Insight into synthesis and characterisation of Ga0.9Fe2.1O4 superparamagnetic NPs for biomedical applications

A Ga3+-substituted spinel magnetite nanoparticles (NPs) with the formula Ga0.9Fe2.1O4 were synthesized using both the one-pot solvothermal decomposition method (TD) and the microwave-assisted heating method (MW). Stable colloidal solutions were obtained by using triethylene glycol, which served as a NPs stabilizer and as a reaction medium in both methods. A narrow size distribution of NPs, below 10 nm, was achieved through selected nucleation and growth. The composition, structure, morphology, and magnetic properties of the NPs were investigated using FTIR spectroscopy, thermal analysis (TA), X-ray diffraction (XRD), transmission electron microscopy (TEM), X-ray photoelectron spectroscopy (XPS), and magnetic measurements. NPs with the expected spinel structure were obtained in the case of the TD method, while the MW method produced, additionally, an important amount of gallium suboxide. The NPs, especially those prepared by TD, have superparamagnetic behavior with 2.02 μB/f.u. at 300 K and 3.06 μB/f.u. at 4.2 K. For the MW sample these values are 0.5 μB/f.u. and 0.6 μB/f.u. at 300 K and 4.2 K, respectively. The MW prepared sample contains a secondary phase and very small NPs which affects both the dimensional distribution and the magnetic behavior of NPs. The NPs were tested in vitro on amniotic mesenchymal stem cells. It was shown that the cellular metabolism is active in the presence of Ga0.9Fe2.1O4 NPs and preserves an active biocompatible cytoskeleton.


Nanoparticle synthesis
The thermal decomposition approaching, as described by Cai and Wan 42 and recently optimized by Solano et al. 13 , was chosen for the synthesis of GaFeO NPs.First, a suspension of Ga(acac) 3 (0.64 mmol) and Fe(acac) 3  (1.28 mmol) in 50 mL TEG into an ultrasonic bath was realized.Then the mixture was directly added into a round-bottomed flask equipped with a condenser, a magnetic stirrer, a thermograph, and heating mantle.The system was slowly heated, 1 °C/min, up to 280 °C and kept at this temperature for three different periods of time.According to the dwell time at the reflux temperature, the samples were labeled as GaFeO#1/0 for 0 min, GaFeO#1/1 for 60 min, GaFeO#1/2 for 120 min and GaFeO#1 for 180 min.All samples were cooled at room temperature (ambient rate).Black homogeneous colloidal suspensions containing GaFeO NPs dispersed in TEG were obtained.These colloids are stable at room temperature, even after storage for several months.The GaFeO NPs were separated and washed thoroughly by magnetic precipitation and using a mixture (4:1 vol.) of ethyl acetate and ethanol.Finally, the NPs were dispersed in ethanol forming a stable black dispersion (160 mM).
The microwave process started from the same suspension of Ga(acac) 3 and Fe(acac) 3 in TEG, as previously described, which was now transferred into a microwave vial.After ultrasonication, the mixture was heated under magnetic stirring by microwave radiation (300 W) up to 220 °C, using a MW oven Discover Explorer Hybrid from CEM.The final temperature was achieved in 10 min with the maximum temperature ramp of 20 °C/min.This temperature was maintained for 10 min then the solution was cooled inside the microwave by the external air flow.The nanocrystals were separated by centrifugation and thoroughly washed with the mixture of ethyl acetate and ethanol thus yielding the GaFeO#2 sample.Redispersion in ethanol produced a suspension which remained stable for several months.

Characterization techniques
Bright field and High-Resolution Transmission Electron Microscopy images (TEM/HRTEM) coupled with selected area electron diffraction (SAED) were obtained using a 300 kV Tecnai G 2 F30 S-TWIN transmission electron microscope from FEI.The X-ray diffraction (XRD) measurements were taken using a Bruker D8 Advance diffractometer with Cu X-ray tube and incident beam Ge (111) monocromator (λ = 1.54056Å).The resulting powders were characterized by thermogravimetric and differential thermal analysis (TG-DTA) using a Mettler Toledo TGA/SDTA851 system, with a platinum crucible, at a heating rate of 10 K/min in air.The chemical nature of the nanoparticle surfaces was analysed by Fourier Transform Infrared Spectroscopy (FT-IR) using a Tensor 27 Bruker FTIR spectrophotometer.The qualitative and quantitative sample compositions were investigated using X-Ray Photoelectron Spectroscopy (XPS) assisted by Ar ions etching.The XPS spectra were recorded using a SPECS spectrometer working with an Al anode (1486.6 eV) as X-rays source.In order to avoid the artificial reduction of the different oxidation stats of elements under the Ar ions beam, the etching was performed by using ions accelerated at a maximum voltage of 1000 V with a filament current of 10 mA.The total content of Fe and Ga was determined through an inductively coupled plasma optical emission spectrometer (ICP-OES) Spectroflame FMD-07 (Spectro Analytical Instruments).The standard solutions used for ICP-OES calibration were purchased form Merck. Ultrapure water (18.2 mΩ cm resistivity) obtained in laboratory with the Millipore equipment (Bedford, USA) was used for dilutions.The magnetization properties were studied using the vibrating sample magnetometer (VSM Lake Shore) and a PPMS 9-QD system (sensitivity 10 -6 emu, 9-T maximum magnetic field, and temperature range 1.9-1300 K).Magnetization (M) versus applied field (H) was measured for the samples at 4.2 and 300 K. Zero field-cooled (ZFC) and field-cooled (FC) magnetization versus temperature (M-T) measurements were conducted in a 10 Oe magnetic field within the temperature range of 2-400 K.

Biocompatibility
The study was carried out with the approval of the "Academy of Romanian Scientists " Ethics Committee, with the approval registration number 778/13.12.2021.The amniotic fluid stem cells used in this research comes from pregnant women's that test for fetal aneuploidy.After obtaining the results the second cell culture sample used as backup was donate for research upon inform consent of the pregnant women.In this study were used the cells that presents a normal karyotype.All the experiments were carried out in accordance with relevant guidelines and regulations, in accordance with EU regulations and the principles of the Declaration of Helsinki, and upon approval of ethics committee.AFSCs were kindly provided by Genetic Lab S.R.L. diagnoses laboratory, upon written informed consent of the patients, in agreement with national and European Union law.The primary cultures were obtained by centrifugation of amniotic fluid at 1050 rpm for 10 min.The cells were then cultured for 10 days without passages in AmnioMax medium, with medium change every two days (ThermoFischer Scientific, Waltham, Massachusetts, USA).After 10 days, the primary culture was passage and cultured in differentiation specific media supplemented with growth factors.Endothelial differentiation of AFSC was done by culture in M200 medium supplemented with 10% fetal bovine serum (FBS), 40 ng/mL vascular endothelial growth factor (VEGF), 20 ng/mL insulin growth factor (IGF-1), 10 ng/mL epidermal growth factor (EGF), 10 ng/mL basic fibroblast growth factor (bFGF), 100 μg/mL penicillin, 100 μg/mL streptomycin, and 50 μg/ mL neomycin (all purchased from Thermo Fischer Scientific, Waltham, Massachusetts, USA).
The biocompatibility of GaFeO NPs was tested using human amniotic fluid stem cells (AFSC) after 72 h by biochemical assay (MTT assay) and after 5 days by fluorescent microscopy.For the cytotoxicity assessment, MTT assay (CellTiter Non-Radioactive Cell Proliferation Assay, Promega) was employed according to the manufacturer's guidelines.Briefly, AFSC were grown in 96-well plates, with a seeding density of 3000 cells/well, in different experimental conditions.Then, 15  www.nature.com/scientificreports/bromide) was added and incubated at 37 °C.After 4 h, 100 µL solution II (solubilization solution) was added and pipetted vigorously to solubilize formazan crystals.After 1 h, the absorbance was recorded at 570 nm using a TECAN spectrophotometer.Fluorescent microscopy was performed using RED CMTPX fluorophore (Life Technologies, Invitrogen, USA), a long-term living cell tracker.RED CMTPX dye was added in the culture medium (DMEM medium, Sigma-Aldrich, USA) at a final concentration of 5 mM and incubated for 30 min to allow the dye to penetrate the cells.Afterwards, the cells were washed with phosphate buffered saline (PBS) and visualized by fluorescence microscopy.The photomicrographs were taken with a Carl Zeiss digital camera using Axio-Vision 4.6 software.Cellular organization in the presence of GaFeO#1 and GaFeO#2 was assessed by looking at cytoskeleton tubulin filaments using immunocytochemistry.The AFSCs were cultured and then twice washed with PBS.They were fixed for 20 min with 4% paraformaldehyde in PBS, rewashed two times with PBS, permeabilized with 0.3% Triton X-100 in PBS (PBS-T) for 2 × 15 min, blocked with 4% normal goat serum in PBS-T (which was used for all subsequent washing and dilution of antibodies) for 1 h, treated with primary antibodies (tubulin, mouse monoclonal, 1:4000) versus negative control (dilution vehicle only) for 3 h at room temperature, washed 3 × 2-3 min, treated with secondary antibody (AlexaFluor 488 goat anti-mouse, Invitrogen, 1:1000) for 1 h at room temperature, washed 3 × 2-3 min and 2 × 1 min with purified water, mounted on glass microscopy slides with ProLong Gold anti fade with DAPI (Invitrogen P36935), and examined with an inverted fluorescence microscope.

Results and discussion
Today it is generally accepted that the formation mechanism of metal oxide NPs by wet-chemical methods consists in two main stages as described by LaMer's model: a nucleation step followed by the particle's growth step 43 .LaMer et al. described the formation of seeds by increasing the precursor monomer concentration up to the supersaturation level.The NPs are generated by a depletion of the monomer concentration as a result of their continuous aggregation onto the seeds.The chemical nature of the solvent influences both the nucleation and growth processes.This can be explained by the stability of the metal organic complexes that depends on the functional group of the solvent [42][43][44] .According to Solano et al. the use of TEG as reaction medium induces a limiting effect for the growth of MFe 2 O 4 NPs during heating stage (150 min) 13 .In this vein, our synthesis has been conducted using TEG as reaction medium for both approaches.To study the influence of the reaction time at the reflux temperature for the solvothermal approach, the synthesis has been done in standard conditions but up to a reaction time of 180 min.

Morphological characterization
The TEM and HRTEM images for the NPs synthesized by solvothermal approach at different reaction times from 0 to 1.5 h are presented in Fig. S1 in the Supplemental data.When the reaction mixture reaches the TEG boiling temperature (T = 280 °C, t = 0 h), the HRTEM image suggests that the nucleation step just occurred.After reflux for 1 h, the formation of NPs can be noticed, although they are not homogeneous in size.
The TEM and HRTEM images, SAED patterns, and particle diameter distribution for the GaFeO#1 sample (3 h) are presented in Fig. 1a-c.The GaFeO NPs present rounded shapes, low aggregation, and an average diameter of 5.1 nm.The HRTEM image in Fig. 1b shows various crystalline planes according to the orientation of this nanoparticle with respect to the direction of the electron beam.The following interplanar distances d hkl have been directly observed: 0.296, 0.252, 0.242, and 0.148 nm, which correspond to the (220), (311), (222), and (440) plane families of the spinel structure, respectively.The electron diffraction patterns present rings indexed with hkl reflections of Ga 0.9 Fe 2.1 O 4 spinel structure, as indicated in Fig. 1c, thus confirming the formation of well-crystallized NPs.
In the case of the GaFeO#2 sample, the same shapes and low aggregation of the NPs can be observed.The image is shown in Fig. 2a.The inset presents the corresponding diffraction patterns, but their indexation cannot be realized in a simple manner.Figures 1d and 2b show the normalized diameter distributions for the two samples.The corresponding insets represent the normalized volume distributions.A bimodal distribution can be noticed in the case of the microwave-prepared sample, GaFeO#2, with the maxima at 1.8 and 2.7 nm.For both samples the diameters of at least 400 NPs were determined for the statistics.The number and sizes of bins was determined by using the Sturges criterion.

Structural characterization
The crystalline structure of the GaFeO NPs was studied using X-ray diffraction (XRD), and the corresponding diffraction peaks are shown in Fig. 3a,b.All the reflections could be indexed to the spinel structure of Ga 0.9 Fe 2.1 O 4 (JCDD PDF 074-2226).Additional peaks representing supplementary phases are not observed here.
To study the influence of reaction time on the nucleation and growth of the NPs, different solvothermal syntheses were performed at various dwell times.Figure 3 shows the diffraction patterns of the reaction products corresponding to different dwell times.After a reflux of 1 h, as for GaFeO#1/1 sample, the Ga 0.9 Fe 2.1 O 4 diffraction patterns (spinel) show low-intensity Bragg reflections.It is an indication that nucleation has already started, and the nanoparticle growth process is ongoing.At longer reflux times (GaFeO#1/2 sample), grain growth continues, as illustrated by the increase in intensity of all Bragg reflections.After 3 h, as indicated by the diffraction patterns of the sample labeled GaFeO#1, the spinel structure is fully formed.
The average particle size of the fully synthesized sample was calculated using the Scherrer method and the (311), (440), (511) reflection peaks, resulting in a mean value of approximately 5.1 nm.The value calculated from the XRD data is the same as the mean particle size value obtained from TEM images (Fig. 1d).This indicates that the particles are mainly formed by nanocrystallites.
On the other hand, the XRD peak broadenings observed in the case of the GaFeO#2 sample indicate smaller dimensions than for the GaFeO#1 sample.Here, the calculated value through the Scherrer method is around 2.3 nm.It should be mentioned that these very broad peaks, as shown in Fig. 3b, have an uncertain baseline with a generally decreasing slope at high 2θ values.This is an indication that, besides the crystalline part, a large amount of amorphous material is contained within this sample.To further check the above-mentioned situation the Rietveld refinement was conducted by using the mentioned PDF 074-2226 file for Ga 0.9 Fe 2.1 O 4 spinel  structure.The convergence was achieved only for GaFeO#1 sample where the amorphous phase is minimal.The resulted fit is shown in Fig. S2 from the Supplemental data file.In case of this sample, all the determined lattice parameters and angles indicate the formation of the undistorted cubic spinel structure.The determined mean crystallite dimension was found to be around 5.7(4) nm, in accordance with its other evaluations.

Compositional characterization
The presence of polyol ligands on the surface of the GaFeO NPs, as a residue resulting from the preparation routes of the two samples, is evidenced by both TG-DTA and FTIR measurements.Figure 4 shows the TGA and DTA curves for GaFeO #1 and #2 NPs.The thermal analyses reveal two successive stages of weight loss: one in the temperature ranges of 25-200 °C, and the other between 200 and 350 °C.The first stage, with mass losses of 2.5 wt% for GaFeO#1 and 5.9 wt% for GaFeO#2, is attributed to the elimination of physically adsorbed solvents, such as ethanol or ethyl acetate.For both samples, the second stage represents the most significant weight loss, namely 9.1% for GaFeO#1 and 19.1% for GaFeO#2.It is attributed to the exothermic removal of residual TEG from the nanoparticle surfaces.In correlation with its smaller nanoparticle sizes and high specific surface area, both thermal processes are enhanced for the GaFeO#2 sample.

FT-IR spectroscopy
In order to understand and analyze the chemical nature of the organic species stabilizing the NPs, FT-IR spectroscopy was used.The results are presented in Fig. 5.The presence of TEG molecules on the surface of the NPs is confirmed through the characteristic strong absorption peaks at 1050-1100 cm −1 representing the symmetric and asymmetric stretching modes (ν sym and ν asym ) specific to the C-O-C bonds.It should be mentioned that these organic residual groups are either covalently bonded or chemisorbed onto the particle surfaces where they form a protective shell 45 .The FT-IR spectra also present high intensity bands at 764, 620, 598 and 443 cm −1 which can be attributed to Fe-O and Ga-O metal-oxygen interactions 46,47 .

The XPS analysis of samples
The sample compositions were determined by the XPS technique.The XPS recorded spectra of Fe 2p core-level doublet recorded for the GaFeO#1 and GaFe#2 samples, together with the corresponding deconvolutions and the fitted envelope, are shown in Figs.6a and 7a respectively.For both 2p (3/2) and 2p (1/2) core-level lines the deconvolutions were realized by considering the Fe 2+ and Fe 3+ states as components 48 .The low energy peaks (labels A) belong to Fe 2+ while the higher ones (labels B) belong to the more oxidized Fe 3+ state.At higher energies, for each Fe oxidation state two shake-up satellite features appear in spectra around 713.3 and 717.5 eV for   (3/2) peaks and at 727.6 and 732.6 eV for (1/2) peaks, respectively 48 .The restrictions used to fit the Fe 2p XPS spectra refer to the relation between areas of the two components, A 1/2 = A 3/2 /2, and to the spin-orbit doublet energy separation which was adjusted between 13.2 and 13.4 eV.The ratios between (1/2) and (3/2) linewidths were set between 1 and 1.1.
For the analysis of the gallium presence in samples the Ga 3d core-level was chosen because it is more sensitive to chemical shifts while the Ga 2p core-level, even if more intense, shows very small chemical shifts.The deconvoluted spectra of Ga 3d core-level lines are presented in Figs.6b and 7b, for the GaFeO#1 and GaFeO#2 samples, respectively.In these figures, besides Ga 3d doublet lines (labeled A), associated to Ga 0.9 Fe 2.1 O 4 , and positioned at 18.93 (5/2) and 19.37 (3/2) eV, another Ga 3d doublet (labeled B), was evidenced at 17.19 (5/2) and 17.63(3/2) eV respectively.The latter doublet arises from Ga + ions belonging to some Ga 2 O sub-oxide phase.The spin-orbit splitting of Ga 3d core-level lines was set at 0.44 eV for both quantifications.
To summarize, both samples contain [Fe 2+ O] T [Ga 3+ 0.9 Fe 3+ 1.1 O 3 ] O and Ga 1+ 2 O secondary sub-oxide phase, but in different relative concentrations.As it can be seen in Fig. 7b a larger quantity of Ga sub-oxide is formed in case of microwave synthesis route (GaFeO#2) than in the solvothermal preparation of GaFeO#1 sample (Fig. 6b).
The analysis of C 1s spectra is in accordance with the expected residual carbon and carbon compounds as found by TG and DTG and other complementary methods 38 .The C 1s core-level spectra are shown in Figs.S3 and S4 from the Supplemental data material.For XPS spectra calibration of both samples the C-C/CH peaks were positioned at a BE of 284.6 eV.The presence of low intensity C 1s peaks at BE greater than of 288 eV corresponds to the presence of different organic residues.
The higher decomposition rate of Ga(acac) 3 relative to Fe(acac) 3 45,49 causes an unbalance in the formation of compounds with different stoichiometry.In case of GaFeO#1 the solvothermal reaction at 280 °C is almost completed and the residual compounds are only formed by small amounts of gallium sub-oxide Ga 2 O, TEG chains and residual carbon or C-H moieties anchored on the particle surfaces.In general, XPS is a very useful technique in the compositional determination of surfaces.In the case of nanostructured materials that can contain several phases, through this method, when associated with Ar ions etching, complex quantitative compositional determinations can be obtained 50 .For the compositional analysis all the integral intensities were calibrated by using the relative sensitivities, transmission and electronic mean free path factors as given in CASA software database.Each of the two samples were subject to consecutive etchings with Ar ions until the shape and intensity of XPS spectra remained unchanged.At this stage, the quantitative composition of samples can be calculated by taking into account the escape depths of each identified component [50][51][52] .To account this, the calculated integral intensities were divided to the corresponding escape depths expressed in nm.Following Refs. 50,53,54the calculated escape depths are as following: 0.88 nm for Ga 0.9 Fe 2.1 O 4 (Fe 2p 3/2 kinetic energy-KE), 1.84 nm for Ga 2 O (Ga 3d 5/2 KE), 2.64 nm for C 6 H 14 O 4 (C 1s KE) and 2.24 nm in case of residual carbon (C 1s KE).Line positions and normalized integral intensities are presented in Tables S1 and S2 from the Supplemental data.Here the Table 1 summarizes the compositional analysis of GaFeO samples.In order to check the XPS calculated compositions the results from ICP-AES determinations are also presented here.

Magnetic characterization
The magnetic properties of the NPs were determined using the superconducting quantum interface device (SQUID) magnetometry.Figure 8a,b show the magnetization behavior of samples GaFeO#1 and GaFeO#2 as a function of the applied magnetic field, respectively.The values of the saturation magnetization (M S ) and coercivity field (H C ) for sample GaFeO#1, at 300 K, are 43.88 emu/g and 65 Oe while, at 4.2 K, the corresponding values are found to be 66.5 emu/g and 327 Oe respectively.The ratio M S /M R , with M R representing the remanent magnetization, was calculated as 0.03 at 300 K and 0.25 at 4.2 K for the GaFeO#1 sample, as it can be seen from the inset of Fig. 8a.In contrast, for the GaFeO#2 sample, the values were 0.02 at 300 K and 0.03 at 4.2 K (inset of Fig. 8b).These values, along with the low coercive fields observed in both samples, suggest that the nanoparticle ensembles exhibit superparamagnetic behaviour within this temperature range.In the case of the GaFeO#2 sample, the very small variation, with respect to the temperature, of the M R /M S ratio should be noted.The saturation magnetizations were calculated in the limit as 1/B approaches 0. The low values of the saturation magnetization indicate that in the case of the GaFeO#2 sample, there is a rather large number of nanoparticles without magnetic ordering associated with their poor crystallinity.These small particles have surface noncorrelated magnetic moments, in accordance with the analysis of the XRD patterns.When expressed in Bohr magnetons per formula unit (μ B /f.u.), the magnetizations of the GaFeO#1 sample are 2.02 μ B /f.u. at 300 K and 3.06 μ B /f.u. at 4.2 K.For the GaFeO#2 sample, the corresponding values are 0.5 and 0.6 μ B /f.u. at 300 K and 4.2 K, respectively.In calculating these values, the paramagnetic contributions were extracted.For the two samples, the corresponding concentrations of material showing magnetic ordering were considered, as indicated in Table 1.On the origin of these values, especially in the case of sample 1, the following qualitative considerations can be made: (i) The arrangement of cations in the spinel structure indicates a stoichiometry of 0.9 ions per formula unit (f.u.) of Ga 3+ and 1.1 ions per f.u. of Fe 3+ in the octahedral positions, while the tetrahedral positions are exclusively occupied by Fe 2+ ions (1 ion per f.u.).(ii) Due to the presence of Ga ions, assumed to be uniformly distributed in the network within the octahedral positions, the overlap of the 3d 5 wave functions of the Fe 3+ ions with oxygen and gallium is very limited, thus favoring their antiferromagnetic ordering.(iii) In this case, the observed magnetization can only arise from the ferromagnetic coupling of Fe 2+ ions (3d 6 in the high-spin configuration) in tetrahedral positions, facilitated by superexchange interactions mediated by oxygen ions.According to this hypothesis, the maximum number of Bohr magnetons is 4 μ B , which corresponds to the value of 3.06 μ B obtained at low temperatures.The difference is attributed to surface states, where the magnetic moments are likely canted into a glassed state.
The magnetization field dependence of an ensemble of superparamagnetic particles can be effectively described using the following equation: here: V(D) represents the volume of the magnetically ordered nanoparticles, which is expressed as a function of their diameters, H stands for the applied external magnetic field, L is Langevin's function, M s represents the saturation magnetization, which is normalized to the magnetic content of the samples, as detailed in Table 1.Also, f(D) is the normalized distribution of magnetic diameters, and it is described by a lognormal distribution function that includes the mean diameter D 0 and its dispersion σ.6][57][58][59][60] .The paramagnetic contribution was first extracted before fitting, and then it was reintroduced.
Thus, for GaFeO #1 sample the field dependence of the magnetization, M = f(H), is represented in Fig. 9 by a continuous line.It is superposed to the experimental demagnetizing data.The fit resulting parameter values are: D 0 = 5.1 nm, M S = 45.1 emu/g, and σ = 0.38.It can be seen that D 0 value is very close to value derived from the TEM data while σ is slightly larger than the value resulted from the statistical analysis of TEM images (Fig. 1a).In the case of sample GaFeO#2, achieving a proper fit was only possible by reducing the mass concentration of magnetic material within the composite to a range of approximately 0.4-0.6%.This reduction means that, out of the total composition of approximately 0.85%, slightly more than half of the nanoparticles possess cores with magnetic ordering capable of superparamagnetic rotation.This difference is likely attributed to the presence of very small nanoparticles, measuring below 2 nm.Even if these nanoparticles maintain the correct stoichiometric composition, their magnetic moments become frozen due to surface effects, resulting in a minimal contribution of this nanoparticle sub-assembly to the overall magnetization.As mentioned, within this procedure a linear paramagnetic term was extracted from the reordered data before fitting procedure.This alignment between experimental and calculated curves is illustrated in Fig. S5.Here, the determined values for M s , D 0 and σ are 8.0 emu/g, 2.5 nm and 0.65, respectively for a GaFeO#2 mass concentration of 0.5%.However, these values are not correct since they don't reflect the entire behaviour of the ensemble.
Additionally, an alternative set of solutions can be derived at higher concentrations, approximately 0.75%.Under these conditions, the calculated magnetic core diameters fall below 1 nm.This scenario represents the point where the model's practical relevance is maximally challenged.It's worth noting that NPs exhibit a core-shell structure in terms of magnetic moments.At extremely small external dimensions, the magnetically ordered core essentially loses its significance, as it is overshadowed by the frozen magnetic moments within the outer shell.Additional details are given in the Supplemental.Thus, the number of parameters becomes too large, and the fitting process gave multiple solutions, mostly unphysical.
Field-cooled (FC) and zero-field-cooled (ZFC) measurements are presented in Fig. 10a for GaFeO#1 and 10(b) for GaFeO#2.The results support the assumption of superparamagnetic behavior in the samples.The ZFC magnetization increases with temperature until it reaches a maximum value at the blocking temperature (TB).Above this temperature, thermal energy surpasses the magnetic anisotropy energy barrier, ΔE, causing the nanoparticles to exhibit superparamagnetic behavior 61 .
The spontaneous temperature evolution of the magnetization of an ensemble of superparamagnetic nanoparticles, during heating in a zero magnetic field and after a field cooling process, represents the so-called thermoremanent magnetization (TRM).TRM is a measure of the sequential deblocking process occurring within the ensemble of nanoparticles, each with a specific volume distribution.It can be measured directly or determined by the difference between the temperature dependencies of the magnetization curves for field cooling (FC) and zero-field cooling (ZFC) 62 .The first derivative of TRM provides insight into the distribution of energy barriers, denoted as ΔE, which must be overcome during the thermal deblocking of magnetic moments 63 .For of a monodomain NP of volume V the height of the energy barrier, is given by the relation �E = K eff V = k B Tln t m τ 0 .Here K eff is the effective anisotropy energy density, T is the temperature, k B is the Boltzmann constant, τ 0 represents a microscopic relaxation time, and t m is a relaxation time seen as the measurement time.Since τ 0 is of the order of 10 -9 s and t m has values between 10 and 100 s the value of the logarithm is between 23 and 27 62,64 .
The first derivatives of TRM functions are presented in Fig. 11 for both GaFeO#1 and GaFeO#2 samples.The discrepancy between the two samples is obvious and it is due to their different magnetic behaviors.Considering a value of 25 for ln t m τ 0 the values of the axial effective anisotropy constant K eff were calculated 62 as 1.38 × 10 5 and 9.1 × 10 5 J/m 3 for GaFeO#1 and GaFeO#2 samples, respectively.Supposing that these values are slowly depending on the deblocking temperatures of various volume fractions from a given nanoparticle ensemble, a reconstruction of the volume distributions can be done starting from the energy barriers distributions given in Fig. 11.
The same distribution of the heights of the energy barriers can be reconstructed from the diameter measurements in the TEM images.For comparison, in Fig. 12a,b are represented the superposed energy barriers distributions, E = k eff V , reconstructed from both TRM and TEM data for GaFeO#1 and GaFeO#2 samples respectively.
As can be seen from Fig. 12a, in the case of sample GaFeO#1, the overlap of the two distributions is very good, which means that, practically, the entire set of NPs in the mentioned sample is represented by crystallites, each of them having ferromagnetic ordering.Regarding GaFeO#2 sample, as shown in Fig. 12b, the superposition Figure 9.The magnetization curve of the GaFeO#1 NPs at room temperature (□) together with the best fit obtained by using Eq. 1 (continuous line).The magnetizations were calculated referred to the magnetic content of this sample as given in Table 1.appears only for the peak positioned at low energy values of the k eff V product.It is evident that several types of nanoparticles (NPs) are present within the sample, resulting in a bimodal distribution: NPs with small dimensions exhibit ferromagnetic ordering (as indicated by the blue curve), while the even smaller ones possess spin-canted states (especially for GFO#2 sample).Additionally, a significant portion consists of amorphous NPs with slightly larger dimensions, lacking magnetic ordering.These findings align with the results obtained from XRD and SAED.Consequently, this approach has allowed us to test the validity and limitations of the superparamagnetic model.

Toxicity studies
Data from stem cell studies suggest that amniotic fluid cells are a rich source of stem cells that could be used in tissue engineering for various applications in regenerative medicine 65 .AFSCs are a heterogeneous population of mesenchymal cells of fetal origin that, with higher proliferation and differentiation plasticity, are able to differentiate into all three germ layers, do not form tumors when injected in vivo, and are free of ethical issues concerning their employment 66 .The in vitro experiments performed on AFSCs revealed that the tested GaFeO#1 and GaFeO#2 have good compatibility with eukaryotic cells.The MTT assay demonstrated that human AFSCs present normal metabolism and growth in the presence of GaFeO#1 and GaFeO#2, with the measured values of absorbance at 570 nm being similar to those of control cells after 72 h of incubation (Fig. 13).A slight increase in proliferation is observed at 8.3 mg/mL and 16.66 mg/mL of GaFeO#2 NPs.
For a better analysis of biocompatibility, we used a second method to assess cytotoxicity.Thereby, fluorescence microscopy using RED CMTPX fluorophore as a cell tracker showed that AFSCs are viable and exhibit normal growth and proliferation capacity in the presence of GaFeO#1 and GaFeO#2 NPs (Fig. 14a-j).The results obtained after 5 days of incubation indicate normal morphology of human AFSCs.The shape and adherence properties are comparable to those of the control cells (Fig. 14a-f).No dead cells or cell fragments are evidenced in the presence of Ga 0.9 Fe 2.1 O 4 NPs, suggesting that the cells are viable.These observations sustain the fact that the GaFeO#1 and GaFeO#2 NPs are safe for use in further tissue engineering applications.
Furthermore, the biocompatibility of GaFeO#1 and GaFeO#2 NPs was investigated in a more complex manner, by evaluated the tubulin cytoskeleton organization using immunocytochemistry.The results are shown in   The control cells display a normal fibroblastic morphology with a fusiform shape, which enables a good cell attachment, spreading and motility as seen in Fig. 15a. Figure 15c-f represent the human AFSC in the presence of GaFeO#1 and GaFeO#2 NPs.They show no major changes in tubulin filaments reorganization.Nevertheless a slight elongation could be observed at 8.3 mg/mL and 33.33 mg/mL concentrations for both compounds.They preserve their initial architecture, with elongated extensions suggesting an active cytoskeleton represented by the network of microtubules, actin and intermediated filaments of proteins.
To conclude, all the assays showed that GaFeO#1 and GaFeO#2 NPs are biocompatible with human AFSC for the tested concentrations.

Conclusions
In summary we have prepared ultra-small NPs of Ga 0.9 Fe 2.1 O 4 , below 6 nm, by two different methods: thermal decomposition and microwave assisted process.In both cases acetylacetonates were used as metal precursors and triethylene glycol as high boiling solvent and surfactant.The resulted products are different from the common ferrite compound through the cation distribution and stoichiometry with Ga 3+ in octahedral sites and only Fe 2+ in the tetrahedral ones.This is sustained by XRD, XPS, ICP-AES and magnetization measurements.The presence of a TEG layer on the nanoparticle surface is responsible for the long-term NPs' stability in different solvents, such as: ethanol, hexane and water.The TD method assures the formation of crystalline, free of agglomeration NPs with a ferromagnetic coupling inside the matrix, while the particles obtained using MW method are smaller in size with a low degree of crystallinity and contain a large fraction of non-magnetic NPs as well as gallium suboxide.Thus, even if it is more laborious, the TD method produces GaFeO NPs with better quality with a good magnetic ordering.Also, the biocompatibility assessment showed that NPs are biocompatible with human AFSC.In vitro experiments on amniotic mesenchymal stem cells showed that the cellular metabolism of the cells is active in the presence of Ga 0.9 Fe 2.1 O 4 NPs being biocompatible with human AFSC.Additionally, µL of solution I (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium Vol:.(1234567890)Scientific Reports | (2023) 13:18175 | https://doi.org/10.1038/s41598-023-45285-y

Figure 2 .
Figure 2. (a) TEM image and SAED patterns (inset) for GaFeO#2 sample and (b) the bimodal normalized diameter distribution with the corresponding normalized volume distribution in the inset.

Figure 8 .
Figure 8. Magnetization of samples (a) GaFeO#1 and (b) GaFeO#2 as a function of the applied magnetic field, determined at room temperature and 4.2 K.

Figure 10 .
Figure 10.FC and ZFC curves highlighting changes in magnetization of NPs as a function of temperature for (a) GaFeO#1 and (b) GaFeO#2 samples.

Figure 12 .
Figure 12.Compared distributions of the k eff V product as resulted from TEM and magnetic measurements for (a) GaFeO#1 and (b) GaFeO#2 samples.

Fig. 15 .
Fig.15.The control cells display a normal fibroblastic morphology with a fusiform shape, which enables a good cell attachment, spreading and motility as seen in Fig.15a.Figure15c-f represent the human AFSC in the presence of GaFeO#1 and GaFeO#2 NPs.They show no major changes in tubulin filaments reorganization.Nevertheless a slight elongation could be observed at 8.3 mg/mL and 33.33 mg/mL concentrations for both compounds.They preserve their initial architecture, with elongated extensions suggesting an active cytoskeleton represented by the network of microtubules, actin and intermediated filaments of proteins.To conclude, all the assays showed that GaFeO#1 and GaFeO#2 NPs are biocompatible with human AFSC for the tested concentrations.

Figure 13 .
Figure 13.The measured values of the absorbance at 570 nm of the AFSC after 72 h of incubation in the presence of different quantities of GaFeO NPs as compared to the control cells.Here GF1 and GF2 means GaFeO#1 and GaFeO#2 samples respectively.

Table 1 .
The compositional analysis of GaFeO samples.