Comparative Evaluation of Heating Ability and Biocompatibilityof Different Ferrite-Based Magnetic Fluids for HyperthermiaApplication

Pallab Pradhan,1 Jyotsnendu Giri,2 Gopal Samanta,3 Haladhar Dev Sarma,4 Kaushala Prasad Mishra,4

Jayesh Bellare,1 Rinti Banerjee,1 Dhirendra Bahadur2

1 School of Biosciences and Bioengineering, Indian Institute of Technology, Mumbai 400076, India

2 Department of Metallurgical Engineering and Materials Science, Indian Institute of Technology, Mumbai 400076, India

3 Department of Veterinary Surgery and Radiology, West Bengal University of Animal and Fishery Sciences,Kolkata 700037, India

4 Radiation Biology and Health Science Division, Bhabha Atomic Research Centre, Mumbai 400085, India

Received 12 November 2005; revised 31 March 2006; accepted 10 April 2006Published online 21 August 2006 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jbm.b.30630

Abstract: In this study, lauric acid-coated, superparamagnetic, nanoparticle-based magnetic

fluids of different ferrites (Fe3O4, MnFe2O4, and CoFe2O4) were prepared and compared in

terms of heating ability and biocompatibility to evaluate the feasibility of use in hyperthermia

treatment of cancer. All the magnetic fluids prepared had particles of average sizes 9–11 nm.

Heating ability of these magnetic fluids was evaluated by calorimetric measurement of specific

absorption rate (SAR) at 300 kHz frequency and 15 kA/m field. Fe3O4 and MnFe2O4 showed

higher SAR (120 and 97 W/g of ferrite, respectively) than CoFe2O4 (37 W/g of ferrite). In vitrostudy on BHK 21 cell lines showed dose-dependent cell viability for all the magnetic fluids.

Threshold-biocompatible ferrite concentration for all the magnetic fluids was 0.1 mg/mL.

Above 0.2 mg/mL, CoFe2O4 was more toxic than the other magnetic fluids. On intravenous

injection of different doses (50, 200, and 400 mg/kg body weight) of magnetic fluids in mice, no

significant changes in hematological and biochemical parameters were observed for Fe3O4 and

MnFe2O4. With CoFe2O4, an increase in SGPT levels at a dose rate of 400 mg/kg body weight

was observed, indicating its mild hepatotoxic effect. However, histology of different vital

organs showed no pathological changes for all the three magnetic fluids. Further, long term invivo evaluation of biocompatibility of the lauric acid-coated ferrites is warranted. This study

shows that lauric acid-coated, superparamagnetic Fe3O4 and MnFe2O4 may be used for

hyperthermia treatment and are to be preferred over CoFe2O4. ' 2006 Wiley Periodicals, Inc.

J Biomed Mater Res Part B: Appl Biomater 81B: 12–22, 2007

Keywords: magnetic nanoparticle; magnetic fluid; biocompatibility; specific absorption rate;

hyperthermia

INTRODUCTION

In recent years, magnetic particles have generated a lot of in-

terest in the biomedical applications, like MRI contrast

enhancement, magnetic separation, drug delivery, and hyper-

thermia.1–3 One of the prospective applications of magnetic

nanoparticles is hyperthermia treatment of cancer.4,5 In clini-

cal hyperthermia, efforts are made to optimize the thermal ho-

mogeneity at 42–468C in the tumor tissue by advanced

therapy and thermometry systems. Magnetic materials with

Curie temperature (Tc) between 42 and 608C are the best can-

didates for effective treatment, since they act as in vivo tem-

perature control switches, preventing over-heating.6 For

hyperthermia, superparamagnetic as well as ferrimagnetic par-

ticles, such as Fe3O4 or c-Fe2O3, have been explored exten-

sively.4,7,8 The systems M1�xZnxFe2O4 (M ¼ Mn, Co), which

have versatile magnetic properties and Curie temperature (Tc)obtained by substitution of nonmagnetic cations, have not yet

been explored.9,10 The heating ability and biocompatibility of

these magnetic fluids need to be investigated. Heating ability

of magnetic fluid which is measured by specific absorption

rate (SAR) depends on physical (particle size, shape, and dis-

tribution) as well as magnetic properties of particles.11 There

Correspondence to: D. Bahadur (e-mail: dhirenb@iitb.ac.in)

Contract grant sponsor: Department of Biotechnology (DBT) and Department ofScience and Technology (DST), India

' 2006 Wiley Periodicals, Inc.

12

are several research papers on SAR of superparamagnetic as

well as ferrimagnetic Fe3O4 and c-Fe2O3 particles, but suffi-

cient data is not available on other ferrites. In case of other fer-

rites, magnetic anisotropy energy plays an important role. A

high value of SAR for hyperthermia treatment of cancer is

beneficial, as the required dose can be minimized and thereby

toxicity due to overdoses can be avoided.

Biocompatibility of magnetic fluid is also an important

issue for suitability of the magnetic fluids for hyperthermia

treatment. Magnetic fluids are stable colloidal suspensions of

magnetic nanoparticles in carrier liquid. For suspension sta-

bility, magnetic nanoparticles are usually coated with sub-

stances, like surfactants or long chain polymers. Hence, the

biocompatibility of magnetic fluids is determined by both the

core ferrite material and the coatings.12 Fe3O4 has been

reported to be biocompatible.13 The other substituted ferrites

(M1�xZnxFe2O4 and Fe1�xMxFe2O4, M ¼ Co, Mn), where

the Fe is substituted by Mn, Co, and Zn ions, having versatile

magnetic properties could have superior magnetic properties,

but there is scanty data about their biocompatibility. In this

study, we have chosen two substituted ferrites (MnFe2O4,

CoFe2O4), where the maximum substitution of Fe ion for

Fe3O4 has been done by two metal ions (Mn and Co) for

evaluation of properties of relevance to hyperthermia. Coat-

ing materials also influence the biocompatibility of core

magnetic particles. The coating of long chain polymer mole-

cules, phospholipids, and other surfactants on ferrite particles

serves as protective layer on the nanoparticles and minimizes

the direct exposure of the ferrite surface to the biological

environment. Several surfactant and polymer-based coating

materials, like citrate, polyaspartic acid, dextran, for mainly

Fe3O4-based magnetic fluids, have been tested for biocom-

patibility.13–17 The dextran coating has been found to be

promising as far as biocompatibility is concerned.4,13,18 But

the use of dextran coating has been restricted for hyperther-

mia treatment of cancer because of loss of dextran shell by

enzymatic degradation in vivo.4 However, we have chosen

lauric acid (dodecanoic acid) for the stabilization of magnetic

particles in aqueous-based magnetic fluids. It is a fatty acid

having twelve (C-12) carbon chains, which gives stable and

dilution-insensitive, water-based magnetic fluid.19 Further,

C-12 is a natural medium chain fatty acid and is already

approved for use in pharmaceuticals and food industry.20

Therefore, the objective of this study was to develop and

characterize lauric acid-coated aqueous-based magnetic flu-

ids of different ferrites (Fe3O4, MnFe2O4, and CoFe2O4) and

compare their heating ability and biocompatibility for feasi-

bility of use in hyperthermia treatment of cancer.

MATERIALS AND METHODS

Materials

The salts used for magnetic fluid synthesis were manga-

nese(II) chloride (MnCl2�H2O), iron(III) chloride (FeCl3�H2O),

iron(II) chloride (FeCl2), cobalt chloride (CoCl2), tetrame-

thylammonium hydroxide, and ammonium hydroxide (spe-

cific gravity, 0.96). All the chemicals were from Aldrich

Chemie, USA, except bases (Thomas Baker, India). All

chemicals were analytical grades with 99.99% purity and

were used without further purification.

Synthesis of Magnetic Fluids

Magnetic fluids of different substituted ferrites were prepared

by coprecipitation technique in N2 atmosphere. A typical

detailed synthesis process was discussed elsewhere.19 In brief,

stoichiometric quantities of metal chlorides were dissolved in

milli-Q water. Metal ion solution (0.05M) was quickly added

into the base solution (NH4OH/tetramethyl ammonium hy-

droxide) with vigorous stirring, and the pH was adjusted to

11–12. The slurry was aged between 90 and 958C for 15–

20 min. The beaker containing the as-prepared slurry was

kept for 5 min on a permanent magnet (0.2 T surface fields)

to accelerate the settling. Then the salt solution was decanted

and the precipitate was repeatedly washed to remove any im-

purity ions. Lauric acid (30–40 wt % of ferrite) was added to

the precipitate. This mixture containing lauric acid and water

was heated (�1008C) with stirring till magnetic fluid forma-

tion. The magnetic fluid was subjected to purification for fur-

ther use. A portion of the wet particles was washed with

acetone and dried at room temperature for characterization,

while the rest was used for magnetic fluid preparation.

Purification of Magnetic Fluids

For purification, magnetic fluids were fast filtered through

Whatman 2 (qualitative) paper. Then excess lauric acid and

NH4+ ions were removed by dialysis against deionized water

using a cellulose membrane (molecular cut off, 12.4 kDa) for

72 h. The resultant magnetic fluids were centrifuged at

3500g for 15 min at room temperature, and purified magnetic

fluids were used for further characterization.

Determination of Ferrite Concentration in Fluid

Known volumes of magnetic fluid were weighed and kept in

an oven at 100–1208C for 3 h for drying. The dried coated

ferrites were weighed. These coated ferrites were subjected

to thermogravimetric analysis (TGA) up to 6008C. From

these data, weight fraction of ferrite, lauric acid, and water

was calculated. The dried coated particles were used for

magnetization measurement.

Characterizations

Determination of phase purity and identification was done by

X-ray diffraction (XRD) studies, using Philips powder dif-

fractometer PW1710, with Cu Ka radiation. The crystallite

size was determined from the X-ray line broadening using

Scherrer’s formula,

D ¼ 0:9k=b cos h ð1Þ

13COMPARATIVE EVALUATION OF DIFFERENT FERRITE-BASED MAGNETIC FLUIDS

Journal of Biomedical Materials Research Part B: Applied BiomaterialsDOI 10.1002/jbmb

where D is the average crystalline size, k is the X-ray wave-

length used, b is the angular line width of half maximum in-

tensity, and h is the Bragg’s angle in degrees. The size and

morphology of the coated particles in the ferrofluid were

observed using a transmission electron microscope (TEM)

(CM 200, Philips). For TEM observation, magnetic fluids

were diluted 10 times, and the samples were deposited drop-

wise onto a copper grid coated with carbon film and dried.

The hydrodynamic diameter of magnetic particles was deter-

mined using photon correlation spectroscopy (PCS) by Zeta

Plus (Brookhaven Instrument, USA). TGA of the uncoated

and coated magnetic particles was done using a Thermowage

L 81 (Germany) TGA unit with an incremental rate of tem-

perature of 78C/min.

Magnetic measurement of samples (powder and liquid) at

room temperature was done by VSM (Vibratory Sample

Magnetometer, Lake Shore, model-7410).

Measurement of Heating Ability by SAR

The heating ability of the different magnetic fluids was mea-

sured from the time-dependent calorimetric measurements.

The magnetic fluids of different concentrations were diluted,

so that resultant fluid contained 20 mg ferrite per milliliter.

Three milliliters of the sample was taken in a glass sample

holder, with suitable arrangements to minimize the heat loss.

The RF generator used for this experiment has a variable

field up to 20 kA/m with a fixed frequency of 300 kHz.

In this study, an AC magnetic field of 15 kA/m field and

300 kHz frequency was used. The SAR was calculated using

following equation:

SAR ¼ CDTDt

1

mFe

ð2Þ

where C is the specific heat of sample and sample holder,

DT/Dt is the slope of the time-dependent temperature curve

and mFe is mass of ferrite in grams.

In Vitro Cytocompatibility Study

In vitro biocompatibility study of different magnetic fluids

was done using BHK 21 (Syrian Baby Hamster Kidney) cell

lines purchased from National Centre for Cell Science, Pune,

India. The cells were grown in basal medium eagle (Sigma,

USA) supplemented with 10% fetal bovine serum, (Sigma),

10% tryptose phosphate broth (Becton Dickinson, USA), and

1% antibiotic antimycotic solutions (Himedia, India) and

incubated at 378C temperature under 5% CO2 under satu-

rated humid environment. Nearly confluent cells in 25-cm2

tissue culture flask (Nunc, USA) were trypsinized by trypsin-

EDTA solution and centrifuged at 1000g for 10 min. The cell

pellet was resuspended in fresh media. 1 3 104 cells were

added to each well of a 96-well tissue culture plate (Nunc)

and incubated in a CO2 incubator (Jouan IG150, France). Af-

ter 24 h, the magnetic fluid samples were added in five differ-

ent ferrite concentrations ranging from 0.05 to 0.6 mg/mL of

culture medium.

After 48 h, sulphorhodamine B (Sigma-Aldrich Chemie,

USA) assay was conducted similar to the procedure by Skehan

et al., with slight modifications.21 In brief, cells were fixed by

adding 50 mL of ice-cold 50% trichloroacetic acid (Loba

Chemie, India) slowly to the medium and incubated at 48C for

1 h. Then, the plates were washed five times with deionized

water and dried in air. Hundred microliters of 0.4% sulforhod-

amine B (Sigma) dissolved in 1% acetic acid was added to the

fixed cells and kept at room temperature for 20 min, after

which they were washed with 1% acetic acid to remove the

unbound dye. The plates were dried and 100 mL of 10 mM Tris

base (Sigma) was added to each well and kept for 20 min to

solubilize the dye. Thereafter, the plates were placed on a

shaker to allow mixing, and the absorbance (OD) of each well

was read in a plate reader (Thermo Electron Corporation,

USA) at 560 nm. Cell viability was measured as follows:

% viability ¼ Absorbance of sample/absorbance of control

3 100

In Vivo Biocompatibility Study

Animal Model and Experimental Design. For in vivobiocompatibility study, 84 Swiss female mice, weighing

about 20–25 g within an age group of 6–8 weeks, were taken.

They were maintained under standard laboratory conditions,

with supplementation of food and water ad libitum. The ani-mal experiment protocol was approved by the Institutional

Animal Ethics Committee, and all the experiments were done

in accordance with the guidelines of the Committee for the

Purpose of Control and Supervision of Experiments on

Animals, India. For this study, all the mice were divided into

different groups, having three animals in each group. Groups

consisted of control (common for all the three magnetic

Figure 1. XRD patterns of Fe3O4, MnFe2O4, and CoFe2O4 fattyacid-coated nanoparticles.

14 PRADHAN ET AL.

Journal of Biomedical Materials Research Part B: Applied BiomaterialsDOI 10.1002/jbmb

fluids), different magnetic fluids in three different doses

killed at three different time points.

The magnetic fluids were sterilized by autoclaving at 1218Cat 15 lb pressure for 30 min before in vivo application. These

magnetic fluids were injected intravenously through the tail

vein at three different dosages (50 mg, 200 mg, and 400 mg/kg

body weight). After 1, 6, and 24 h, animals were killed under

chloroform anesthesia. Blood samples were collected from the

heart in two different vials (one with EDTA for whole blood

and other without any anticoagulant for serum). Tissue samples

were collected from liver, lung, spleen, heart, and kidney, and

preserved in 10% formalin.

Hematological Parameters. Whole blood samples were

analyzed for total red blood cells (RBC), white blood cells

(WBC), hematocrit (HCT), hemoglobin (Hb), mean corpus-

cular Hb (MCH), MCH concentration (MCHC), and mean

corpuscular volume (MCV) in an automated hematology an-

alyzer (Sysmex, KX-21), and serum samples were analyzed

for blood urea nitrogen (BUN) by diacetylmonoxime method

and for serum glutamate pyruvate transaminase (SGPT) by

2,4-dinitrophenyl hydrazine colorimetric method, using com-

mercial kits (Span Diagnostic, India).

Histological Studies. Small fragments of each organ

were fixed in 10% formalin solution, dehydrated in series

dilution of ethanol, embedded in paraffin, and sections of

�3–5 mm were made with a microtome. Thereafter sections

were stained with hematoxylene–eosin solution. Observations

were made under an Olympus BX 51 microscope and photo-

graphs were taken in differential interference contrast (DIC)

mode.

Prussian Blue Staining. Qualitative study of magnetic

nanoparticle distribution in different organs was done by

Prussian blue staining of the organs. Iron deposition in differ-

ent organs was detected by the characteristic blue color of

Prussian blue. In brief, deparaffinized and hydrated sections

were immersed in a mixture containing equal volume of 20%

aqueous solution of HCl and 10% aqueous solution of potas-

sium ferrocyanide (K4Fe(CN)6,�3H2O) for 20 min, and coun-

terstaining was done by nuclear fast red for 5 min.

Photographs were taken in DIC mode using Olympus BX 51

microscope.

Statistical Analysis

Statistical analysis was done by analysis of variance

(ANOVA) and Newman–Keul’s test using Microsoft Excel

2000 software. A p value < 0.05 was taken as the cut off for

significance.

RESULTS

Characterization

Figure 1 shows the typical XRD pattern of the Fe3O4,

MnFe2O4, and CoFe2O4 nanoparticles. All the peaks corre-

spond to the cubic ferrite. The crystallite size of these ferrites

was measured from XRD line (311) broadening, using Scher-

rer’s formula. The average crystalline size is 9–11 nm. Fig-

ure 2 shows the TGA curve of the fatty acid-coated dried

sample of MnFe2O4 as a typical example. The mass profile

shows two distinct weight losses at different temperatures.

The dw/dt versus temperature curve shows that the major

weight loss temperatures are around 260 and 3408C. Similar

weight loss profile with temperature (TGA) was also

Figure 2. TGA curve and dw/dt versus temperature plot of fatty

acid-coated MnFe2O4 particles, showing different percentage of

weight loss and maximum temperature of decomposition.

Figure 3. Typical TEM photographs of lauric acid-coated (a) Fe3O4, (b) MnFe2O4, and (c) CoFe2O4

magnetic fluids. Bar ¼ 100 nm.

15COMPARATIVE EVALUATION OF DIFFERENT FERRITE-BASED MAGNETIC FLUIDS

Journal of Biomedical Materials Research Part B: Applied BiomaterialsDOI 10.1002/jbmb

observed for other ferrites (data not shown). The morphology

of the fatty acid-coated magnetic particles of the three mag-

netic fluids is shown in the Figure 3(a–c). The particles are

spherical in shape and well dispersed. The sizes of the par-

ticles in the three ferrofluids vary from 8 to 16 nm, and the

mean size is about 9–11 nm. Table I shows ferrite concentra-

tion, lauric acid content, average physical diameters, and

average hydrodynamic diameters of three magnetic fluids

and magnetization at 18 kOe field for three different mag-

netic fluids.

The particle size distribution pattern as obtained from the

PCS measurement is unimodal for all the magnetic fluids

(figure not shown). The magnetization loop of MnFe2O4

nanoparticles is depicted in Figure 4 as a typical example.

The result shows zero coercivity and zero remanence, and

the magnetization does not saturate at 20 kOe (see inset

of Figure 4). The magnetization at 200 Oe field is shown in

Table II.

Measurement of Heating Ability by SAR

Figure 5 shows time-dependent temperature rise of magnetic

fluids under AC magnetic field of 15 kA/m field and 300 kHz

frequency. SAR was calculated from the initial slope of the

time–temperature curve of different fluids, using Eq. (2).

Taking the value of K1 from literature,22 superparamagnetic

critical size (DP), Neel relaxation time (sN) of three ferrite

particles, and optimum particles size (DI) for maximum SAR

value, the product of x and sN with the ferrite composition

were calculated and summarized in Table II.

The superparamagnetic critical sizes (DP) are estimated

from the equation KV ¼ 25kT (where k ¼ Boltzmann’s con-

stant, V ¼ particle volume, K ¼ anisotropy constant, and T ¼absolute temperature). Since cobalt ferrite has higher mag-

netic crystalline anisotropy constant (1.8 3 106 erg/cm3)

than does the manganese ferrite/magnetite (1.4 3 105erg/

cm3 for Fe3O4 and 3.3 3 104 erg/cm3 for MnFe2O4), the esti-

mated superparamagnetic critical size for CoFe2O4, Fe3O4,

and MnFe2O4 are 14, 25, and 50 nm, respectively. For the

entire calculation, we have assumed that the particles are

spherical, and the average particle size was considered as the

magnetic core diameter. Anisotropy (K) of these spherical

particles is considered to be equal to magnetocrystalline ani-

sotropy (K1). It is important to note that the magnetic core di-

ameter for nanoparticles is always less than that of physical

diameter (obtained from TEM). The SAR values for different

magnetic fluids are given in Table II. Fe3O4 and MnFe2O4

showed SAR values of 120 and 97 W/g of ferrite, respec-

tively, which were much higher than CoFe2O4 (37 W/g

of ferrite). Figure 6 shows the variation of specific magnet-

ization at 200 Oe and SAR for the three magnetic fluids.

SAR of the Fe3O4 and c-Fe2O3 is reported with unit per gram

of iron in literature.23 Since we have substituted ferrite

particles (Fe2+ substituted by Co2+/Mn2+) to compare our

SAR values with the reported values, we have expressed

SAR per gram of metal ions, where the molecular weight of

metal ions is considered as the weight average (56.57) of

three metal ions. Thus the conversion factor between two

units will be 1.40 [SAR (W/g) of metal ions ¼ 1.40 3 SAR

(W/g) of ferrite)].

In Vitro Cytocompatibility Studies

In vitro cytocompatibility study showed dose-dependent cell

viability for all the magnetic fluids. Up to 0.1 mg/mL con-

centration, cell viability percentage was more than 90% for

all the magnetic fluids (Figure 7). At 0.2 mg/mL of ferrite

concentration, decrease in cell viability resulted with more

than 30% loss of viability for Fe3O4 and MnFe2O4 and more

than 50% loss of viability for CoFe2O4. CoFe2O4 showed

higher loss of cell viability than do Fe3O4 and MnFe2O4

TABLE I. Concentration of Ferrite, Lauric Acid Content, Physical Diameter, Hydrodynamic Diameter, and Magnetizationof Different Magnetic Fluid Samples

Magnetic Fluid

Composition

Ferrite

Concentration

(mg/mL)

Lauric Acid

Per Gram of

Ferrite

Avg. Physical

Diameter by

TEM (nm)

Hydrodynamic

Diameter by

PCS (nm)

Magnetization

at 18 kOe

(emu/g)

Fe3O4 27.12 0.100 9–10 100 64.54

MnFe2O4 32.20 0.115 10–11 70 58.18

CoFe2O4 27.74 0.110 9–10 90 59.56

Figure 4. Plot of magnetization versus magnetic field of MnFe2O4

nanoparticles. (Inset shows expanded scale.)

16 PRADHAN ET AL.

Journal of Biomedical Materials Research Part B: Applied BiomaterialsDOI 10.1002/jbmb

above 0.2 mg/mL ferrite concentration. But MnFe2O4

showed comparable cell viability as Fe3O4, for all the con-

centrations tested.

In Vivo Biocompatibility Studies

General Examination. The animals injected with mag-

netic fluids were active. No clinical signs of discomfort/dis-

tress were observed. No adverse reactions occurred during or

after the injection of magnetic fluids.

Hematological Parameters. There was no significant

change in hematological parameters, like total RBC, WBC,

HCT, Hb, MCH, MCHC, and MCV, in mice treated with any

of the magnetic fluids as compared to control group (p >

0.05; ANOVA, data not shown). None of the magnetic fluid-

treated mice showed any significant changes in BUN levels

as compared to control group up to the dose rate of 400 mg/

kg body weight (p > 0.05; ANOVA) (Figure 8). Though

SGPT levels increased with increased doses and intervals for

CoFe2O4, the results were not statistically significant (p >

0.05; ANOVA and Newman–Keul’s test) (Figure 9). How-

ever, no such increasing trend in SGPT levels was observed

for Fe3O4- and MnFe2O4-treated mice.

Histological Studies. There were no pathological

changes in organs like liver, lung, spleen, heart, and kidney

because of the administration of magnetic fluids upto 400

mg/kg body weight intravenously after 24 h. Figure 10 shows

the histological photomicrographs of liver [Figures 10 and

1(a–d)], lung, [Figures 10 and 2(a–d)] and spleen [Figures 10

and 3(a–d)] for 400 mg/kg body weight dose group at 24-h

intervals. In magnetic fluid-treated organs, tissue architecture

remained unchanged and no inflammatory or pathological

changes were detected even after 24 h of administration of

400 mg/kg body weight of ferrites intravenously. Numerous

Kupffer cells with engulfed particles were seen in the liver.

Clusters of particles were found in the liver, lung, and spleen

of magnetic fluid-treated mice. No pathological changes or

clusters of particles were observed in the heart and kidney.

Prussian Blue Staining. Micrographs from Prussian blue-

stained sections of Fe3O4-treated mice [Figure 11(a–e)] show

that there was extensive accumulation of magnetic nanopar-

ticles particularly in liver, lung, and spleen, which is evident

by the presence of blue color. But there was no accumulation

of magnetic nanoparticles in the heart and kidney. In spleen,

the particles were mostly accumulated in the red pulp area.

Similar distribution pattern was also observed for other mag-

netic fluids (figures not shown). As expected, Prussian blue

coloration was not observed for the control group (figures

not shown).

TABLE II. Variation of Magnetocrystalline Anisotropy Constant (K1) and Estimated Particles Size (DI) for Maximizationof SAR for the Different Magnetic Fluids

Ferrite

Composition

K1a

(joule/m3)

DP

(nm)

DI at 300

kHz (nm)

Dexp

(nm)

m at 200

Oe [emu/g]

SAR

(W/g ferrite)

SAR

(W/g metal)

xsNValueb

Fe3O4 1.4 3 104 25 15 10 19.92 120 168 0.0197

MnFe2O4 3.3 3 103 50 24 10 18.58 97 135.8 0.00327

CoFe2O4 1.8 3 105 14 7 10 6.56 37 51.8 1.33 3 107

DP, superparamagnetic critical size; DI, calculated particle size to satisfy the maximum SAR condition (xsN ¼ 1); Dexp, particle size of ferrofluid samples; m, particle magnetic

moment at 200 Oe.a K1 taken from reported value22.b Considered value of x ¼ 1.88 3 106 s�1 (i.e. 300 kHz) and the particle size Dexp.

Figure 5. Time–temperature curve of lauric acid-coated different

ferrite-based magnetic fluids.

Figure 6. Variation of specific magnetization (at 200 Oe) and SAR(at 15 kA/m) for three magnetic fluid samples.

17COMPARATIVE EVALUATION OF DIFFERENT FERRITE-BASED MAGNETIC FLUIDS

Journal of Biomedical Materials Research Part B: Applied BiomaterialsDOI 10.1002/jbmb

DISCUSSION

Magnetic fluids of different compositions (Fe3O4, MnFe2O4,

and CoFe2O4) were synthesized and characterized to evalu-

ate the possibilities of their successful use in hyperthermia

treatment. From XRD pattern of the Fe3O4, MnFe2O4, and

CoFe2O4 nanoparticles, it is evident that single phases were

formed in each of these ferrites. The broadness of the peak in

XRD of these ferrites is the signature of nanocrystallinity.

The arrangement of fatty acid molecules on the magnetic

nanoparticle in the stable water-based magnetic fluid may be

in bilayers with primary and secondary layers.24 In the pri-

mary layer, the carboxylic acid groups of the fatty acid are

chemisorbed to the ferrite surface and thus help in stabilizing

the nanoparticles. A secondary fatty acid layer is present on

the primary fatty acid molecules, with the interpenetration of

two layers equivalent to four carbon–carbon bonds. The sec-

ondary fatty acid layer provides a hydrophilic outer shell

exposed to the surrounding polar solvent, resulting in stable

water-based magnetic fluid. From the TGA curve, two dis-

tinct weight losses at different temperatures were found. dw/dt versus temperature curve reveals that the major weight

loss temperatures are around 260 and 3408C. These weight

losses may be due to desorption and evaporation of second-

ary (outer) and primary (inner) layers of fatty acid molecules.

The broad weight loss at higher temperature at around 3408Ccorresponds to the inner layer (primary) fatty acid molecule,

which is chemisorbed on the particle surface having higher

bond energy. The relatively sharp weight loss at 2608C is

due to the outer layer (secondary) fatty acid molecules with

lower bond energy. It is to be noted that three different fatty

acid-coated ferrite nanoparticles (Fe3O4, MnFe2O4, and

CoFe2O4) show the same weight loss profile with tempera-

ture (TGA), indicating presence of double layer of lauric acid

on nanoparticles for all the three magnetic fluids.

From the TEM micrographs, it can be seen that the par-

ticles are spherical in shape. The mean particles sizes are

about 9–11 nm. This is in good agreement with the size

measured form the X-ray line (311) broadening using Scher-

rer’s formula.

Magnetization loops of ferrite nanoparticles show zero

coercivity and zero remanence, and the magnetization does

not saturate at 20 kOe, indicating superparamagnetic nature

of ferrite particles. All the ferrite nanoparticles have room

temperature-specific magnetization values less than that of

their corresponding bulk state. The reduction of magnetic

properties of nanosized particles may be due to several addi-

tional factors, like formation of magnetic dead layer on the

surface, existence of random canting of particle surface spin,

nonsaturation effects due to random distribution of particle

size, deviation from the normal cation distribution, absorbed

water etc.25–27 It is interesting to note that the specific

magnetization at 1.8 kOe for CoFe2O4 is higher than for

MnFe2O4 whereas at lower field (200 Oe) it reverses. The

mean particles size of all the ferrites are nearly same; thus,

one can expect that the same degree of reduction of magnet-

ization may be due to reduction of size. Sato et al. observed

that the extent of reduction of saturation magnetization for

different ferrite nanoparticles (having same particle size)

largely depends on the crystalline magnetic anisotropy con-

stant K1.28 Among the three ferrites studied, CoFe2O4 has the

highest K1 value whereas MnFe2O4 has the lowest K1 value

(Table II). The higher specific magnetization (1.8 kOe) of

CoFe2O4 nanoparticles than of MnFe2O4 nanoparticles may

be taken into account by K1. The magnetization at lower field

(200 Oe), that is, the initial susceptibility, depends on several

parameters like particle size and anisotropy. The initial sus-

ceptibility also depends on the composition of ferrite par-

ticles, since the magnetocrystalline anisotropy (K1) depends

on the presence and types of cations at the different sub-

Figure 8. Effect of different magnetic fluids on BUN levels following

i.v. administration at the dose rate of 400 mg/kg body weight for differ-ent intervals in mice. Results are expressed as mean6 SD; n¼ 3.

Figure 9. Effect of different magnetic fluids on SGPT levels following

i.v. administration at the dose rate of 400 mg/kg body weight for differ-

ent intervals in mice. Results are expressed as mean6 SD; n¼ 3.Figure 7. Percent viability of BHK 21 cells 48 h after incubation of

different magnetic fluids at different ferrite concentrations. Results

are expressed as mean 6 SD; n ¼ 3.

18 PRADHAN ET AL.

Journal of Biomedical Materials Research Part B: Applied BiomaterialsDOI 10.1002/jbmb

lattices (A and B sites). Thus, the magnetization of Fe3O4 at

200 Oe (19.92 emu/g) is much higher than CoFe2O4 (6.56

emu/g). It is to be noted that both particles have the same aver-

age particle size (9–11 nm) as confirmed by XRD and TEM.

SAR value of Fe3O4 measured at an AC magnetic field

(H) of 15 kA/m and a frequency (f) of 300 kHz is a moder-

ately good value as compared to the values reported in litera-

ture.23 Report on SAR values of MnFe2O4 and CoFe2O4 is

scanty in literature, as these ferrites are hardly exploited for

the hyperthermia application. However, MnFe2O4 has shown

fairly good SAR value as compared to Fe3O4, whereas

CoFe2O4 has much lower value than Fe3O4. For superpara-

magnetic particles, SAR results from relaxation processes.

These processes in magnetic fluids are either due to reorien-

tation of the magnetization within the particle (Neel losses)

or due to frictional losses (Brown losses). The Brownian

relaxation time can be calculated using following formula:

sB ¼ 4pgr3

kTð3Þ

Figure 10. Histological micrograph of mice (1) livers, (2) lungs, and (3) spleens for (a) control and

after 24 h of i.v. administration of (b) Fe3O4, (c) MnFe2O4, and (d) CoFe2O4 at the dose rate of400 mg/kg body weight.

19COMPARATIVE EVALUATION OF DIFFERENT FERRITE-BASED MAGNETIC FLUIDS

Journal of Biomedical Materials Research Part B: Applied BiomaterialsDOI 10.1002/jbmb

and the Neel relaxation time is

sN ¼ s0 expKV

kTð4Þ

where g is the viscosity of carrier liquid, r the hydrodynamic

radius of particles, k the Boltzmann’s constant, s0 the time

constant (�10�9 s), V the particle volume, and K is the ani-

sotropy constant. Thus the power loss corresponds to Neel

and/or Brownian relaxation and is calculated approximately

from the following equation:

P ¼ ðmHxsÞ2=½2skT qVð1þ x2s2Þ� ð5Þ

where m is the particle magnetic moment, x the angular fre-

quency, H the AC field amplitude, and q is the density of fer-

rite. From Eq. (5), it is clear that the power loss (P) dependson the (a) particle magnetic moment (m), (b) AC magnetic

field amplitude (H), and (c) AC magnetic field frequency

(x). Thus the SAR is proportional to the square of magnetic

moment. The variation of SAR for different magnetic fluids

follows the same pattern as the specific magnetizations vary

at lower field (200 Oe), that is, 16 kA/m. On the other hand,

at a particular value of m and H, P will reach a maximum

value when xs ¼ 1.29 In our systems (dh ¼ 70–100 nm, f ¼300 kHz), contribution due to Brownian losses is very small,

since the product of x and sB is far apart from 1. This was

also confirmed by a polyacrylic gel experiment in which the

magnetic fluids were immobilized in polyacrylic gel and

thereby the viscosity of magnetic particles becomes too high

to show the Brownian loss (result not shown here). If we con-

sider the SAR due to Neel losses, the relaxation time depends

on V and K. For magnetic particles of particular composition,

Neel relaxation time (sN) depends on the particle volume

(i.e. size). To maximize P (SAR) of a particular ferrite parti-

cle at any AC field frequency (f or x), the particle volume

should be such that the product of x and sN becomes 1. With

varying composition of the ferrite particles, the K value will

change and it will decide sN. With the knowledge of K, theoptimum particle size could be calculated for maximum P(SAR) and is summarized in Table II. The optimum particle

sizes (DI) to maximize SAR at 300 kHz AC frequency for

Fe3O4, MnFe2O4, and CoFe2O4 are 15, 25, and 7 nm, respec-

tively. The average particle size of our ferrite nanoparticles

is in between 9 and 11 nm (measured from XRD and TEM)

which are bigger/smaller than their optimum sizes (DI). The

effect of particle size on sN, as well as on SAR, reflects on

the value of xsN and hence it influences its degree of devia-

tion from the maximum power loss condition (i.e. xsN ¼ 1).

From the table it is found that though there is not much dif-

ference in low field specific magnetization for MnFe2O4 and

Fe3O4, the SAR is relatively higher for Fe3O4. This could be

due to the deviation of particle sizes from their optimum

sizes. Thus, the value of xsN for MnFe2O4 (0.00327) is

smaller than that of Fe3O4 (0.0197) and much away from 1

(maximum power loss condition).

From the earlier discussion it is clear that the SAR of the

Fe3O4 and MnFe2O4 can be enhanced to higher value by sim-

Figure 11. Prussian blue-stained sections of mice (a) liver, (b) lung, (c) spleen, (d) heart, and (e) kid-

ney after 24 h of i.v. administration of Fe3O4 magnetic fluid at the dose rate of 400 mg/kg bodyweight. [Color figure can be viewed in the online issue, which is available at www.interscience.

wiley.com.]

20 PRADHAN ET AL.

Journal of Biomedical Materials Research Part B: Applied BiomaterialsDOI 10.1002/jbmb

ply increasing the particle size to their optimum size (DI)

(15 nm for Fe3O4 and 25 nm for MnFe2O4, respectively). In

case of CoFe2O4, SAR could be enhanced further by reduc-

ing the particle size to 7 nm (DI). But, initial susceptibility

(magnetization at 200 Oe) of CoFe2O4 will be further

reduced because of size reduction and thereby lower value of

SAR will be observed. Thus, even with optimum particle size

(DI), one can expect very low SAR value for CoFe2O4

because of high anisotropy value.

From in vitro biocompatibility study, it is evident that the

cytotoxicity of magnetic fluids depends on the concentration of

different ferrites. For all the magnetic fluids, 0.1 mg/mL is the

threshold concentration beyond which decrease in cell viability

is observed. Threshold concentrations for dextran-coated Fe3O4

has been reported to be in the range of 1–2 mg/mL,4 which is

higher than lauric acid-coated Fe3O4 reported here. This differ-

ence in threshold concentration might be due to the use of dif-

ferent coatings (i.e. lauric acid) in this study. Moreover, reports

on cytotoxicity of other ferrites, like MnFe2O4 and CoFe2O4,

having lauric acid coating are scanty in literature. When lauric

acid-coated MnFe2O4 is compared with lauric acid-coated

Fe3O4, no significant difference in viability is found upto

0.6 mg/mL (highest concentration tested) (p > 0.05; ANOVA).

But CoFe2O4 has shown comparatively higher loss of cell via-

bility (>10% viability) at higher doses (>0.2 mg/mL), indicating

more cytotoxicity as compared to Fe3O4 and MnFe2O4.

For in vivo evaluation of biocompatibility, different hemato-

logical parameters and histology of different organs were eval-

uated. Effects of different magnetic fluids on renal and liver

functions were evaluated by changes in BUN and SGPT levels,

respectively, for different doses and intervals. All the magnetic

fluids have been found to be nonnephrotoxic, as revealed by no

significant change in BUN levels for magnetic fluid-treated

groups as compared to control group (p > 0.05; ANOVA).

Fe3O4 and MnFe2O4 did not cause any significant change in

SGPT levels as compared to control groups (p > 0.05; ANOVA

and Newman–Keul’s test), indicating their nontoxicity to liver.

But for CoFe2O4, an increasing trend (nonsignificant, p > 0.05;

ANOVA and Newman–Keul’s test) in SGPT levels could be

due to the mild hepatotoxic effect of CoFe2O4. However, histol-

ogy of organs did not reveal any pathological changes for all

the three magnetic fluids, indicating their biocompatibility.

Prussian blue staining shows extensive accumulation of

magnetic particles mainly in liver, lung, and spleen, which is

attributed to the phagocytosis of magnetic particles by the

phagocytic cells of these organs. In spleen, accumulation of

particles was mainly in the red pulp, which is in agreement

with other reports.13,18 As spleen white pulp does not have

mononuclear phagocytes, no accumulation of particles was

found in this region. However, unlike spleen, clusters of par-

ticles are distributed homogenously throughout the liver and

lung. On the contrary, kidney and heart did not show any

accumulation of magnetic particles. However, no pathologi-

cal changes were detected even after massive accumulation

of magnetic particles in liver, lung, and spleen, indicating

their biocompatibility.

To the best of our knowledge, very few groups have eval-

uated the biocompatibility of MnFe2O4- and CoFe2O4-based

magnetic fluid. Lacava et al.30,31 reported that ionic-, citrate-,

and tartarate-coated manganese ferrite-based magnetic fluid

were not biocompatible. However, in our study lauric acid-

coated manganese ferrite-based magnetic fluids did not show

any toxic effects at 400 mg/kg body weight up to 24 h. This

difference might be due to the different coating materials

used (i.e. lauric acid) in this study. Kuckelhaus et al.12,32 did

not find any morphological alteration following intravenous

injection of citrate-coated cobalt ferrite-based magnetic fluid.

In our study, though no morphological changes in vital

organs have been detected after magnetic fluid injection, an

increase in SGPT levels has been found after CoFe2O4-based

magnetic fluid injection, which indicates that this composi-

tion is mildly hepatotoxic. But Fe3O4- and MnFe2O4-based

magnetic fluid has been found to be nontoxic from both the

morphological and hematological point of view.

SUMMARY

In this study, lauric acid-coated different ferrites (Fe3O4,

MnFe2O4, and CoFe2O4) based magnetic fluids have been

successfully prepared and characterized, and their heating

ability and biocompatibility have been comparatively eval-

uated. CoFe2O4-based magnetic fluid has lower SAR

(thereby heating ability) than Fe3O4- and MnFe2O4-based

magnetic fluids. However, Fe3O4- and MnFe2O4-based mag-

netic fluids have comparable SAR values. The SAR of the

Fe3O4 and MnFe2O4 systems can be further enhanced to a

maximum value by increasing the particles size to the opti-

mum size (DI). But even after optimizing particle size, higher

SAR values for CoFe2O4 system could not be achieved

because of its high anisotropy energy. All the magnetic fluids

have dose-dependent effect on cell viability. Threshold cyto-

compatible concentration of all the magnetic fluids is 0.1 mg/

mL. However, CoFe2O4-based magnetic fluid is more cyto-

toxic to BHK 21 cells than the others at higher concentra-

tions (0.2 mg/mL). Lauric acid-coated Fe3O4- and MnFe2O4-

based magnetic fluids are biocompatible in vivo whereas

CoFe2O4-based magnetic fluid is mildly hepatotoxic at

higher doses. However, in vivo biocompatibility study was

evaluated only for a short term of 24 h in the present study.

Further, longer term in vivo biocompatibility needs to be

evaluated for the evidence of inertness of the surfactant-

coated magnetic particles prior to use in hyperthermia.

From this study, it can be concluded that lauric acid-

coated, superparamagnetic, MnFe2O4-based magnetic fluid is

comparable to Fe3O4-based magnetic fluid as far as biocom-

patibility and SAR are concerned. But CoFe2O4-based mag-

netic fluid has lesser SAR and biocompatibility as compared

to Fe3O4 and MnFe2O4. Hence Fe3O4- and MnFe2O4-based

magnetic fluids are suitable for hyperthermia application,

whereas CoFe2O4 is not suitable for this purpose. However,

further optimization of particle size to maximize SAR, effi-

cacy of hyperthermia treatment, and longer term in vivo bio-

21COMPARATIVE EVALUATION OF DIFFERENT FERRITE-BASED MAGNETIC FLUIDS

Journal of Biomedical Materials Research Part B: Applied BiomaterialsDOI 10.1002/jbmb

compatibility study are required prior to their use in hyper-

thermia application (work in progress).

REFERENCES

1. Bahadur D, Giri J. Biomaterials and magnetism. Sadhana2003;28:639–656.

2. Pankhurst QA, Connolly J, Jones SK, Dobson J. Applicationsof magnetic nanoparticles in biomedicine. J Phys D: ApplPhys 2003;36:167–181.

3. Gupta AK, Gupta M. Synthesis and surface engineering ofiron oxide nanoparticles for biomedical applications. Biomate-rials 2005;26:3995–4021.

4. Jordan A, Wust P, Scholz R, Tesche B, Fahling H, MitrovicsT, Vogl T, Cervos-Navarro J, Felix R. Cellular uptake ofmagnetic fluid particles and their effects on human adenocar-cinoma cells exposed to AC magnetic fields in vitro. Int JHyperthermia 1996;12:705–722.

5. Jordan A, Scholz R, Wust P, Fahling H, Felix R. Magneticfluid hyperthermia (MFH): Cancer treatment with AC mag-netic field induced excitation of biocompatible superparamag-netic nanoparticles. J Magn Magn Mater 1999;201:413–419.

6. Giri J, Ray A, Dasgupta S, Datta D, Bahadur D. Investigationon Tc tuned nano particles of magnetic oxides for hyperther-mia applications. Biomed Mater Eng 2003;13:387–399.

7. Kawashita M, Tanaka M, Kokubo T, Inoue Y, Yao T, Ham-ada S, Shinjo T. Preparation of ferrimagnetic magnetitemicrospheres for in situ hyperthermic treatment of cancer.Biomaterials 2005;26:2231–2238.

8. Hergt R, Hiergeist R, Zeisberger M, Glockl G, Weitschies W,Ramirez LP, Hilger I, Kaiser WA. Enhancement of AC-lossesof magnetic nanoparticles for heating applications. J MagnMagn Mater 2004;280:358–368.

9. Grasset F, Mornet S, Demourgues A, Portier J, Bonnet J,Vekris A, Duguet E. Synthesis, magnetic properties, surfacemodification and cytotoxicity evaluation of Y3Fe5-xAlxO12 (0� x � 2) garnet submicron particles for biomedical applica-tions. J Magn Magn Mater 2001;234:409–418.

10. Giri J, Sriharsha T, Bahadur D. Optimization of parametersfor the synthesis of nano-sized Co1–xZnxFe2O4, (0 � x � 0.8)by microwave refluxing. J Mater Chem 2004;14:875–880.

11. Hergt R, Hiergeist R, Hilger I, Kaiser WA, Lapatnikov Y, Mar-gel S, Richter U. Maghemite nanoparticles with very high AC-losses for application in RF-magnetic hyperthermia. J MagnMagn Mater 2004;270:345–357.

12. Kuckelhaus S, Reis SC, Carneiro MF, Tedesco AC, OliveiraDM, Lima ECD, Morais PC, Azevedo RB, Lacava ZGM.In vivo investigation of cobalt ferrite-based magnetic fluid andmagnetoliposomes using morphological tests. J Magn MagnMater 2004;272:2402,2403.

13. Lacava LM, Garcia VAP, Kuckelhaus S, Azevedo RB, Sade-ghiani N, Buske N, Morais PC, Lacava ZGM. Long-termretention of dextran-coated magnetite nanoparticles in theliver and spleen. J Magn Magn Mater 2004;272:2434, 2435.

14. Lacava LM, Lacava ZGM, Azevedo RB, Chaves SB, GarciaVAP, Silva O, Pelegrini F, Buske N, Gansau C, Da Silva MF,Morais PC. Use of magnetic resonance to study biodistribu-tion of dextran-coated magnetic fluid intravenously adminis-tered in mice. J Magn Magn Mater 2002;252:367–369.

15. Freitas MLL, Silva LP, Azevedo RB, Garcia VAP, LacavaLM, Grisolia CK, Lucci CM, Morais PC, Da Silva MF, Buske

N, Curi R, Lacava ZGM. A double-coated magnetite-basedmagnetic fluid evaluation by cytometry and genetic tests.J Magn Magn Mater 2002;252:396–398.

16. Sadeghiani N, Barbosa LS, Silva LP, Azevedo RB, MoraisPC, Lacava ZGM. Genotoxicity and inflammatory investiga-tion in mice treated with magnetite nanoparticles surfacecoated with polyaspartic acid. J Magn Magn Mater 2005;28:466–468.

17. Petri-Fink A, Chastellain M, Juillerat-Jeanneret L, Ferrari A,Hofmann H. Development of functionalized superparamag-netic iron oxide nanoparticles for interaction with human can-cer cells. Biomaterials 2005;26:2685–2694.

18. Lacava LM, Garcia VAP, Kuckelhaus S, Azevedo RB,Lacava ZGM, Silva O, Pelegrini F, Gansau C, Buske N, Mor-ais PC. Magnetic resonance and light microscopy investiga-tion of a dextran coated magnetic fluid. J Appl Phys 2003;93:7563–7565.

19. Giri J, Pradhan P, Banerjee R, Datta D, Bahadur D. Synthesis,characterization and in vitro evaluation of water based tempera-ture sensitive ferrofluids for hyperthermia applications. Pre-sented at fifth International conference on the scientific andclinical applications of magnetic carriers, Lyon, May 20–22,2004.

20. Kinderlerer JL. Degradation of the lauric acid oils. Int Bio-deter Biodegrad 1994;33:345–354.

21. Skehan P, Storeng R, Scudiero D, Monks A, McMahon J,Vistica D, Warren JT, Bokesch H, Kenney S, Boyd MR. Newcolorimetric cytotoxicity assay for anticancer-drug screening.J Natl Cancer Inst 1990;82:1107–1112.

22. Bozorth RM, Tilden EF, Williams AJ. Anisotropy and magne-tostriction of some ferrites. Phys Rev 1955;99:1788–1798.

23. Andra W. Magnetic hyperthermia. In: Andra W, Nowak H,editor. Magnetism inMedicine. Berlin: Wiley; 1998. p 450–470.

24. Shen L, Laibinis PE, Hatton TA. Aqueous magnetic fluids sta-bilized by surfactant bilayers. J Magn Magn Mater 1999;194:37–44.

25. Coey JMD. Noncollinear spin arrangement in ultrafine ferri-magnetic crystallites Phys Rev Lett 1971;27:1140–1142.

26. Pankhurst QA, Polland RJ. Origin of the spin-canting anom-aly in small ferrimagnetic particles. Phys Rev Lett 1991;67:248–250.

27. Chen JP, Sorensen CM, Klabunde KJ, Hadjipanayis GC,Delvin E, Kostikas A. Size-dependent magnetic properties ofMnFe2O4 fine particles synthesized by coprecipitation. PhysRev B: Condens Matter 1996;54:9288–9296.

28. Sato T, Iijima T, Seki M, Inagaki N. Magnetic properties ofultrafine ferrite particles. J Magn Magn Mater 1987;65:252–256.

29. Debye P. Polar Molecules. New York: The Chemical CatalogCompany; 1929.

30. Lacava ZGM, Azevedo RB, Martins EV, Lacava LM, FreitasMLL, Garcia VAP, Rebula CA, Lemos APC, Sousa MH,Tourinho FA, Da Silva MF, Morais PC. Biological effects ofmagnetic fluids: Toxicity studies. J Magn Magn Mater 1999;201:431–434.

31. Lacava ZGM, Azevedo RB, Lacava LM, Martins EV, GarciaVAP, Rebula CA, Lemos APC, Sousa MH, Tourinho FA,Morais PC, Da Silva MF. Toxic effects of ionic magnetic flu-ids in mice. J Magn Magn Mater 1999;194:90–95.

32. Kuckelhaus S, Garcia VAP, Lacava LM, Azevedo RB,Lacava ZGM, Lima ECD, Figueiredo F, Tedesco AC, MoraisPC. Biological investigation of a citrate-coated cobalt–ferrite-based magnetic fluid. J Appl Physics 2003;93:6707,6708.

22 PRADHAN ET AL.

Journal of Biomedical Materials Research Part B: Applied BiomaterialsDOI 10.1002/jbmb