CoFe₂O₄ nanoparticles synthesized with natural templates

Abstract

CoFe₂O₄ nanoparticles were synthesized using a simple precipitation method in the presence of three natural templates: flax, gauze and sisal. Each preparation gave rise to two different types of samples. The structural, morphological and magnetic properties of the nanoparticles were characterized, and their usefulness as nano-heaters in magnetic hyperthermia was explored. The specific loss powers obtained for all CoFe₂O₄ nanoparticles in water shared similar magnitudes and were attributed to a dominant Brown relaxation. The nanoparticles obtained after template calcination at 550 °C display morphologies determined by the threaded templates, with enhanced magnetic anisotropy and associated magnetic coercivities similar to that of nanofibers or nanorods.

---

Introduction

Magnetic nanostructured materials have a wide range of applications both in nanodevices, magnetic information storage media^1–4 and, more recently, as an important component for electromagnetic radiation absorbers.⁵ In addition to the important technological advances mentioned, magnetic nanoparticles are currently a major tool for biomedical applications such as drug delivery, magnetic particle imaging (MPI), magnetic resonance imaging (MRI) and magnetic hyperthermia.^6–9 The different biomedical applications explore different interactions with the magnetic nanoparticles (MNP) and, consequently, require different particles. For instance, the best particles for MPI are superparamagnetic particles¹⁰ and for MRI the MNP should be characterized by a high magnetic moment,¹¹ but for magnetic hyperthermia the central requirement is a high magnetic loss. One way to tune the magnetic behaviour of the nanoparticles (NP), besides changing their composition, is to control their size, shape and structure, a route that has been explored by using different methods of preparation and processing. One such example is morphology control of the β-FeOOH nanostructures for pH-controlled anti-cancer drug delivery using a simple gelatine assisted hydrothermal method, in which gelatine acts as a surfactant.¹² Also, magnetite nanoparticles with different shapes, such as nanoflowers and nanocubes have been synthesized for optimizing magnetic hyperthermia efficiency.^13–15

For application in magnetic hyperthermia, magnetite has been the most explored magnetic iron oxide compound, but nanoparticles of CoFe₂O₄ and other ferrites have also been investigated.^9,16,17 Cobalt ferrite is ferrimagnetic with a high saturation magnetization and high magnetic anisotropy. These properties make CoFe₂O₄ nanoparticles good candidates for magnetic hyperthermia since anisotropic magnetic NP are expected to be associated with high heating efficiency.

Magnetic anisotropy depends on particle size and shape as well as on their interactions. Consequently, the magnetic properties can be tailored by acting on the individual NP or by synthesizing nanostructured aggregates with a different geometrical arrangement of the constituent NP. New synthesis methods have been explored to enhance magnetic anisotropy by controlling the CoFe₂O₄ nanoaggregates size and morphology.^18–20 Highly anisotropic structures such as nanowires prepared using an anodic alumina template and nanofibers produced by electrospinning have been successfully obtained.^4,20

In this work, we explore a simple method to obtain anisotropic CoFe₂O₄ NP: the synthesis of CoFe₂O₄ in the presence of natural fibers. This method was already tested in a recent work²¹ using medicinal cotton with promising results. An analogous synthesis methodology using cotton and sponge as templates has been reported to influence the saturation magnetization and coercivity of NiFe₂O₄.²²

In order to confirm the role of the template and study its influence on the NP properties, three different fibers, flax, gauze (aligned cotton threads) and sisal, were explored for the synthesis of CoFe₂O₄ nanoparticles, and the physical properties of the nanoparticles were investigated.

Experimental

The CoFe₂O₄ nanoparticles were synthesized by a standard precipitation method in a gelatine medium, followed by hydrothermal treatment^21,23 in the presence of three different natural fiber templates: flax, gauze and sisal.

Stoichiometric amounts of iron nitrate (Fe(NO₃)₃·9H₂O) and cobalt nitrate (Co(NO₃)₃·6H₂O) from Sigma-Aldrich were dissolved in distilled water. After complete dissolution of the salts, 40 mL of a gelatine solution was added, followed by the addition of an adequate volume of ammonia solution (25%, Scharlau), leading to the formation of a brownish precipitate. The mixture was transferred to an autoclave where the template (around 3.0 g) was immersed and then placed in an oven at 150 °C for 3 hours. After cooling down to room temperature, the mixture consisted of a wet dark template immersed in a liquid suspension containing nanoparticles. The dark template was isolated, and the nanoparticles in the liquid suspension were separated by filtration. The synthesis procedure branches at this point, as schematically presented in Fig. 1, to give two different batches of nanoparticles: the ones obtained directly from filtration of the suspension and those impregnating the template. The two products, template and filtered nanoparticles, were separately washed with distilled water until a neutral pH was obtained. The nanoparticles obtained from the filtration procedure were dried at 60 °C. The template with the embedded nanoparticles was heated at 550 °C for 6 hours in air to eliminate the template. These calcination conditions were selected after a detailed study for each template, being the ones that provided elimination of the template (less than 1% residue).

Fig. 1 Scheme of the synthesis method after precipitation, showing the branching of the process after thermal treatment. This methodology gave rise to two different batches of CoFe₂O₄ samples: nanoparticles obtained from solution filtration (S-NP) and nanoparticles obtained after template calcination (T-NP). The same procedure was used for the three templates: sisal, flax and gauze.

The structural characterization of the CoFe₂O₄ nanoparticles was carried out by powder X-ray diffraction (XRD) with a Philips Analytical PW 3050/60 X'Pert PRO diffractometer equipped with an X'Celerator detector and using CuKα radiation as the incident beam from 10 to 80° (2θ) with a scan step time of 40 s. The average crystallite sizes were calculated from the broadening of the X-ray diffraction peaks using the Scherrer formula.

Fourier transform infrared (FTIR) spectra were recorded on a Shimadzu, IRAffinity-1 spectrometer with samples dispersed in KBr pellets. The spectra were collected in the wavenumber range 400–4000 cm⁻¹, with a resolution of 4 cm⁻¹.

The morphology of the templates and synthesized samples was investigated by Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy (TEM) using a JEOL 7001F and Hitachi 8100 with digital image acquisition microscopes, respectively. The particles size distribution was determined from the TEM images using the program ImageJ, which measured the size for around 100 nanoparticles for each sample.

The magnetic characterization of the nanoparticles was performed using a SQUID magnetometer (QD-MPMS) and ⁵⁷Fe Mössbauer spectroscopy. Magnetization measurements were carried out as a function of temperature between 10 K and 400 K at 2.0 mT after cooling the sample from room temperature in zero magnetic field (zero field cooled – ZFC) and after cooling under the measurement field (field cooled – FC). Hysteresis curves were obtained at different temperatures for applied magnetic fields up to 5.5 T. The ⁵⁷Fe Mössbauer spectra were collected at room temperature in transmission mode using a conventional constant-acceleration spectrometer and a 50 mCi ⁵⁷Co source in a Rh matrix. The velocity scale was calibrated using an α-Fe foil. The spectra were fitted to Lorentzian lines considering distributions of the magnetic hyperfine field using the WinNormos Program.

To check their suitability for magnetic hyperthermia applications, induction heating of the nanoparticles was evaluated using an Easy Heat 0224 device (Ambrell) under an AC field of 13.9 kA m⁻¹ amplitude and 274 kHz frequency, applied for periods of 100 s, using the methodology and experimental set-up previously described in detail.¹² For these measurements, 30 mg of each sample were dispersed in 3 mL of distilled water or paraffin (in glass vials), and the liquid or solid mixture was placed inside the magnetic field coils, separated by isolating materials. The temperature inside the vials was monitored during the process using a Photon Control fiber optic temperature sensor. In the case of nanoparticles immobilized in paraffin, the solid samples were obtained after ultrasonic dispersion in liquid paraffin followed by rapid cooling to room temperature.

Results and discussion

Structural and morphological characterization

The synthesis methodology produced two sets of CoFe₂O₄ nanoparticles from each preparation: those obtained from the liquid suspension, referred as S-NP, and those obtained after template calcination, referred as T-NP (Fig. 1).

Regarding the S-NP, the powder X-ray diffraction patterns revealed the formation of a CoFe₂O₄ spinel type structure (Fig. 2), with similar mean particle sizes in all cases (∼4.5 nm estimated by the Scherrer equation). These nanoparticles are smaller than the ones obtained with cotton as template (7 nm), which were prepared with the same methodology but without using the gelatinous medium during the synthesis.²¹ The difference in NP size is attributed to the effective role of the gelatine medium in controlling the particle size distribution, as reported for magnetite NP synthesis.²³

Fig. 2 XRD patterns of the nanoparticles obtained from solution filtration (S-NP) with different templates. All patterns exhibit the characteristic peaks of the spinel structure.

The powder X-ray diffraction patterns of the samples obtained after template calcination (T-NP) are presented in Fig. 3 and show that the CoFe₂O₄ spinel type structure was maintained for all samples. When compared with the patterns for the S-NP (Fig. 2), it is clear that T-NP samples display a higher crystallinity and larger mean particle size, which was calculated to be 21 nm when flax and gauze were used and 15 nm in the case of sisal. For the CoFe₂O₄ NP obtained using cotton, a slightly higher mean particle size (24 nm) was reported.²¹

Fig. 3 XRD patterns of the samples obtained after template elimination (T-NP) for the different templates. All patterns exhibit the characteristic peaks of the spinel structure.

FTIR spectra of the samples obtained after template calcination are presented in Fig. 4. All samples show the characteristic bands of the spinel structure below 580 cm⁻¹, corresponding to the stretching vibration of the tetrahedral and octahedral sites. The well-defined band between 550 and 580 cm⁻¹ is assigned to the intrinsic vibrations of the tetrahedral sites (M–O). The band corresponding to the octahedral sites is observed below 400 cm⁻¹.²⁴ No other bands are identified in the FTIR spectra of the T-NP samples, confirming total elimination of the templates. The spectrum of a sample obtained from solution filtration is also included in Fig. 4. The characteristic bands of the spinel structure are also observed, as well as bands between 3250 and 3500 cm⁻¹ and around 1625 cm⁻¹, which are assigned to absorbed water molecules on the surface of the nanoparticles.

Fig. 4 FTIR spectra of the CoFe₂O₄ nanoparticles obtained after template calcination and one sample obtained from solution filtration.

Representative TEM results of the CoFe₂O₄ S-NP (with the corresponding size distribution histograms) are presented in Fig. 5, and images from the T-NP are illustrated in Fig. 6. CoFe₂O₄ S-NP are homogeneous, with mean sizes 3.8(1) and 3.9(1) for sisal and gauze, respectively, and 4.5(1) for flax. These values are compatible with those determined by XRD data (∼4.5 nm). In the case of the T-NP, the images show that instead of individual particles, the samples consisted of coalesced nanoparticles, with smaller agglomerates in the case of sisal.

Fig. 5 Representative TEM images (left) and the corresponding size distribution histograms (right) of CoFe₂O₄ S-NP obtained from solution using the different templates (sisal, flax and gauze).

Fig. 6 Representative TEM images of CoFe₂O₄ T-NP obtained after template elimination. The two amplifications clarify the differences in the sample morphology.

TEM images show that the particles obtained from calcined flax and gauze reflect the template's thread morphology, showing sintered elongated shape particles. However, the ones obtained after sisal calcination do not show a threaded morphology and are organized in smaller aggregates with smaller NP. In order to understand the effect of the templates in the morphology of the products, SEM images were also obtained for each template before use and compared with SEM images of CoFe₂O₄ NP obtained after template calcination (Fig. 7). These images show that while flax and gauze have similar morphologies, consisting of threads with ∼10 μm diameter, sisal has a more compact morphology with rough, thicker fibers with diameters larger than 100 μm. The fiber morphology explains the different aspects of the particles observed on the TEM images of CoFe₂O₄ T-NP obtained with sisal when compared with the two other templates. The thicker sisal fibers require a very large number of particles to be covered, explaining why such a structure is not observed (Fig. 6). The formation of smaller T-NP in the case of sisal (average diameter ∼ 15 nm) when compared with T-NP from the other templates (∼21 nm) is attributed to the significant roughness of the sisal fibers.

Fig. 7 SEM images of the templates before use (left) taken with amplification ×400 and ×10 000 (inset) and of the CoFe₂O₄ T-NP samples obtained after template calcination (right) at amplification ×20 000.

Magnetic characterization

All CoFe₂O₄ nanoparticles were characterized by ⁵⁷Fe Mössbauer spectroscopy and SQUID magnetometry.

The Mössbauer spectra of the CoFe₂O₄ S-NP obtained from solution are shown in Fig. 8, and a representative spectrum of the calcined T-NP (identical for the three samples) is shown in Fig. 9. The Mössbauer parameters obtained from the experimental data analysis are listed in Table 1.

Fig. 8 ⁵⁷Fe Mössbauer spectra collected at 290 K for CoFe₂O₄ S-NP obtained from solution (left), showing that each spectrum is well fitted using two sub-spectra associated with two different iron sites. The distribution of the magnetic hyperfine field, B_hf, of each sub-spectrum is also shown (right).

Fig. 9 ⁵⁷Fe Mössbauer spectra collected at 290 K for the CoFe₂O₄ NP obtained after gauze calcination (left), showing that it is well fitted using two sub-spectra associated with two different iron sites. One sub-spectrum required a distribution of the magnetic hyperfine field, B_hf, shown in the right figure.

Table 1 Mössbauer hyperfine fitting parameters for all CoFe₂O₄ nanoparticles (S-NP and T-NP). B_hf: magnetic hyperfine field; 〈B_hf[σ]〉: average magnetic hyperfine field and standard deviation of the distribution of magnetic splittings; δ: isomer shift; ε: quadrupole shift; Γ: line width; I: relative area. Uncertainties in I are less than 2%

CoFe2O4	Template	Site	Bhf/T	〈Bhf[σ]〉/T	δ/(mm s^−1)	ε/(mm s^−1)	Γ/(mm s^−1)	I/%
From solution (S-NP)	Gauze	1	—	30[15]	0.24(2)	0.08(3)	0.38	49.1
		2	—	32[16]	0.39(2)	−0.05(2)	0.38	50.9
	Flax	1	—	30[16]	0.19(4)	0.03(4)	0.38	38.1
		2	—	28[16]	0.41(3)	0.00(3)	0.38	61.9
	Sisal	1	—	32[15]	0.21(2)	0.06(4)	0.38	37.8
		2	—	31[16]	0.39(2)	−0.04(2)	0.38	62.2
After template elimination (T-NP)	Gauze	1	48.4(2)	—	0.24(1)	−0.00(1)	0.43(1)	34.1
		2	—	46[7]	0.37(1)	0.04(1)	0.36	65.9
	Flax	1	48.7(1)	—	0.24(1)	0.01(1)	0.44(1)	34.5
		2	—	46[8]	0.37(1)	0.02(1)	0.37	65.5
	Sisal	1	48.2(1)		0.24(1)	0.00(1)	0.46(1)	32.8
		2	—	46[7]	0.36(1)	0.03(1)	0.33	67.2

---

The spectra of the S-NP samples are quite different from those produced after template calcination, as expected because of the different mean particle sizes. For the S-NP (Fig. 8), all spectra required two distributions of the magnetic hyperfine field, B_hf, to be properly fitted: one with a lower isomer shift value (δ), indicating the presence of Fe³⁺ in tetrahedral sites, and the other one with a higher δ value, associated with Fe³⁺ ions in octahedral sites. The magnetic field distributions are broad, typical of very small nanoparticles (∼4 nm, as deduced from TEM), and include important fractions of NP with zero or very low magnetic field values.

The three spectra corresponding to the T-NP samples are similar. As can be seen from Fig. 9, where the spectrum for the NP obtained after gauze calcination is shown, and from Table 1, the spectra are well resolved by two sextets, both assigned to Fe³⁺, one with δ = 0.24 mm s⁻¹ (site 1), which is well defined and attributed to iron in tetrahedral sites, and the other with δ = 0.37 mm s⁻¹ (site 2), which is distributed and assigned to iron in the octahedral sites of the AB₂O₄ spinel structure. The relative amounts obtained for octahedral and tetrahedral sites are almost identical for the three samples, showing that the template has no role in the cation's distribution. There is no evidence of Fe²⁺ ions in the three samples. Therefore, assuming the correct stoichiometry of the spinel phase (CoFe₂O₄), cobalt ions remain as Co²⁺. From the relative areas of each iron site (approximately 34% and 66% for the three samples), and assuming the same f value (Lamb–Mössbauer factor) for both sites, the following average occupancies can be deduced: (Fe_0.68Co_0.32)_tet[Fe_1.32Co_0.68]_octO₄, with tet and oct referring to tetrahedral and octahedral sites, respectively. This formulation is similar to that previously reported for 8 nm CoFe₂O₄ nanoparticles obtained by a coprecipitation method and heated at high temperature for 2 hours.²⁵ A slightly different distribution was reported for 5.5 nm and 6 nm CoFe₂O₄ nanoparticles synthesized by forced hydrolysis in polyol²⁶ and thermal decomposition,²⁷ respectively. Thus, it seems that the synthesis method also has a marked impact on the cation distribution in the spinel lattice of CoFe₂O₄ nanoparticles, as previously stated.²⁸ Nevertheless, the deduced distribution for the samples obtained after template calcination (performed at 550 °C) corresponds to an almost statistical distribution of Co²⁺ and Fe³⁺ in tetrahedral and octahedral sites, showing a high inversion degree of the spinel structure.

The obtained distribution is slightly different from that reported for the NP obtained with a similar methodology but using cotton,²¹ where a higher number of Co²⁺ was found in octahedral sites. As previously mentioned, the synthesis of the NP in the presence of sisal, flax and gauze was performed in a gelatine medium. This medium has a positive role in the synthesis, leading to a narrow particle size distribution and low oxidation degree of magnetite nanoparticles.²³ Thus, the different distributions of iron between octahedral and tetrahedral sites found in this study is attributed to the use of gelatine during synthesis, affecting the cation's distribution in the spinel structure.

The magnetization results for the CoFe₂O₄ nanoparticles obtained from the solution (S-NP) are shown in Fig. 10. The ZFC/FC curves show the expected behaviour for a system of magnetic nanoparticles, small enough to be considered as single domain NP. For high temperatures, the thermal energy is high, and the particles can overcome the energy barrier that separates different magnetic moment orientations, being in a superparamagnetic state. At low temperatures, the particles remain in a blocked state. The temperature separating the superparamagnetic from the blocked state is called the blocking temperature (T_B). The energy barrier, E, associated with the blocking temperature depends on the NP volume, V, and on the magnetic anisotropy of the particle and is written as a function of the effective magnetic anisotropy constant, K_eff, as E = K_effV. The ZFC curves display maxima just below the temperature at which the ZFC and FC curves overlap. These maxima define the NP blocking temperature (T_B) and are around 250 K for the sample obtained with flax and close to 290 K for samples prepared with sisal and gauze. The position of the ZFC maxima is changed by dipolar interactions, shifting to higher temperatures with increasing strength.²⁹ However, this influence is less marked in high anisotropy materials like CoFe₂O₄. Since the measurements were carried out in powder samples, where dipolar interactions are not negligible, the values obtained for the maxima can be considered the higher limit for the blocking temperature of each sample. From TEM results, the average diameter for the S-NP obtained with flax is higher than the average diameter for S-NP obtained with gauze or sisal, suggesting a higher blocking temperature for flax. However, the experimental results are the opposite, as T_B for flax S-NP is lower than the T_B value for the other two samples by 14%. Consequently, the difference in size does not explain the relationship among the blocking temperatures of the three S-NP distributions, implying a smaller effective anisotropy constant for S-NP from flax since dipolar interactions are expected to be similar for the three S-NP with similar distributions and measured in the same conditions.

Fig. 10 Temperature dependence of the magnetization (top) and hysteresis curves at 300 K (bottom) for the CoFe₂O₄ S-NP samples in the presence of gauze, flax and sisal.

The effective anisotropy constant can be obtained from the relationship between the energy barrier and blocking temperature, where the logarithm of the ratio between the characteristic time of measurement, τ, and the period of oscillation of the NP, τ₀, is on the order of 30 for SQUID measurements.³⁰ Using the experimental values for T_B and the NP volumes calculated from the mean sizes determined by TEM, the values 2.2 × 10⁶ J m⁻³, 3.9 × 10⁶ J m⁻³ and 4.2 × 10⁶ J m⁻³ were obtained for K_eff(flax), K_eff(gauze) and K_eff(sisal), respectively. The different values can be explained by an important surface contribution to the effective anisotropy for small NP. An estimation of this influence for one cobalt ferrite nanoparticle 4 nm in diameter, considering a surface layer thickness of 0.4 nm (half of the ferrite lattice parameter assumed as the average diameter of CoFe₂O₄ molecule), leads to the conclusion that the number of atoms at the surface is about the same as the number in the inner volume of the particle. However, this value greatly increases if the surface is more than one molecule thick (for a double thickness, it would be around 4 times the number of particles in the inner volume). Therefore, the differences in T_B can be explained by the increase of the surface contribution to K_eff when the NP size decreases, attaining a higher value for the smaller NP.

The relevant magnetic parameters for the S-NP are summarized in Table 2. At 300 K, the three samples display similar saturation magnetization values (M_s), approximately 55% of the saturation magnetization of bulk CoFe₂O₄ (81 A m² kg⁻¹ at 300 K),³¹ and small coercive field values (μ₀H_c). Because of the small size of these NP (3.8–4.5 nm) and the T_B values observed, these S-NP should be superparamagnetic at 300 K. The non-zero coercivity measured at this temperature is attributed to a remaining small fraction of blocked NP at 300 K.

Table 2 Magnetic parameters at 300 K for the CoFe₂O₄ obtained from the solution (S-NP). M_s: saturation magnetization; μ₀H_c: coercive field

Template	Ms/(A m^2 kg^−1)	μ0Hc/mT
Gauze	43	0.5
Flax	40	1.2
Sisal	47	0.5

---

For the nanoparticles obtained after template calcination (T-NP), the magnetization results (Fig. 11 and Table 3) are consistent with a ferrimagnetic behaviour, as expected since bulk CoFe₂O₄ is ferrimagnetic. The saturation magnetization approaches the value of bulk CoFe₂O₄, being 77% of this value at 300 K. The coercive field (μ₀H_c) values of these nanoparticles (T-NP) are on the order of 70 mT at room temperature, much higher than the ones for the S-NP (Table 2), demonstrating the influence of the template in increasing nanoparticle coercivity. The high coercivity is a consequence of a higher magnetic anisotropy,³² associated with the morphology of the samples after template calcination. At room temperature, the coercivity values are comparable to the ones reported for nanorods,³³ nanowires^4,34 and nanoribbons,³⁵ reaching 1.2 T to 1.3 T at 10 K. The shape of the hysteresis curves at 10 K also indicate the coexistence of two contributions with different coercivities; the component with a lower coercivity being more important in the case of sisal.

Fig. 11 Temperature dependence of the magnetization (top) and hysteresis curves at 300 K (middle) and 10 K (bottom) for the CoFe₂O₄ nanoparticles obtained after each template calcination (T-NP).

Table 3 Magnetic parameters at 10 K and 300 K for the CoFe₂O₄ nanoparticles obtained after template elimination (T-NP). M_s: saturation magnetization; μ₀H_c: coercive field

	Template	Ms/(A m^2 kg^−1)	μ0Hc/T
10 K	Gauze	72	1.3
	Flax	68	1.2
	Sisal	64	1.2
300 K	Gauze	67	0.090
	Flax	62	0.086
	Sisal	60	0.068

---

Magnetic hyperthermia

The magnetic hyperthermia performance of all samples was evaluated by induction heating, as previously described in the experimental part. The temperature evolution for the first 100 s was fitted considering a constant magnetic heating power supplied by the nanoparticles and assuming linear energy exchanges between samples and environment through conduction and convection processes in the surrounding materials.²¹

Several measurements were performed for each sample, and the mean specific loss power (SLP) was evaluated. The results are summarized in Table 4. To evaluate magnetic losses, minor hysteresis cycles were performed for a maximum DC magnetic field that equals the AC field amplitude used in the hyperthermia experiments. The area, A, of the hysteresis cycle was determined, and the hysteresis losses power (HLP) was estimated as HLP = fA, with f being the frequency of the AC magnetic field. To evaluate the contribution of the rotation of the magnetic moments within the particles (Néel relaxation or hysteresis losses) and the contribution of viscous dissipation due to rotation of the NP in water (Brown relaxation), dispersions of the NP in water and in solid paraffin were used. The results indicated that the SLP values drastically decrease when the particles are immobilized in paraffin (also observed for NP with higher average sizes⁹) and the SLP values are almost zero in the case of the samples produced after template calcination. Therefore, for the experimental conditions used, Brown relaxation is the predominant mechanism for the magnetic heat release of all produced CoFe₂O₄ NP dispersed in water, being higher for the sample with larger nanoparticles (flax).

Table 4 Mean specific loss power (SLP) values for all NP, obtained from fitting the experimental temperature variation and magnetic parameters extracted from minor hysteresis loops performed at 300 K between ±20 mT: H_c coercivity; HLP hysteresis loss power

		Hc/mT	HLP/(W g^−1)	SLP/(W g^−1)
				Water	Paraffin
From solution (S-NP)	Gauze	0.0	0.2	6	2.3
	Flax	0.0	0.4	8	2.1
	Sisal	0.0	0.0	5	2.2
After template calcination (T-NP)	Gauze	1.2	0.6	8	0.6
	Flax	1.6	0.4	6	0.6
	Sisal	0.4	0.0	7	0.5

---

The Néel relaxation time was calculated for different effective anisotropy constants (K_bulk (2.0 × 10⁵ J m⁻³),¹⁸ 5K_bulk and 10K_bulk) using and the Brown relaxation time was determined in water using τ_B = 4πηR_H³/k_BT, considering different hydrodynamic radii, R_H, and using 8.90 × 10⁻⁴ Pa s for water viscosity, η, at 25 °C.³⁶ The logarithm of the corresponding frequencies (log(1/τ)) as a function of the NP diameter is plotted in Fig. 12. In this figure, the Néel relaxation frequency is plotted for three different values of the anisotropy constant (referred to as the bulk value of cobalt ferrite), and the Brown relaxation frequency is calculated assuming three different hydrodynamic radii, chosen to vary between the nanoparticle radius and about three times the radius of the NP produced in this work. Assuming that the relaxation mechanisms are independent, the relaxation of the NP characterized by a certain diameter is dominated by the process corresponding to the higher frequency. According to the plotted results, the predominance of the Brown relaxation process over Néel relaxation for NP with diameters between 3.8 nm and 4.5 nm (shaded region) requires K_eff values higher than 5K_bulk. The curves plotted for different hydrodynamic radii show the slow variation of the Brown relaxation with this parameter. The conclusions will not change if the hydrodynamic radius is changed within the same order of magnitude.

Fig. 12 Dependence of the reciprocal of relaxation time (relaxation frequency) with NP average diameter. The Brown relaxation curves were determined assuming three different hydrodynamic radii (lines), and the Néel relaxation plots were calculated for one, five and ten times the CoFe₂O₄ bulk anisotropy constant (symbols). The shaded region corresponds to the sizes of the NP obtained from the solution.

The magnetic heating efficiency is low for all samples, as previously reported for other CoFe₂O₄ nanoparticles^8,9,21,37 and in agreement with the magnetic losses calculated from DC hysteresis measurements. For the S-NP, the HLP values are much lower than the SLP ones, as expected due to the large contribution of the Brown mechanism to SLP.

In the case of samples T-NP, the values of the coercive field are about three to four times the amplitude of the AC magnetic field used in the hyperthermia experiments. Thus, no rotation of the magnetic moments of the particles is achieved. This explains the low SLP values obtained and their good agreement with the HLP values, since rotation is not expected to occur in both cases.

Conclusions

CoFe₂O₄ nanoparticles were successfully synthesized using three different natural templates: flax, gauze and sisal.

The NP obtained from the solution filtration have an average diameter of 3.8 to 4.5 nm. The heating efficiency of the S-NP was about 7 W g⁻¹ for dispersions in water. For dispersions in paraffin, where NP are immobilised, the SLP values are much lower, indicating that Brown relaxation dominates over Néel relaxation for suspensions in water. The calculation of the relaxation times indicates an anisotropy constant on the order of 10⁶ J m⁻³, much higher than the value for bulk ferrite. This is a result that is consistent with the values obtained from magnetization measurements in S-NP powders, which can be explained by the important contribution of the surface in small NP.

The NP obtained after template calcination show morphologies that are determined by the threaded templates, leading to high aspect ratio nanostructures and anisotropic physical behaviours. In particular, the coercivity of these nanoparticles is strongly enhanced compared to other NP obtained by conventional methods. The coercive fields reach 1.3 T at 10 K and decrease to 70 mT at room temperature, which are values comparable to the ones obtained in CoFe₂O₄ nanofibers or nanorods. The high coercivity of these nanoparticles does not imply good efficiency as magnetic hyperthermia heaters, since the AC field amplitudes in this technique are usually below 30 kA m⁻¹ (within the biological safety range for clinical applications) and impose an upper limit close to 35 mT for the coercivity of suitable NP. This general result indicates that the optimization of the NP heating efficiency requires careful design of their size and shape.

In summary, this work presents a very simple method to produce CoFe₂O₄ nanoparticles with high magnetic coercivities. The results obtained are important and motivate future work to explore the use of this type of natural template for the production of NP with tailored coercivity. One possible pathway is to organize the template threads, inducing a preferable alignment of the template constituent fibers by adequate processing prior to its addition to the synthesis.

Acknowledgements

This study was funded by the Portuguese Foundation for Science and Technology (FCT) under the project PTDC/CTM-BIO/2102/2012 (including grants for S. G. Mendo and A. F. Alves), UID/MULTI/00612/2013 and UID/MULTI/04046/2013.

References

1. R. M. Anderson, C. R. Vestal, A. C. S. Samia and Z. J. Zhang, Appl. Phys. Lett., 2004, 84, 3115.
2. M. Lenglet and B. Lefez, Solid State Commun., 1996, 98, 689.
3. J. W. D. Martens, J. Appl. Phys., 1986, 59, 3820.
4. J.-S. Jung, J.-H. Lim, K.-H. Choi, S.-L. Oh, Y.-R. Kim, S.-H. Lee, D. A. Smith, K. L. Stokes, L. Malkinski and C. J. O'Connor, J. Appl. Phys., 2005, 97, 10F306.
5. P.-B. Liu, Y. Huang and X. Sun, ACS Appl. Mater. Interfaces, 2013, 5, 12355–12360.
6. G. Baldi, D. Bonacchi, M. C. Franchini, D. Gentili, G. Lorenzi, A. Ricci and C. Ravagli, Langmuir, 2007, 23, 4026.
7. D. H. Kim, S. H. Lee, K. N. Kim, K. M. Kim, I. B. Shim and K. Lee, J. Magn. Magn. Mater., 2005, 293, 287.
8. M. Veverka, P. Veverka, O. Kaman, A. Lančok, K. Závěta, E. Pollert, K. Knížek, J. Boháček, M. Beneš, P. Kašpar, E. Duguet and S. Vasseur, Nanotechnology, 2007, 18, 345704.
9. I. Andreu, E. Natividad, C. Ravagli, M. Castro and G. Baldi, RSC Adv., 2014, 4, 28968.
10. N. Panagiotopoulos, N. R. L. Duschka, M. Ahlborg, G. Bringout, C. Debbeler, M. Graeser, C. Kaethner, K. Lüdtke-Buzug, H. Medimagh, J. Stelzner, T. M. Buzug, J. Barkhausen, F. M. Vogt and J. Haegele, Int. J. Nanomed., 2015, 10, 3097.
11. H. Shokrollahi, A. Khorramdin and G. Isapour, J. Magn. Magn. Mater., 2014, 369, 176.
12. X. Zhang, J. Ge, B. Lei, Y. Xuea and Y. Du, CrystEngComm, 2015, 17, 4064.
13. P. Guardia, A. Riedinger, S. Nitti, G. Pugliese, S. Marras, A. Genovese, M. E. Materia, C. Lefevre, L. Manna and T. Pellegrino, J. Mater. Chem. B, 2014, 2, 4426.
14. G. Zhen, B. W. Muir, B. A. Moffat, P. Harbour, K. S. Murray, B. Moubaraki, K. Susuki, I. Madsen, N. Agron-Olshina, L. Waddington, P. Mulvaney and P. G. Hartley, J. Phys. Chem. C, 2011, 115, 327–334.
15. P. Hugounenq, M. Levy, D. Alloyeau, L. Lartigue, E. Dubois, V. Cabuil, C. Ricolleau, S. Roux, C. Wilhelm, F. Gazeau and R. Bazzi, J. Phys. Chem. C, 2012, 116, 15702.
16. X. L. Liu, E. S. G. Choo, A. S. Ahmed, L. Y. Zhao, Y. Yang, R. V. Ramanujan, J. M. Xue, D. D. Fan, H. M. Fan and J. Ding, J. Mater. Chem., 2014, 2, 120.
17. S. A. Shah, A. Majeed, K. Rashid and S.-U. Awan, Mater. Chem. Phys., 2013, 138, 703.
18. D. Peddis, C. Cannas, A. Musinu, A. Ardu, F. Orrù, D. Fiorani, S. Laureti, D. Rinaldi, G. Muscas, G. Concas and G. Piccaluga, Chem. Mater., 2013, 25, 2005.
19. M. Artus, L. B. Tahar, F. Herbst, L. Smiri, F. Villain, N. Yaacoub, J.-M. Grenèche, S. Ammar and F. Fiévet, J. Phys.: Condens. Matter, 2011, 23, 506001.
20. R. E. Lu, K. G. Chang, B. Fu, Y. J. Shen, M. W. Xu, S. Yang, X. P. Song, M. Liu and Y. D. Yang, J. Mater. Chem. C, 2014, 2, 8578.
21. S. G. Mendo, A. F. Alves, L. P. Ferreira, M. M. Cruz, M. H. Mendonça, M. Godinho and M. D. Carvalho, New J. Chem., 2015, 39, 182.
22. D.-Y. Li, Y.-K. Sun, P.-Z. Gao, X.-L. Zhang and H.-L. Ge, Ceram. Int., 2014, 40, 16529.
23. A. F. Alves, S. G. Mendo, L. P. Ferreira, M. H. Mendonça, P. Ferreira, M. Godinho, M. M. Cruz and M. D. Carvalho, J. Nanopart. Res., 2016, 18, 27.
24. W. B. White and B. A. DeAngelis, Spectrochim. Acta, Part A, 1967, 25, 985.
25. N. Sivakumar, A. Narayanasamy, K. Shinoda, C. N. Chinnasamy, B. Jeyadevan and J.-M. Greneche, J. Appl. Phys., 2007, 102, 013916.
26. M. Artus, L. B. Tahar, F. Herbst, L. Smiri, F. Villain, N. Yaacoub, J.-M. Grenèche, S. Ammar and F. Fiévet, J. Phys.: Condens. Matter, 2011, 23, 506001.
27. D. Peddis, N. Yaacoub, M. Ferretti, A. Martinelli, G. Piccaluga, A. Musinu, C. Cannas, G. Navarra, J.-M. Greneche and D. Fiorani, J. Phys.: Condens. Matter, 2011, 23, 426004.
28. S. Gyergyek, D. Makovec, A. Kodre, I. Arcon, M. Jagodic and M. Drofenik, J. Nanopart. Res., 2010, 12, 1263.
29. J. García-Otero, M. Porto, J. Rivas and A. Bunde, Phys. Rev. Lett., 2000, 84, 167.
30. A. J. Rondinone, A. C. S. Samia and Z. J. Zhang, Appl. Phys. Lett., 2000, 76, 3624.
31. R. E. Vandenberghe, R. Vandenberghe, E. De Grave and G. Robbrecht, J. Magn. Magn. Mater., 1980, 15–18, 1117.
32. J. L. Dormann, D. Fiorani and E. Tronc, Magnetic relaxation in fine-particle systems, in Advances in Chemical Physics, ed. I. Prigogine and S. A. Rice, John Wiley & Sons, Hoboken, New Jersey, 1997, vol. 98, pp. 287–288.
33. G. B. Ji, S. L. Tang, S. K. Ren, F. M. Zhang, B. X. Gu and Y. W. Du, J. Cryst. Growth, 2004, 270, 156.
34. A. Pirouzfar and S. A. S. Ebrahimi, J. Magn. Magn. Mater., 2014, 370, 1.
35. P. Jing, J. Du, C. Jin, J. Wang, L. Pan, J. Li and Q. Liu, J. Mater. Sci., 2016, 51, 885.
36. M. L. Huber, R. A. Perkins, A. Laesecke, D. G. Friend, J. V. Sengers, M. J. Assael, I. N. Metaxa, E. Vogel, R. Mareš and K. Miyagawa, J. Phys. Chem. Ref. Data, 2009, 38, 101.
37. Ö. Çelik, M. M. Can and T. Firat, J. Nanopart. Res., 2014, 16, 2321.

---