Heat dissipation in Sm3+ and Zn2+ co-substituted magnetite (Zn0.1SmxFe2.9-xO4) nanoparticles coated with citric acid and pluronic F127 for hyperthermia application

In this work, Sm3+ and Zn2+ co-substituted magnetite Zn0.1SmxFe2.9-xO4 (x = 0.0, 0.01, 0.02, 0.03, 0.04 and 0.05) nanoparticles, have been prepared via co-precipitation method and were electrostatically and sterically stabilized by citric acid and pluronic F127 coatings. The coated nanoparticles were well dispersed in an aqueous solution (pH 5.5). Magnetic and structural properties of the nanoparticles and their ferrofluids were studied by different methods. XRD studies illustrated that all as-prepared nanoparticles have a single phase spinel structure, with lattice constants affected by samarium cations substitution. The temperature dependence of the magnetization showed that Curie temperatures of the uncoated samples monotonically increased from 430 to 480 °C as Sm3+ content increased, due to increase in A-B super-exchange interactions. Room temperature magnetic measurements exhibited a decrease in saturation magnetization of the uncoated samples from 98.8 to 71.9 emu/g as the Sm3+ content increased, which is attributed to substitution of Sm3+ (1.5 µB) ions for Fe3+ (5 µB) ones in B sublattices. FTIR spectra confirmed that Sm3+ substituted Zn0.1SmxFe2.9-xO4 nanoparticles were coated with both citric acid and pluronic F127 properly. The mean particle size of the coated nanoparticles was 40 nm. Calorimetric measurements showed that the maximum SLP and ILP values obtained for Sm3+ substituted nanoparticles were 259 W/g and 3.49 nHm2/kg (1.08 mg/ml, measured at f = 290 kHz and H = 16kA/m), respectively, that are related to the sample with x = 0.01. Magnetic measurements revealed coercivity, which indicated that hysteresis loss may represent a substantial portion in heat generation. Our results show that these ferrofluids are potential candidates for magnetic hyperthermia applications.

MNPs play an important role as heat exchangers in a ferrofluid in the RF magnetic hyperthermia. Use of MNPs causes fewer side effects as damage to healthy tissues than normal cancer treatment methods. This method decreases the cancerous cells and augments their sensitivity to chemotherapy and radiation additionally by rising the local temperature of targeted tissues to an interval from 42 to 46 °C 13,14 . The heating efficiency of MNPs is measured through the specific loss power (SLP) that is also referred to as specific absorption rate (SAR). It has been found that the SLP of MNPs can be modified by tuning saturation magnetization, effective anisotropy, and particle size 15 .
MNPs tend to agglomerate because of their large surface to volume ratio and magnetic dipole-dipole interactions 1 . In order to reduce and/or avoid sedimentation and to enhance biocompatibility and functionalization, it is necessary to coat MNPs with surfactants or polymers. The stability of ferrofluids against agglomeration is related to a competition between various interactions, such as Van der Waals, magnetic dipole-dipole interactions, viscous drag force from the carrier fluid, and electrostatic and steric repulsion resulting from the surfactants in the coating 16,17 .
Achieving stability of ferrofluids in polar media (e.g., water) is a more complicated challenge than non-polar media (e.g., oil). Electrostatic repulsive forces between MNPs due to a high electric surface charge density may allow achieving long-term stability for water-based ferrofluids. Steric stabilization may not be sufficient for obtaining a stable colloidal suspension of particles with large magnetic core which introduce strong magnetic attraction forces. In such cases, strong electrostatic repulsion forces between the particles can be obtained by coating them with a highly charged material. To this end, citric acid (CA) is an appropriate candidate to coat particles, specifically for magnetic nanoparticle with strong magnetic interaction. CA has a high biocompatibility and introduces both, electrostatic and steric repulsion effects. CA has three carboxyl groups and a hydroxyl group that chemisorb to the iron oxide surface of the nanoparticles by forming a carboxylate complex with the Fe ions [17][18][19] .
Pluronic F127 is a biocompatible triblock polymer of amphiphilic nature which is composed of two hydrophilic chains of polyethylene oxide (PEO) and a hydrophobic chain of polypropylene oxide (PPO) 20 . At low temperatures and/or low concentrations in aqueous solution, PEO-PPO-PEO copolymers are present as individual unimers. By increasing copolymer concentration and/or solution temperature, thermodynamically stable micelles are formed. The corresponding critical micelle concentration (CMC) is temperature dependent 21 . To stabilize magnetic nanoparticles by coating them with pluronic F127 as used for magnetic diagnostic (i.e. magnetic resonance imaging (MRI)) and therapy (i.e. magnetic hyperthermia), pluronic F127 is normally accompanied by oleic acid and/or other polymers more than CA or in addition to CA 22,23 .
In this work, due to high biocompatibility of Zn 2+ and anticancer activity of samarium complexes 12,24 , Sm 3+ and Zn 2+ co-substituted magnetite Zn 0.1 Sm x Fe 2.9-x O 4 (x = 0.0, 0.01, 0.02, 0.03, 0.04 and 0.05) MNPs, were synthesized via co-precipitation route at 80 °C. In a simple process, these MNPs were coated by CA and pluronic F127, respectively. The effect of substitution of iron by Sm 3+ ions on the physical properties of the uncoated nanoparticles and their ferrofluids were studied.
Reduction of drug dose in all medical treatments is a goal. It is desirable to attain the target temperature in MH with as small amount of MNPs as possible to be delivered in tumors, which needs SLPs as high as possible 17 . Also synthesis of stable suspensions of large (d > 20 nm) magnetic core-shell nanostructures is another advantage 25 , which achieved in this work. Additionally maximum specific absorption rate (259 W/g) and intrinsic loss power (3.49 nHm 2 /kg) were achieved for aqueous ferrofluids, which were related to the sample with x = 0.01, in a concentration as low as 1.08 mg/ml.  3 .6H 2 O (1 M) were separately dissolved in deionized double distilled water and stirred on a magnetic stirrer at room temperature to get clear solutions, which were then mixed together (40 ml). In this way, 40 ml of 3 M NaOH solution was added into each mixture abruptly. All obtained precipitates were dark green and were heated and stirred on a magnetic stirrer at 80 °C till their hues changed to black. The dark green precipitates were included iron (II), zinc and samarium compositions, which some Fe 2+ ions were partially oxidized to Fe 3+ in presence of air oxygen by oxidizing solution. These compositions formed via separate reactions occur during the formation of Sm-Zn co-substituted magnetite nanoparticles. The obtained black precipitates were magnetic and responded to an NdFeB permanent magnet strongly. The precipitates were decanted magnetically and were washed with deionized double distilled water several times to eliminate excess ions and get a neutral pH. For powder characterization, a small amount of each washed precipitate was dried in air at room temperature. These samples were named S .00 , S .01 , S .02 , S .03 , S .04 and S .05 for x = 0.0, 0.01, 0.02, 0.03, 0.04 and 0.05, respectively. The nanoparticles were coated to reduce their toxicity and to modify their surface. Approximately 2 g of each of the washed as-precipitated samples were dispersed in 300 mL milli-Q water and sonicated for 15 min, using a FRITSCH ultrasonic bath. A CA solution (1.7 g in 25 ml of milli-Q water) was added to each solution and finally each mixture was stirred further for another 10 min. Each mixture was heated and stirred at 80 °C for According to the data obtained from the different analyses on the samples, although the nanoparticles in slurries were large in size and had strong magnetic interactions, the ferrofluids were stable in an aqueous environment (pH 5.5). Stabilization might be related to the provided thermal energy and due to steric repulsive through the polymer coating. The PPO part of the pluronic F127 adsorbed on the surfaces of the nanoparticles. The PEO parts formed a water soluble shell around the particles, generating a repulsive force for entropy reasons. Along with the CMC, we also expect that the solubility was temperature dependent. At high temperatures the value of the CMC is normally higher than at low temperatures 21,26 . Images of the ferrofluids and a schematic representation of the coating process are presented in Figs. 1 and 2, respectively. The XRD patterns of the samples were taken at room temperature using an X-ray diffractometer (Philips, X'PERT model), with Cu-Kα radiation (λ = 1.5406 Å), at a scanning rate of 0.04° per 1 s, and their full Rietveldrefined patterns were fitted by the MAUD program. The lattice constants of the uncoated nanoparticles were calculated by least-squares method 27 . The mean crystallite sizes of the nanoparticles were estimated from the broadening of the XRD peaks, using Scherrer's formula, D = 0.9λ/β cos(θ), where D is the mean crystallite size, λ is the used X-ray wavelength, θ is the Bragg angle and β is the full width at half-maximum (FWHM) intensity of the (311) peak, respectively.

Materials and methods
The morphology, particle size and size distribution of coated samples were studied using a transmission electron microscope (JEOL JEM-2100F model). The mean particle size from TEM images was calculated by Image J software. The colloidal properties of the nanoparticles in aqueous suspension (mean hydrodynamic size and polydispersity index (PDI)) were studied by dynamic light scattering (DLS, Horiba Scientific) at pH 5.5 and T = 25 °C. Fourier transform infrared (FTIR) spectra were recorded with a Jasco spectrometer (6300 model) between 4000 and 400 cm −1 . The Curie temperature of the samples was determined by DTG/M/method 28 in a weak magnetic static magnetic field. In our previous study, we elucidated the Curie temperature measurements 29 . Magnetic measurements were performed at room temperature, using a vibrating sample magnetometer (VSM) (Lake Shore Cryotronics, 7407 model) with a maximum applied magnetic field of ± 18 kOe.
To determine heating efficiency, the initial temperature slope dT/dt, an alternating magnetic field was generated in an induction coil, using a high-frequency induction machine (EFD Induction, Germany) (Fig. 3). The water-cooled coil was made of copper tube which has 3 turns and a mean diameter of 5 cm. The rms field strength and the field frequency were 16 kA/m and 290 kHz, respectively. The temperature was measured with a fiber optic thermometer with precision of 0.1 °C (FOtemp, OPTOcon, Germany) and the probe was kept in the center of the ferrofluid (Fig. 3). The nanoparticles suspensions (a 0.5 ml of ferrofluid put in a 2 ml Cryovial) were thermally isolated with polyurethane foam and placed at the center of the copper coil. The heat efficiency related to the specific loss power (SLP) defined as heat power dissipation per unit mass of nanoparticles, is expressed by: where m i and C i are the mass and specific heat capacity of each component of the magnetic ferrofluid, and m is the mass of the magnetic nanoparticles 30 . Specific heat capacities of water and magnetite are 4180 and 937 J/kg K, respectively 31 . The specific heat capacity of water was used for aqueous ferrofluids with a particle concentration lower than 2% in the suspension 30 . Since SLP is measured for different magnetic field strengths and frequencies and in different laboratories, the intrinsic loss power (ILP) is determined by using the following equation 32 : Also it can be seen that, by increasing the Sm 3+ content, the main diffraction peaks were initially shifted to lower diffraction angles, which is due to, based on Vegard's law 33   www.nature.com/scientificreports/ dral (A) and octahedral (B) sites), respectively 34,35 . The site preferences of Sm 3+ and Zn 2+ ions are octahedral and tetrahedral sites, respectively. By increasing the Sm 3+ ion content, the main XRD diffraction peaks were shifted to larger angles, due to either internal strains which can be responsible for both increase and decrease of lattice parameters 36,37 or due to redistribution of cations between A and B sites in order to relax the strain 38 . As the expansion of a lattice is not unlimited, larger rare earth ions have a limited solubility in the spinel structure. XRD patterns of the S. 01 , S .03 , FS .01 and FS .03 samples are presented in Fig. 4b. As can be seen, coating of the nanoparticles with CA and F127 did not affect the positions of the diffraction peaks. However, the intensity of the peaks decreased which may be attributed to a lower crystallinity due to the presence of the ligands on the surface of the nanoparticles. Figure 5 shows some typical Rietveld-refined XRD patterns of as-prepared samples. As can be seen, the experimental and standard data were matched very well and no unwanted phases were observed. In the Rietveld analysis, the inverse spinel has been applied. It was purposed that zinc ions were in the A sites and samarium ions were in the B sites. Mean crystallite sizes, lattice constants of the samples, refinement parameters www.nature.com/scientificreports/ and quality of the patterns, such as goodness of fit (Sig), weighted profile R-factor (R wp ), Bragg R-factor (R b ) and the expected R-factor (R exp ) are tabulated in Table 1. FTIR spectra of Zn 0.1 Sm x Fe 2.9-x O 4 nanoparticles coated with CA are illustrated in Fig. 6a. The broad band spectrum around 3411 cm −1 is attributed to the O-H band groups of water absorbed on the surface of nanoparticles. The band around 2900 cm −1 is related to the stretching vibration mode of the C-H bond. The 1720 cm −1 peak is due to the symmetric C=O stretching vibration mode from the -COOH group of CA which shifted to a lower wavenumber 1622 cm −1 . The band around 1622 cm −1 is assigned to the binding of CA radicals on the surface of nanoparticles through the chemisorption of carboxylate citrate ions. The band around 1461 cm −1 is a characteristic band of the asymmetric stretching vibration mode of CO from the carboxylic group 39 . The intense band observed around 570 cm −1 in all FTIR spectra is assigned to stretching vibrational mode of Fe-O bonds on the tetrahedral and octahedral sites 40 . As can be seen, by increasing the Sm 3+ content, this band shifted to lower wavenumbers which is attributed to the substitution by large Sm 3+ on the octahedral site which affected distances of Fe-O bonds on the octahedral sites. Therefore, the bands were ascribed to the CA-coated MNPs. CA binds to the surface of nanoparticles through the carboxylate complex with the surface Fe ions. Figure 6b shows the FTIR spectra of coated S .00 and S .01 samples with both CA and Pluronic F127, i.e., FS .00 and FS 0.01 , respectively. The band around 1100 cm −1 is attributed to C-O-C stretching vibration mode of the PPO/PPE chains of pluronic F127 41 . The presence of this band confirms the adhesion of pluronic F127 on the surface of nanoparticles. It was reported that the intensity and the position of the band are composition dependent 42 . Figure 7a and b show TEM images of FS .00 nanoparticles for two different magnifications. As can be seen, nearly hexagonal and irregularly shaped were formed. Figure 7c shows the histogram of particle size distribution for the FS .00 sample, yielding an average of 40 nm. As the thicknesses of their shells were not clearly observable, we did not consider its thickness in averaging.  Table 2. According to Néel theory for two sublattice ferrimagnetism, each sublattice is ordered ferromangetically and has a nonzero spontaneous magnetization and magnetizations of the two sublattices are coupled antiferromagnetically. Then, the magnitude of the net magnetization is obtained from: Figure 5. The Rietveld-refined XRD patterns of as-prepared specimens as marked on the patterns. Table 1. The lattice constants, the mean crystallite sizes and Rietveld-refined XRD data of the as-prepared Sm 3+ substituted nanoparticles.

Sample a ± 0.001(Å) D ± 2(nm) Sig (GoF) R wp (%) R b (%) R exp (%)
S. 00  www.nature.com/scientificreports/  www.nature.com/scientificreports/ where M A , M B and T are magnetizations of A, B sublattices and temperature, respectively. As can be seen, by increasing the content of Sm 3+ , the saturation magnetization gradually decreased which is due to replacement of Fe 3+ (5 μ B ) by Sm 3+ (1.5 μ B ) ions 43 in B sites. Therefore, the number of magnetic moments of B sites was reduced, while the magnetization of the A sites remained constant. As can be seen, by increasing the content of Sm 3+ the remanent to saturation magnetization ratio decreased slightly. According to the model of non-interacting randomly distributed uniaxial single domain particles, the M r /M s ratio of the particles is 0.5. Any deviation from this value is indicating a transition to multidomain particles 44 . Table 2 illustrates that the M r /M s ratios of the uncoated samples were less than 0.1, indicating that the particles are multidomains. In order to estimate the effective anisotropy constants K eff of the samples, an empirical equation, based on the law of approach to saturation, was used:   45,46 . Figure 8c shows the variation of magnetization versus magnetic field for high fields that was fitted to M H = M s (T)[1-(a/H)-(b/H 2 )] + cH. All experimental data were fitted very well with R-squared values higher than 99%. As can be seen, by increasing the Sm 3+ content the effective anisotropy constant K eff decreased, which can be attributed to the effect of particle size. By increasing the Sm 3+ content, a correlation between H c and K eff is observed, which is attributed to the direct relationship between H c and K eff 46 : Table 2 illustrates that coercivity of the uncoated samples decreased as the Sm 3+ content increased. Coercivity is dependent on the displacement of domain walls and on the particle sizes 7 . Coercivity also is impressed by other factors such as microstrain and magnetocrystalline anisotropy or interactions due to packing density, which is controlled from the coating 36,47 .
According to Berkov's theory for low anisotropy, coercivity first increases with increasing concentration when dipolar interaction becomes significance, but for more dense packing coercivity decreases again due to the formation of a short range order. Possibly, in the densely packed samples particles come locally in closer contact and therefore a coupling due to exchange interactions may become dominant, so some complicated magnetization structures arise which were not considered in Berkov's model 47 . Figure 9 illustrates magnetization-temperature (M-T) curves of the as-prepared samples in the course of first heating and cooling processes. As can be seen, by increasing the Sm 3+ content, the Curie temperature of the samples increased gradually from 430 °C to 480 °C. The Curie temperature is related to the number and the strength of magnetic interactions (A-A, B-B and A-B super-exchange interactions) as well as distance between paramagnetic ions 48 . In spinel ferrites, the interaction between Fe 3+ ions in A and B sites is the strongest one and thus plays a dominant role in determining T c 7 . There is a weak exchange interaction between the 4f. electrons of Sm 3+ and the 3d electrons of Fe 3+ , but A-O-B super-exchange interactions between Fe 3+ -Fe 3+ is stronger than that of Fe 3+ -Sm 3+ . Furthermore, by increasing Sm 3+ content, electron hopping between Fe 3+ and Fe 2+ ions in A and B sites occurs, which leads to a decrease in A-B super-exchange interaction 7 . Therefore, for a low concentration of Sm 3+ ions, the lattice constant decreased as the Sm 3+ content increased, A-B super-exchange interactions in the samples increased, all resulting in an increase in T c . As M-T curves were recorded in air, samples were oxidized during the determination of the Curie temperatures. The oxidation reaction of substituted magnetite (M z 2+ Fe 1-z 2+ Fe 2 3+ O 4 2-, where M 2+ is a bivalent cation or a combination of cations with different valences so that their net valences are two or Fe 2+ Fe 2-x 3+ M x 3+ O 4 2− , where M 3+ is a trivalent cation or a combination of cations with net valence of three) is a topotactic reaction where the spinel structure is preserved. A metastable defect γ phase structure forms, which can be described by the following formula: where y = 4z/(9-z) and, where y = x/3, 0 < y < 2/3.
By increasing the temperature, the cubic γ phase transforms into a stable rhombohedral hematite phase. Accordingly, the sample decomposes to a spinel and hematite structure [49][50][51] .  During cooling from high temperatures to room temperature, magnetization showed a temperature dependency, which is due to the spinel phase with a new cation rearrangement based on ions preference energies which has ferrimagnetic order. The samples decomposed to hematite and a spinel phase. The Curie temperatures for all samples are summarized in Table 2.
Colloidal properties. Figure 10 shows DLS diagrams of the ferrofluids. The mean hydrodynamic sizes of the FS. 00 , FS. 01 , FS .02 , FS .03 , FS .04 and FS .05 coated nanoparticles, which were dispersed in an aqueous medium, at T = 25 °C and at pH = 5.5 were 582, 489, 491, 518, 445 and 433 nm, respectively. These sizes were much larger than those of both crystallite and particle, which indicates that the magnetic nanoparticles were agglomerated. The hydrodynamic size of the nanoparticles depends on their interactions and the numbers of polymers attached on their surfaces 52 . However, even at concentrations below the CMC, amphiphilic molecules attach onto the surfaces of nanoparticles, depending on their surface properties. For concentrations of the pluronic F127 in water above its CMC, micelles form. Possibly absence of micelles of pluronic F127 at T = 25 °C, the micelle passes its CMC, led to agglomeration and thus to a large hydrodynamic size.
Calorimetrical and magnetic measurements on ferrofluids in an aqueous medium. To consider contributions of Néel and Brownian relaxation mechanisms for heat generation, we estimated the critical sizes as well as the effective anisotropy constant of the samples. For Néel relaxation, it is assumed that magnetic moments rotate, while the crystal structure is fixed in space. The Néel time constant is defined as: where τ 0 = 10 −9 s, k B is the Boltzmann's constant (1.38 × 10 −23 J/K), T is the absolute temperature, V is the particle volume and K eff is effective magnetic anisotropy constant, which may originate from magnetocrystalline, shape and other anisotropies. For the Brownian relaxation mechanism, it is supposed that the magnetic moment is locked to the crystal structure. The magnetic moment rotates in a low viscosity carrier medium, when it aligns with the applied field.
The Brownian time constant is: where η is the viscosity of the carrier liquid, which is 1.01 × 10 −3 kg/sm for water and V H is the hydrodynamic volume 53,54 .
In presence of a magnetic field varying in time, the Brownian relaxation mechanism results in a heat generation in a ferrofluid, as a consequence of viscous friction between particles and the surrounding carrier liquid, which is not limited to superparamagnetic particles 55 .
We presume that magnetocrystalline anisotropy prevails in effective magnetic anisotropy of the nanoparticles. At a high frequency, the critical size is defined by ωτ = 1 56 . In the critical size region the hysteresis loss vanishes abruptly and relaxation effects form remaining loss mechanisms 57 . The estimated effective anisotropy constants (8)  www.nature.com/scientificreports/ of the samples are given in Table 2. For the applied frequency (290 kHz) the critical particle size of the samples were found to be in the range from 19 to 22 nm, where p Néel achieved its maximum. The produced nanoparticles have larger average sizes than the critical size. Therefore, high SLP values are due to nonzero coercivity, i. e., due to hysteresis loss. The hydrodynamic size for our measuring frequency was 11 nm for ωτ B = 1. The hydrodynamic sizes of the prepared samples are marked in Fig. 10, which are very higher than those obtained from the hydrodynamic volumes (V H ), estimated from Eq. (9). Figure 11 shows room temperature VSM curves, which were carried out on the aqueous ferrofluids (FS .01 and FS .02 ). As can be seen although magnetizations of the samples were saturated, but they are so small, which is due to very low concentrations. Saturation magnetization of a ferrofluid in an aqueous medium not only depends on the single cores M S , but depends on concentration. Also VSM data reviled that both ferrofluids have moderate coercivities and then no superparamagnetic behavior, which is due to slightly agglomeration in aqueous medium. As discussed above the same agglomeration was observed by DLS measurements too. So we can deduce that relaxation processes did not play a dominant role in SLP. In an external RF magnetic field, magnetic materials convert the incident RF energy into heat 58 . The heating efficiency of the magnetic ferrofluids was obtained using SLP (W/g) = dT/dt × Σ m i C i /m. The calculated SLP and ILP values are summarized in Table 3. The maximum SLP and ILP were found to be 259 W/g and 3.49 nHm 2 /kg, respectively, for the sample with x = 0.01. The SLP depends on magnetic field parameters and properties of the nanoparticles, such as size, size distribution, saturation magnetization M S , remanent magnetization M R , coercivity H C and effective magnetic anisotropy as well as the concentration of the sample. Hysteresis loss represents the main portion in heat generation of multidomain magnetic materials, while relaxation processes represent the main contribution for single domain superparamagnetic particles 17 . Therefore the high SLP values of the FS .01 and FS .02 samples can be attributed to hysteresis loss first and Brownian mechanism second, due to viscous friction between rotating particles and surrounding aqueous medium. Figure 12a and b illustrate the variation of the temperature in terms of time (T-t) curves for x = 0.01 and 0.02 in an RF magnetic field. According to data from Table 3 and the temperature variations (ΔT) with respect to time, the FS .01 and FS .02 could potentially be good candidates for magnetic hyperthermia application. Gadzhimagomedova et al. reported SLP and ILP values of 8.2 W/g and 0.15 nHm 2 /kg, respectively for samarium doped superparamagnetic magnetite nanoparticles (Sm 0.033 Fe 2.967 O 4 ) coated with PEG 59 , which are very lower than those we obtained in this work, Table 3.

Conclusions
Single phase Sm 3+ and Zn 2+ co-substituted magnetite (Zn 0.1 Sm x Fe 2.9-x O 4 , x = 0.01, 0.02, 0.03, 0.04 and 0.05) nanoparticles were synthesized via co-precipitation method. The comparatively large (d > 20 nm) nanoparticles were coated with citric acid and pluronic F127, using a simple route to obtain core-shell structures and then suspended in water to get stable ferrofluids. The stability of the ferrofluids (pH 5.5) was related to the formation of micelles. The PPO sequence of the pluronic F127 adsorbed on the surface of the nanoparticles, while the PEO chains formed hydrophilic shells around the nanoparticles, which in turn generated repulsive forces due to entropy reasons. The coercivity, (M r /M s ) ratio and the effective anisotropy constant K eff were monotonically  www.nature.com/scientificreports/ decreased by increasing the Sm 3+ content, which related to properties of the nanoparticles, such as size, size distribution and morphology. VSM measurements on all ferrofluids presented nonzero coercivities, which are results of both multidomain cores and agglomeration of the coated nanoparticles in the solution. The highest obtained values of the SLP and ILP were 259 W/g and 3.49 nHm 2 /kg, respectively, which were found for the FS .01 sample at a concentration as low as 1.08 mg/ml. Those high values are due to the nonzero coercivity of the samples that leads to a hysteresis loss portion as the main loss mechanism, although frictional loss may have occurred additionally too. From the SLP and ILP data, we concluded that both FS .01 and/or FS .02 samples are good candidates for magnetic hyperthermia applications potentially to achieve sufficient heating efficiencies at a low magnetic nanoparticles concentration. www.nature.com/scientificreports/