nano-particles dispersed in different mediums shows very similarvalues of SLP, even when the viscosity of the mediums ismore than two orders of magnitude different. This resultconfirms that the Néel relaxation is the dominant mechanism,which is the expected one for butter oil due to its largeviscosity that inhibits the physical movement of the nano-particles in the MFH condition. Notice that if the particleshave a coherent inversion of the magnetic moment, and themagnetization response is linear with the alternating magneticfield, the Rosensweig’s model predicts the optimum SLPwhen the characteristic time of the experiment are in corre-spondence with the relaxation time of the system as: 2πfτ=1[5]. In the present MNPs system the coherent inversion of themagnetic moment is supported by the rigid core/shell cou-pling, and the agreement between the magnetic activationvolume and the volume measured by TEM. Moreover as thehyperthermia experiments were performed with larger nano-particles dispersion the system approaches to the hypothesisof non-interacting nanoparticles. Therefore, because theBrown mechanism is avoided for these systems, the Néelrelaxation time can be adjusted by a fine tuning of theeffective energy barrier until it reaches the optimum value forthe frequency used in the MFH experiments. Notice that to do

a fine tuning of the Néel relaxation time, the size of thenanoparticles is not a proper parameter since small variationsof it results in large changes of the energy barrier. Instead, thecore/shell NPs shown to be ideal systems since they can bedesigned with comparable size and magnetization and theyallow the fine adjustment of the magnetic anisotropy byadjusting the composition of a thin magnetic shell. For thecore/shell MNPs studied in this work, the optimum condi-tions were observed near of 50% of Zn/(Zn+Co) atomicshell ratio, when the MNPs are dispersed in butter oil,whereas the same MNPs displayed optimum heating near60% when they are dispersed in water and DMEM. Theexperimentally SLP observed values were successfullydescribed by a simple model considering non-interactingmagnetic moments where the Néel relaxation is the onlyactive mechanism to produce magnetic losses, with theeffective magnetic anisotropy as the single adjusted para-meter. The obtained value of the magnetic anisotropy fromthe SLP fitting was in excellent agreement with that calcu-lated from dc-magnetization, showing a systematic decreasingwhen the Zn concentration increases. These strongly coupledcore/shell MNPs provide a successful method to producehighly efficient heaters for viscous media such as the intra-cellular medium, through the control of the shell composition.This method could be in principle applied to MNPs of dif-ferent sizes to improve their heat efficiency, which is a keyrequisite for many MFH-based therapeutic applications.

4. Conclusions

The strategy to tune the effective magnetic anisotropy of MNPspresented in this work, is based on controlling the phase com-position of magnetically coupled core/shell nanoparticles. Thisallows to find the optimal composition, for a given particle size,that maximizes the heating efficiency of the system, even inhighly viscous media. We showed that by changing the shellcomposition of ∼10.7 nm Fe3O4/ZnxCo1−xFe2O4 core/shellnanoparticles, the effective magnetic anisotropy can be changedfrom 11×105 to 4×105 erg cm−3 when the Zn atomic ratio(Zn/(Zn+Co)) increases. In this way, proper condition tooptimize the SLP in MFH can be tuned for each working fre-quency. Thereby, the modulation of the composition of the thinshell coating and the interface core/shell nanoparticle couplingpermits not only the optimization of the heating power, main-taining the overall nanoparticle size and magnetization value,but it also improves the reproducibility of MNPs heatingproperties when they are dispersed in media of different visc-osity. Moreover, by combining in a single nanoparticle a softmagnetic core with a hard magnetic shell, the magnetic aniso-tropy can be enhanced, which allows to reach the optimumheating condition for lower nanoparticles size. Our results opennew possibilities to overcome current limitations in clinicalMFH, optimizing the heating power and extending the work-ability of the MFH to applications that require nanoparticles ofsmaller sizes.

9

Nanotechnology 32 (2021) 065703 F Fabris et al

Acknowledgments

The authors thank to Daniel E Fregenal and Guillermo CBernardi for the collaboration with the PIXE measurements.The authors acknowledge financial support of Argentiniangovernmental agency ANPCyT (Project No. PICT-2016-0288and PICT-2018-02565) and UNCuyo (Project No.06/C527and 06/C528). The authors gratefully acknowledge the EU-commission financial support under the: H2020-MSCA-RISE-2016, SPICOLOST PROJECT No 734187.

ORCID iDs

Myriam H Aguirre https://orcid.org/0000-0002-1296-4793Roberto D Zysler https://orcid.org/0000-0003-0687-5898Elin L Winkler https://orcid.org/0000-0002-9575-7879

References

[1] Mahmoudi K, Bouras A, Bozec D, Ivkov R and Hadjipanayis C2018 Magnetic hyperthermia therapy for the treatment ofglioblastoma: a review of the therapy’s history, efficacy andapplication in humans Int. J. Hyperth. 34 1316–28

[2] Maier-Hauff K, Ulrich F, Nestler D, Niehoff H, Wust P,Thiesen B, Orawa H, Budach V and Jordan A 2011 Efficacyand safety of intratumoral thermotherapy using magneticiron-oxide nanoparticles combined with external beamradiotherapy on patients with recurrent glioblastomamultiforme J. Neurooncol. 103 317–24

[3] Thorat N D and Bauer J 2020 Functional smart hybridnanostructures based nanotheranostic approach for advancedcancer treatment Appl. Surf. Sci. 527 146809

[4] Usov N A and Liubimov B Y 2012 Dynamics of magneticnanoparticle in a viscous liquid: application to magneticnanoparticle hyperthermia J. Appl. Phys. 112 023901

[5] Rosensweig R E 2002 Heating magnetic fluid with alternatingmagnetic field J. Magn. Magn. Mater. 252 370–4

[6] Lima E, De Biasi E, Zysler R D, Vasquez Mansilla M,Mojica-Pisciotti M L, Torres T E, Calatayud M P,Marquina C, Ricardo Ibarra M and Goya G F 2014 Relaxationtime diagram for identifying heat generation mechanisms inmagnetic fluid hyperthermia J. Nanopart. Res. 16 1–11

[7] Liu E, Yin Y, Sun L, Zhai Y, Du J, Xu F and Zhai H 2017Increasing spin polarization in Fe3O4 films by engineeringantiphase boundary densities Appl. Phys. Lett. 110 142402

[8] Patil U S, Adireddy S, Jaiswal A, Mandava S, Lee B R andChrisey D B 2015 In vitro/in vivo toxicity evaluation andquantification of iron oxide nanoparticles Int. J. Mol. Sci. 1624417

[9] Hergt R, Dutz S and Röder M 2008 Effects of size distributionon hysteresis losses of magnetic nanoparticles forhyperthermia J. Phys.: Condens. Matter 20 385214

[10] Gonzales-Weimuller M, Zeisberger M and Krishnan K M 2009Size-dependant heating rates of iron oxide nanoparticles formagnetic fluid hyperthermia J. Magn. Magn. Mater. 3211947–50

[11] Cruz M M, Ferreira L P, Ramos J, Mendo S G, Alves A F,Godinho M and Carvalho M D 2017 Enhanced magnetichyperthermia of CoFe2O4 and MnFe2O4 nanoparticlesJ. Alloys Compd. 703 370–80

[12] Noh S H, Na W, Jang J T, Lee J H, Lee E J, Moon S H, Lim Y,Shin J S and Cheon J 2012 Nanoscale magnetism control via

surface and exchange anisotropy for optimized ferrimagnetichysteresis Nano Lett. 12 3716–21

[13] Di Corato R, Espinosa A, Lartigue L, Tharaud M, Chat S,Pellegrino T, Ménager C, Gazeau F and Wilhelm C 2014Magnetic hyperthermia efficiency in the cellularenvironment fordifferent nanoparticle designs Biomaterials35 6400–11

[14] He S et al 2018 Maximizing specific loss power for magnetichyperthermia by hard–soft mixed ferrites Small 14 1–9

[15] Raland R D, Saikia D, Borgohain C and Borah J P 2017Heating efficiency and correlation between the structural andmagnetic properties of oleic acid coated MnFe2O4

nanoparticles for magnetic hyperthermia application J. Phys.D: Appl. Phys. 50 325004

[16] Lavorato G C, Das R, Xing Y, Robles J, Litterst F J,Baggio-Saitovitch E, Phan M H and Srikanth H 2020 Originand shell-driven optimization of the heating power in core/shell bimagnetic nanoparticles ACS Appl. Nano Mater. 31755–65

[17] Simeonidis K et al 2020 Controlling magnetization reversaland hyperthermia efficiency in core–shell iron–iron oxidemagnetic nanoparticles by tuning the interphase couplingACS Appl. Nano Mater. 3 4465–76

[18] Gangwar A, Kumar S, Meena S S, Sharma A, Viswanadh M K,Neogi K, Muthu M S and Prasad N K 2020 Structural andin-vitro assessment of ZnxFe3−xC (0�x�1) nanoparticlesas magnetic biomaterials Appl. Surf. Sci. 509 144891

[19] Lu A H, Salabas E L and Schüth F 2007 Magneticnanoparticles: synthesis, protection, functionalization, andapplication Angew. Chem.—Int. Ed. 46 1222–44

[20] Zubarev A Y 2019 Magnetic hyperthermia in a system ofimmobilized magnetically interacting particles Phys. Rev. E99 1–6

[21] Fu R, Yan Y, Roberts C, Liu Z and Chen Y 2018 The role ofdipole interactions in hyperthermia heating colloidal clustersof densely-packed superparamagnetic nanoparticles Sci.Rep. 8 1–10

[22] Serantes D et al 2010 Influence of dipolar interactions onhyperthermia properties of ferromagnetic particles J. Appl.Phys. 108 073918

[23] Lima E, De Biasi E, Mansilla M V, Saleta M E, Granada M,Troiani H E, Effenberger F B, Rossi L M,Rechenberg H R and Zysler R D 2013 Heat generation inagglomerated ferrite nanoparticles in an alternating magneticfield J. Phys. D: Appl. Phys. 46 045002

[24] Sanz B, Calatayud M P, De Biasi E, Lima E, Mansilla M V,Zysler R D, Ibarra M R and Goya G F 2016 In silico beforein vivo: how to predict the heating efficiency of magneticnanoparticles within the intracellular space Sci. Rep. 6

[25] Serantes D, Simeonidis K, Angelakeris M,Chubykalo-Fesenko O, Marciello M, Del Puerto Morales M,Baldomir D and Martinez-Boubeta C 2014 Multiplyingmagnetic hyperthermia response by nanoparticle assemblingJ. Phys. Chem. C 118 5927–34

[26] Périgo E A, Hemery G, Sandre O, Ortega D, Garaio E,Plazaola F and Teran F J 2015 Fundamentals and advancesin magnetic hyperthermia Appl. Phys. Rev. 2

[27] Salas G et al 2014 Modulation of magnetic heating via dipolarmagnetic interactions in monodisperse and crystalline ironoxide nanoparticles J. Phys. Chem. C 118 19985–94

[28] Lima E, Brandl A L, Arelaro A D and Goya G F 2006 Spindisorder and magnetic anisotropy in Fe3O4 nanoparticlesJ. Appl. Phys. 99 083908

[29] Vasilakaki M, Trohidou K N and Nogués J 2015 Enhancedmagnetic properties in antiferromagnetic-core/ferrimagnetic-shell nanoparticles Sci. Rep. 5 9609

[30] Skomski R and Coey J M D 1993 Giant energy product innanostructured two-phase magnets Phys. Rev. B 4815812–6

10

Nanotechnology 32 (2021) 065703 F Fabris et al

[31] Kneller E F and Hawig R 1991 The exchange-spring magnet: anew material principle for permanent magnets IEEE Trans.Magn. 27 3560–88

[32] Winkler E L, Lima E, Tobia D, Saleta M E, Troiani H E,Agostinelli E, Fiorani D and Zysler R D 2012 Origin ofmagnetic anisotropy in ZnO/CoFe2O4 and CoO/CoFe2O4

core/shell nanoparticle systems Appl. Phys. Lett. 101 252405[33] Zhao G P, Zhao M G, Lim H S, Feng Y P and Ong C K 2005

From nucleation to coercivity Appl. Phys. Lett. 87 1–3[34] Fullerton E E, Jiang J and Bader S 1999 Hard/soft magnetic

heterostructures: model exchange-spring magnets J. Magn.Magn. Mater. 200 392–404

[35] Lavorato G and Winkler E 2016 Thickness dependence ofexchange coupling in epitaxial Fe3O4/CoFe2O4 soft/hardmagnetic bilayers Phys. Rev. B 94 054405

[36] Klekotka U, Piotrowska B, Satuła D and Kalska-Szostko B2018 Modified ferrite core–shell nanoparticles magneto-structural characterization Appl. Surf. Sci. 444 161–7

[37] Lavorato G C, Lima E, Troiani H E, Zysler R D andWinkler E L 2017 Tuning the coercivity and exchange biasby controlling the interface coupling in bimagnetic core/shell nanoparticles Nanoscale 9 10240–7

[38] Fabris F et al 2019 Controlling the dominant magneticrelaxation mechanisms for magnetic hyperthermia inbimagnetic core–shell nanoparticles Nanoscale 11 3164–72

[39] Makridis A, Curto S, van Rhoon G C, Samaras T andAngelakeris M 2019 A standardisation protocol for accurateevaluation of specific loss power in magnetic hyperthermiaJ. Phys. D: Appl. Phys. 52 255001

[40] Wildeboer R R, Southern P and Pankhurst Q A 2014 On thereliable measurement of specific absorption rates andintrinsic loss parameters in magnetic hyperthermia materialsJ. Phys. D: Appl. Phys. 47 495003

[41] Zhao G P and Wang X L 2006 Nucleation, pinning, andcoercivity in magnetic nanosystems: an analyticalmicromagnetic approach Phys. Rev. B 74 2–5

[42] Leineweber T and Kronmüller H 1997 Micromagneticexamination of exchange coupled ferromagnetic nanolayersJ. Magn. Magn. Mater. 176 145–54

[43] Fabris F et al 2019 Tunnel magnetoresistance in self-assemblies of exchange-coupled core/shell nanoparticlesPhys. Rev. Appl. 11 054089

[44] Fabris F, Xing Y T, Franceschini D F, Sanchez D R,Alzamora M and Nunes W C 2017 Effects of postdepositionheat treatment on the structural and magnetic properties ofCoFe2O4 nanoparticles produced by pulsed laser depositionJ. Appl. Phys. 122 063901

[45] Yao C, Zeng Q, Goya G F, Torres T, Liu J, Wu H, Ge M, Zeng Y,Wang Y and Jiang J Z 2007 ZnFe2O4 nanocrystals: synthesisand magnetic properties J. Phys. Chem. C 111 12274–8

[46] Bruvera I J, Mendoza Zélis P, Pilar Calatayud M,Goya G F and Sánchez F H 2015 Determination of theblocking temperature of magnetic nanoparticles: the good,the bad, and the ugly J. Appl. Phys. 118 184304

[47] Micha J S, Dieny B, Régnard J R, Jacquot J F and Sort J 2004Estimation of the Co nanoparticles size by magnetic

measurements in Co/SiO2 discontinuous multilayersJ. Magn. Magn. Mater. 272–276 2003–4

[48] Livesey K L, Ruta S, Anderson N R, Baldomir D,Chantrell R W and Serantes D 2018 Beyond the blockingmodel to fit nanoparticle ZFC/FC magnetisation curves Sci.Rep. 8 1–9

[49] Lavorato G C, Lima E Jr, Tobia D, Fiorani D, Troiani H E,Zysler R D and Winkler E L 2014 Size effects in bimagneticCoO/CoFe2O4 core/shell nanoparticles Nanotechnology 25355704

[50] Peddis D, Cannas C, Piccaluga G, Agostinelli E and Fiorani D2010 Spin-glass-like freezing and enhanced magnetizationin ultra-small CoFe2O4 nanoparticles Nanotechnology 21125705

[51] Grady K O and Laidler H 1999 The limits to magneticrecording – media considerations J. Magn. Magn. Mater.200 616–33

[52] Laureti S, Varvaro G, Testa A M, Fiorani D, Agostinelli E,Piccaluga G, Musinu A, Ardu A and Peddis D , 2010Magnetic interactions in silica coated nanoporous assembliesof CoFe2O4 nanoparticles with cubic magnetic anisotropyNanotechnology 21 315701

[53] Lavorato G C, Lima E, Troiani H E, Zysler R D andWinkler E L 2015 Exchange-coupling in thermal annealedbimagnetic core/shell nanoparticles J. Alloys Compd. 633333–7

[54] Lavorato G, Winkler E, Ghirri A, Lima E, Peddis D,Troiani H E, Fiorani D, Agostinelli E, Rinaldi D andZysler R D 2016 Exchange bias and surface effects inbimagnetic CoO-core/Co0.5Ni0.5Fe2O4-shell nanoparticlesPhys. Rev. B 94 054432

[55] Hayashi H 1994 Viscoelasticity of Butter (Boston, MA:Springer US)

[56] Rumble J R 2017 CRC Handbook of Chemistry and Physics98th edn (Boca Raton, London, New York: CRC Press,Taylor & Francis Group)

[57] Fröhlich E, Bonstingl G, Höfler A, Meindl C, Leitinger G,Pieber T R and Roblegg E 2013 Comparison of two in vitrosystems to assess cellular effects of nanoparticles-containingaerosols Toxicol. In Vitro 27 409–17

[58] De Biasi E, Zysler R D, Ramos C A and Knobel M 2008 Anew model to describe the crossover fromsuperparamagnetic to blocked magnetic nanoparticlesJ. Magn. Magn. Mater. 320 312–5

[59] Anand M, Banerjee V and Carrey J 2019 Relaxation in one-dimensional chains of interacting magnetic nanoparticles:analytical formula and kinetic Monte Carlo simulationsPhys. Rev. B 99 1–11

[60] Niculaes D et al 2017 Asymmetric assembling of iron oxidenanocubes for improving magnetic hyperthermiaperformance ACS Nano 11 12121–33

[61] Tan R P, Carrey J and Respaud M 2014 Magnetichyperthermia properties of nanoparticles inside lysosomesusing kinetic Monte Carlo simulations: influence of keyparameters and dipolar interactions, and evidence for strongspatial variation of heating power Phys. Rev. B 90 1–12

11

Nanotechnology 32 (2021) 065703 F Fabris et al