Effect of Gd3+ substitution on proton relaxation and magnetic hyperthermia efficiency of cobalt ferrite nanoparticles

In this study, Gadolinium substituted Cobalt Ferrite nanoparticles (CoFe2-xGdxO4, 0 ≤ x ≤ 0.4) were prepared via hydrothermal route using triethylamine as reducing agent at 180°C for 12 h. X-ray diffraction studies revealed the single phase cubic spinel structure for both Cobalt ferrite (CF) and Gadolinium substituted Cobalt Ferrite (CFG) nanoparticles (x ≤ 0.24). An increase in the Specific absorption rate (SAR) was observed with increase in Gd concentration. Further with increase in the molar concentration (x > 0.24), gadolinium hydroxide was observed as the secondary phase, which was also confirmed by the Gd–O stretching vibrations observed in Fourier transform Infrared spectroscopy. The evolution of Gadolinium hydroxide showed a strong influence in the relaxivity (r1) and hyperthermia potential. Field emission scanning electron microscopy revealed that CF and CFG (x ≤ 0.24) nanoparticles were spherical in nature with particle size ranging from 10 to 25 nm, whereas the particle size increases above 30 nm for CFG (0.3 ≤ x ≤ 0.4) nanoparticles along with the presence of columnar shaped particles. Magnetic measurements confirmed the pseudo single domain, ferri-magnetic nature of CF and CFG nanoparticles. The magnetization data revealed a change in direction of magnetization towards easy axis with increasing Gd concentration. The orientation of magnetization direction towards easy axis had induced change in the hyperthermia potential. Proton relaxation studies of CF and CFG nanoparticles revealed that there is a strong interaction between the relaxivities r1 and r2. The specific absorption rate of CF and CFG nanoparticles were observed to be in the range from 91.49 W g−1 to 232.17 W g−1 at applied Hf of 4.19 × 109 Am−1 s−1.


Introduction
The concept of hyperthermia was originated and been in practice over several centuries by ancient Greeks, Romans and Egyptians for the treatment of breast cancer [1][2][3]. However the acceptance of hyperthermia as a treatment for cancer has been seriously considered only in the 20th century after its successful clinical trials [3]. The spatial control in heating of biological tissue remains a challenge with various physical mechanisms such as microwave irradiation, ohmic heating, optical laser irradiation and whole body hyperthermia (water bath heating) [4]. Magnetic nanoparticle hyperthermia has gained its special attention due to their ability to concentrate the heat in the specific target area without affecting the other biological tissues [5][6][7]. Magnetic nanoparticles also play a significant role in the other biomedical fields such as magnetic resonance imaging (MRI) [8,9], targeted drug delivery systems [10][11][12], tissue engineering [13][14][15], labelling and separation of cells [16] and proteins [17,18] and integrated lab on chip technology for monitoring, diagnosis and therapy [19,20]. The contrasting ability, high magnetic moment and heating ability of magnetic materials makes them an ideal candidate for theragnostic applications.
The translation of magnetic hyperthermia into a standard medical procedure requires an improvement in successful administration of nanoparticles into the intra-tumoural region, reaching and maintaining the therapeutic temperature inside the tumour tissue [4,34]. The heating of magnetic materials occurs through Neel and Brownian relaxation losses in superparamagnetic and hysteresis losses in ferri/ferro magnetic materials [3,35,36]. The heating potential of superparamagnetic materials rely only on Neel relaxation as different biological tissues have different viscosities, leading to the loss of Brownian relaxation [3,37]. Whereas, the ferri/ ferro magnetic materials requires application of high magnetic field, frequency and power to achieve increased specific absorption rate [38,39], which is clinically not realisable due to physiological constraints [3]. After various experiments, the maximum threshold of Hf factor for hyperthermia application was estimated to be 5×10 9 Am −1 s −1 to avoid any detrimental effects on healthy tissues due to electromagnetic radiation exposure [40]. The materials should be able to achieve therapeutic heat dose at low applied magnetic field to overcome the physiological constraints. Hergt et al suggested that high specific absorption rate (SAR) value can be achieved with the particles whose size is near the transition from superparamagnetic to ferromagnetic behaviour [41]. Rumenapp and his coworkers reported that the magnetic resonance imaging capability of the superparamagnetic and ferromagnetic materials were almost equal [42].
Though there are several reports available related to hyperthermia, the region between single domain and multi-domain ferromagnetic nature of nanomaterials were not investigated till date. Hence, there is a vast scope in developing manifold magnetic nanoparticles with high SAR to work in the therapeutic region. In this study, the effort has been made to develop a Pseudo domain ferrimagnetic nanomaterial with better contrasting ability and high SAR suitable for theragnostic applications.
The structural properties of the CFG nanoparticles were characterized by PANalytical Xpert Pro x-ray diffraction system (XRD) with Cu-kα source. The morphological characters such as particle shape and size was studied using Field Emission Scanning electron Microscopy (FESEM), FEI Quanta, USA. The Elemental composition was detected using Energy Dispersive Spectroscopy (EDS), Bruker, Germany at an accelerating voltage of 20 KV for samples 0x0.24 and 30 KV for samples x=0.3 and 0.4. The vibrational frequencies of oxides and other organic impurities were studied using Fourier transform infrared spectroscopy (FTIR). Magnetization measurement of the CF and CFG nanoparticles was carried out using Vibrating Sample Magnetometer (VSM), Lakeshore, USA. The T 1 and T 2 relaxation times were measured using Minispec benchtop Time domain NMR Spectrometer, Bruker, Germany, equipped with a 470 mT (20.00 MHz for 1 H) permanent magnet. The T 1 relaxivity (r 1 ) of the nanoparticles were found from the slope of line fitted to 1/T 1 versus Gd concentration in milli Moles (mM) whereas T 2 relaxivity (r 2 ) was obtained from the slope of 1/T 2 versus Fe concentration in mM. The hyperthermic potential of the nanoparticles were measured under AC magnetic field at a frequency of 150 KHz and 35 mT with the help of 30 mm heating coil using MSI Automation Inc., USA equipped with Neoptix fiber sensor system. The specific absorption rate of CF and CFG nanoparticles were calculated from the data obtained from the hyperthermia potential by using initial slope method using the formula,

Cw Ms Mm dT dt
Where Cw is the specific heat capacity (4.187 Jg −1 K −1 ) of water, Mm and Ms are the mass of nanoparticles alone and nanoparticles along with dispersion medium respectively, dT/dt is the rate of change of temperature with respect to time.
The structural, magnetic and electronic properties of both Cobalt ferrite (CF) and Gadolinium substituted Cobalt ferrite (CFG) were also calculated theoretically using density functional theory (DFT). The Structure optimization, magnetic and electronic calculations were performed using Vienna Ab initio simulation package (VASP). The cutoff energy of 520 eV with 5×5×5 K-mesh was used for optimization calculations and tetrahedron method with 9×9×9 K mesh was used for total energy calculations. The generalized gradient approximation GGA+U approach has been applied for all the calculations. We applied U=4.06 eV, J=0.80 eV for Co and U=4.19 eV and J=0.81 eV for Fe. The magnetic moments in octahedral and tetrahedral sites are maintained antiparallel in both the structures.

Results and discussion
The total density of states (TDOS) of cobalt ferrites (CF) and gadolinium substituted cobalt ferrites (CFG) using GGA+U are shown in the figure 1(a). The crystal field (cf) theory states that in tetrahedral site of cobalt ferrites, the e g levels are lower than the t 2g levels due to the direct electrostatic repulsion between d xy , d yz and d zx orbitals and surrounding anion orbitals, while in octahedral sites the order is reversed as the d z2 and d x2−y2 orbitals are repelled directly. The relative strength of crystal field and intra-atomic exchange field (ef) determine the possible high spin (cf<ef) or low spin (cf>ef) electronic configurations [43]. From the TDOS, the band gap of the cobalt ferrite widens as the substitution of gadolinium increases from 1.58 eV to 1.68 eV. This can be explained by Burstein-Moss effect, wherein it states that the increase in the doping concentration in case of n type dopant, the electrons populates within the conduction band moves the fermi level to higher energy [44]. The density of states of individual atoms ( figure 1(b)) indicates that the electrons in the atoms becomes localized with the increase in the doping of gadolinium. The localization of electrons was also evidenced by the decrease in the charge density from 418.80 to 402.79 Coulombs/meter 3 . This decrease in the charge density also indicates the decrease in bond strength and increase in distance between the atoms due to the introduction of gadolinium in to the lattice whose ionic radii is comparatively larger than the iron. This increase in the bond length decreases the overlapping of electrons from different atoms leading to an increase in their band gap.
The x ray Diffraction studies (figure 2(a)) showed a single phase cubic spinel (Fd -3m) structure of cobalt ferrite (JCPDS # 022-1086) for CF and CFG nanoparticles with 0x0.24. Hexagonal phase Gadolinium hydroxide (CIF#4031383) with space group P 63/m was observed as a secondary phase with the increase in molar concentration (x>0.24). The lattice constant of inverse spinel CF and CFG nanoparticles were determined by theoretical (VASP) and experimental (High score Plus) data and the values are depicted in the figure 2(b). Both theoretical and experimental values showed an increase in lattice parameters from 8.5677 Å to 8.7620 Å and 8.3351 Å to 8.4049 Å respectively with the increase in gadolinium concentration (0.0x0.3). From theoretical calculations it was found that the lattice constant 'b' is 8.7105 and 8.7585 for 12% and 16% doping of gadolinium in cobalt ferrite which is comparatively lesser than 'a' and 'c' (where a=c). The change in the lattice constant 'b' indicates the increased strain in the lattice, which is evidenced by the origination of the secondary phase gadolinium hydroxide. The crystallite size of gadolinium substituted cobalt ferrites (0.0x0.4) calculated using Williamson hall (WH) plot was recorded in the figure 2(c). It was observed that the crystallite size does not change appreciably with the increased substitution of gadolinium concentration (0.0x0. 16). Further increase in the gadolinium concentration (0.2x0.4) substantially increases the crystallite size. The increase in lattice constant and crystallite size of gadolinium substituted cobalt ferrites (0.0x0.3) was greatly influenced by the substitution of Gd 3+ (94 pm), a larger ionic radii in the place of smaller Fe 3+ (67 pm) ions. A higher concentration of Gd 3+ (x>0.3) substitution exhibits a reverse trend in lattice parameter, which may be attributed to the saturation limit of Gd 3+ in Cobalt ferrite lattice. The trend of lattice constant also explains the incorporation of Gd 3+ ions into the interstitial sites at lower concentrations, whereas at the higher concentration segregation or accumulation of Gd 3+ on the surface takes place distorting the Cobalt ferrite lattice. The lattice strain of CF, CFG02, CFG04, CFG06, CFG08, CFG10, CFG12, CFG15, CFG20 nanoparticles obtained by WH plot were −0.2, −0.1, −0.1, −0.1, 0, 0, 0.1, 0.2, 0.2 respectively. The induction of lattice strain in the crystal structure was directly proportional to the increase in lattice parameter and crystallite size, which was greatly influenced by the higher gadolinium concentration in the parent cobalt ferrite.
The FESEM images of the synthesized CF and CFG nanoparticles were shown in figure 3. It was observed that CF and CFG (0.04x0.24) nanoparticles were uniform, slightly aggregated and almost spherical in shape. Beyond x0.3 a columnar shaped particles (ranges from 0.4 to 1.0 μm in length) indicating the emergence of hexagonal gadolinium hydroxide along with the spherical particles was observed. The particle size and the size range was measured by using the image J software. The average particle size of the CF and CFG nanoparticles were shown in the figure 2(c). The particle size ranges form 9-20 nm, 11-24 nm, 10-21 nm, 12-25 nm, 15-24 nm, 16-26 nm, 16-30 nm, 24-32 nm, 27-44 nm for CF, CFG02, CFG04, CFG06, CFG08, CFG10, CFG12, CFG15, CFG20 nanoparticles respectively. The particle size monotonically increase with increase in the concentration of gadolinium, which is further explained by the substitution of larger ionic radii of Gd 3+ at Fe 3+ ionic site. The increasing trend of particle size observed by FESEM are in good agreement with the crystallite size obtained from the XRD data.
The energy dispersed along with molecular weight percentage and atomic weight percentage of CF and CFG nanoparticles were shown in EDS ( figure 4). The ratio of molar concentration of the CF and CFG nanoparticles synthesized was 1:(2-x):x:4 (Co:Fe:Gd:O). Elemental analysis revealed that the Gd/Fe ratio in CF and CFG nanoparticles are in good agreement with the theoretical ratio of the synthesized nanomaterials except with the samples CFG15 and CFG20 as shown in the table 1. The atomic percentage of the oxygen detected in all the   samples varied substantially than the expected atomic percentage for CF and CFG (0.04x0.24) nanoparticles. The profound variation of oxygen is mainly due to low accelerating voltage (20 KV) used for analysis with or without the combined effects such as formation of more favorable auger electrons compared to the less favorable x-rays, low fluorescence yield, inability of low energy x-rays to leave the sample, absorption of photons in certain regions of detector and peak overlapping with L, M, N lines of heavier elements [45]. From the elemental analysis of CF and CFG (0.04x0.4) nanoparticles, it is clearly evident that various concentrations of gadolinium is incorporated into cobalt ferrite.
The chemical nature and presence of organic impurities in the synthesized CF and CFG nanoparticles was further confirmed from the vibrational frequencies obtained by FTIR spectroscopy (figure 5). The vibrational frequencies of CF and CFG nanoparticles were depicted in the table 2. Parent cobalt ferrite was confirmed by the presence of a strong characteristic band for Co-O stretching (850.61 cm −1 ) and Fe-O stretching (532.29 cm −1 ) vibrations [46,47]. The substitution of the gadolinium in the parent cobalt ferrite was clearly evident with the emergence of additional stretching vibrations at the region of ∼1400 cm −1 corresponding to the Gd-O linkage for CFG (0.04x0.4) nanoparticles [48]. It was also found that the substitution of gadolinium over x0.3 showed a characteristic vibrational frequency of Gd=O linkage and Gd-O-H linkage at the region of ∼1500 cm −1 and ∼3610 cm −1 [48,49]. The vibrations at the region of ∼1400 cm −1 , ∼1500 cm −1 and   The room temperature hysteresis of the CF and CFG nanoparticles was recorded using vibrating sample magnetometer at applied field from −1.5 to +1.5 tesla. Figure 6 indicates the ferri-magnetic nature of the particles and also revealed a slight shift in the magnetization direction from hard axis to easy axis. The ferrimagnetic behavior is observed when the magnetic moment of the sample interacts with the super exchange interactions [50]. The magnetic saturation of the CF and CFG nanoparticles were determined by using law of approach [51] and are represented in the table 3. The magnetic remanance (Mr) and magnetic saturation (Ms) were found to decrease with an increase in the gadolinium concentration. An abrupt decline in Ms beyond x0.3 is majorly due to the presence of paramagnetic gadolinium hydroxide, which has contributed to the total net magnetization of the CFG nanoparticles. The coercivity of the Cobalt ferrite decreases upto x0.2 and then increases slightly upon further doping. Similarly the anisotropy constant was found to decrease with increase in gadolinium concentration mainly due to the magnetic distortion, weak coupling to crystal structure, weak interaction of Fe-Re. The Mr/Ms and Hr/Hc values indicate that CF and CFG nanoparticles represent pseudo single domain characteristics [52]. It was also observed both theoretically and experimentally that the   magnetization and bohr magneton (μB) decreases with increase in Gd concentration. The difference in the bohr magneton value could be due to the magnetic interactions due to random occupation of the Fe sites by the gadolinium. The crystalline symmetry of nanoparticles was reduced with the strain in lattice with doping. The decrease in the magnetic saturation, magnetic anisotropy and coercivity is mainly due to the random occupation of dopant atoms in substitutional positions. In pristine cobalt ferrite, the magnetic interactions are high thus leading high coercivity and magnetic retentivity.
Proton relaxation studies Different concentration (0.0 to 0.2 mg) of CF and CFG nanoparticles in 0.05% w/v poly acrylic acid (PAA) gel were studied for the T 1 and T 2 relaxation rates at room temperature. Both T 2 and T 1 relaxation rates depicted in the figures 7(a) and (b) clearly showed that they are highly dependent on the concentration of the Iron and Gadolinium respectively. The r 1 and r 2 proton relaxivity values obtained from the slope of line fitted to the 1/T 1 versus Gd concentration and 1/T 2 versus Fe concentration respectively were tabulated in the table 4. It was found that the relaxation rate (1/T 2 ) and relaxivity (r 2 ) decreases with increase in the Gd concentration until x0.24. The decrease in relaxivity (r 2 ) can be due to decreasing saturation magnetization, as the magnetic moment of Gd is anti-parallel and their increased concentration opposes the net magnetic moment induced by Fe-Fe interaction [53]. Whereas, the increase in relaxivity was observed with Gd (x 0.3) due to the evolution of gadolinium hydroxide. The T 1 relaxation rate and relaxivity (r 1 ) of pure cobalt ferrite nanoparticles remains negligible, whereas it gradually increases with increasing concentration of the Gadolinium. The gradual rise in relaxivity (r 1 ) is mainly due to induced signal reduction of T 2 which is further due to weak Re-Fe interactions [53]. The samples CFG15, CFG20 showed a sudden increase in relaxivity (r 1 ) as the Gd 3+ ions are densely populated due to the occurrence of gadolinium hydroxide as a secondary phase. The abrupt rise in relaxivity (r 1 ) with respect to decreased Iron concentration and evolution of gadolinium hydroxide clearly states that Iron quenches the T 1 relaxivity of Gadolinium in cobalt ferrites.  The hyperthermia measurement was performed in 10 ml of CF and CFG nanoparticles dispersion (2 mg ml −1 ) in 0.05% w/v PAA gel medium. The Hf factor used for hyperthermia applications in this work was estimated to be 4.19×10 9 Am −1 s −1 , which is comparatively within the threshold limit [41]. The figures 8(a) and (b) shows the heating curve and specific absorption rate of CF and CFG nanoparticles. It was observed that the heating efficiency and specific absorption rate of the cobalt ferrite increases monotonically with the increase in the dopant (Gd 3+ ) concentration upto x0.24, whereas it decreases with further increase due to the evolution of the gadolinium hydroxide which possess larger particle size and reported to exhibit very weak magnetization. The increase in the specific absorption rate of CF and CFG (x0.24) nanoparticles is mainly due to the increase in particle size and increased poly-dispersivity [54]. The increase in SAR can also be explained by the decrease in the strength of magnetic interactions and anisotropy constant of CF and CFG nanoparticles [54].

Conclusion
The morphological, structural, crystallographic properties were investigated and found that the substitution of the gadolinium over 0.24 M showed an impure phase of gadolinium oxide. Magnetic measurement revealed the change in the direction of magnetization towards easy axis with increasing dopant concentration. The relaxation studies showed significant change in r 1 and r 2 relaxivities with respect to the gadolinium concentration. The hyperthermia efficiency has been greatly increased with the incorporation of gadolinium atom, makes them a potential candidate for Magnetic hyperthermic applications. The incorporation of these magnetic nanoparticles into drug loaded polymeric nanoparticles could help to enhance the site targeting, sensitivity of chemotherapeutic dose and provide adjuvant therapy along with chemotherapy.