Magnetic domains orientation in (Fe3O4/γ-Fe2O3) nanoparticles coated by Gadolinium-diethylenetriaminepentaacetic acid (Gd3+-DTPA)

In this work, the magnetic domains (MDs) orientation was evaluated from magnetite/maghemite nanoparticles (Fe3O4/γ-Fe2O3) NPs coated with Gadolinium (Gd3+) chelated with diethylenetriamine pentaacetate acid (Gd–DTPA). The (Fe3O4/γ–Fe2O3) superparamagnetic cores were configured by adding a DTPA organic layer and paramagnetic Gd as (Fe3O4/γ–Fe2O3)@Gd–DTPA NPs. The cores were obtained by coprecipitation and coated with additional modifications to the synthesis with Gd–DTPA. Analysis of properties showed that particles 9–12 nm, with Gd–DTPA layer thickness ∼10 nm increased their magnetisation from 62.72 to 75.82 emu/g. The result showed that the structure, particle size, composition, thickness and interface defects, as well as the anisotropy, play an important role in MDs orientation of (Fe3O4/γ–Fe2O3)@Gd–DTPA NPs. Magnetic force microscopy (MFM) analysis showed an MDs uniaxial orientation of 90° at magnetisation and disorder at zero conditions and demagnetisation. The MDs interactions showed uniaxial anisotropy defined in the direction of the magnetic field. These addressable and rotational features could be considered for potential applications to induce hydrogen proton alignment in water by longitudinal spin-lattice relaxation T 1 and transversal spin-spin relaxation T 2 as a dual contrast agent and as a theranostic trigger.


Introduction
The configuration of nanostructures for dual contrast agents (T 1 -T 2 ) is still a challenge controlling the MDs rotation and promoting the hydrogen proton magnetic moment alignment [1]. Magnetic resonance imaging (MRI) scanning depends on the contrast agent anisotropy, as addressable MDs with respect to a strong static external magnetic field. Typically, this is achieved from 1.5 to 7 T, and pulsed radio frequency (RF) in the order of MHz [2]. The hydrogen protons of physiological media undergo a relaxation process by orientation and return to the original equilibrium. Their orthogonal components are: spin-lattice relaxation T 1 produced from magnetisation parallel to the static external magnetic field, and spin-spin relaxation T 2 generated from magnetisation decaying on the plane perpendicular to the external field [3]. MRI agents have been based on magnetic moment interactions from paramagnetic ion complexes as gadolinium (Gd 3+ ) for T 1  sodium hydroxide (NaOH, pellets, 97%) were purchased from Sigma-Aldrich. Distilled water was used for all experiments. All chemical reagents were of analytical grade and used without further purification.

(Fe 3 O 4 /γ-Fe 2 O 3 ) NPs synthesis
The (Fe 3 O 4 /γ-Fe 2 O 3 ) NPs were processed by co-precipitation according to the Massart method [22]. Starting from an aqueous solution of HCl, 2 M. From this solution, 25 ml was added to 6.965 g of FeCl 3 ·6H 2 O, 1M (Fe 3+ solution), and 6.25 ml were aggregated to 2.51 g of FeCl 2 ·4H 2 O, 2M (Fe 2+ solution), and each solution was stirred for 30 min. The iron solutions were mixed, from 5 ml of the Fe 3+ solution and 1.25 ml of the Fe 2+ solution. The reaction was heated at 80°C and deoxygenated by argon gas during entire process.
NPs were precipitated by dropwise addition of 30 ml of tetraethylammonium (TEA) hydroxide solution [23]. The solution turned to black. Finally, the NPs were washed several times with distilled water and ethanol, until a pH=7 was reached.

(Fe
The Gd-DTPA solution was added dropwise after (Fe 3 O 4 /γ-Fe 2 O 3 ) precipitation in experimental conditions of inert atmosphere, at 80°C under mechanical stirring for 1 h. The suspension was then cooled to room temperature. The final product was precipitated, and the Gd-DTPA and TEA excess were removed by washing several times with ethanol and water (1:1) until the pH was 7. Finally, the sample was freeze-dried. This sample was denoted as MG-1.

(Fe 3 O 4 /γ-Fe 2 O 3 )@Gd-DTPA NPs synthesis in-situ
The Gd-DTPA solution was mixed directly with iron solutions (Fe 2+ and Fe 3+ ). Subsequently, 30 ml of TEA hydroxide solution was added dropwise. After 1 h, the solution was cooled to room temperature and precipitated. The nanoparticles were washed several times to remove the unabsorbed Gd-DTPA and TEA, until a neutral pH was reached. The final product was lyophilized. This sample was denoted as MG-2.

Physical characterisation
The XRD was performed using a Rigaku SmartLab diffractometer with Cu Kα (λ=1.5406 Å) radiation and a nickel filter. The measurements were carried out at 9 kW (200 mA, 45 kV) at a scanning rate of 0.02°/s, 25-70°r ange at2 .
q The mean crystallite size was estimated with Scherrer equation: D=0.9 λ/β cos Ө, for (311) reflections. The lattice parameters were calculated from the experimental XRD patterns using PowderCell for Windows, Version 2.4. The FWHM values and sizes were evaluated through a modified Scherrer analysis using a pseudo-Voigt function [25]. TEM analysis was performed using a JEOL JEM-2010 microscope at 200 kV. Each sample was dispersed in isopropanol and held in formvar carbon copper grids, 400 mesh. Histograms were obtained from 130 particles for each sample. FTIR analysis was performed in the range of 4000 to 400 cm −1 using a Nicolet 6700 spectrometer. The chemical compositions of the surfaces were analysed by XPS using a Thermo Scientific K-Alpha system equipped with a monochromatic Kα X-ray source. All the signals were calibrated using the adventitious hydrocarbon peak, C 1s, located at 284.6 eV. Furthermore, the powder compositions were recorded by EDS coupled with scanning electron microscopy (SEM) Auriga Zeiss FEG 25 kv. Magnetisation curves were measured using a magnetic property measurement system (MPMS ® 3), Quantum Design, at room temperature under a maximum applied field of 3 T. The topography and MDs analyses were performed by using a scanning probe microscope (SPM) JEOL-JSPM-5200 in the mode atomic force microscope-magnetic force microscopy (AFM-MFM) or Lift Mode. Each powder sample was confined to a carbon adhesive tape and flattened with the pressure of a flat glass. A magnetic tip NSC18, Co-Cr/Al Micromasch with an uncoated radius of 8 nm, coated radius<60 nm and full tip cone angle of 40°was used for MFM characterisation. The magnetisation of the tip was performed using a neodymium magnet. Topography and MFM images were obtained at 180 kHz with a lift height interaction of 5-86 nm, output of 0.011-0.025 Amp/V and H=12 kOe under saturation conditions. The MFM resolution is not high enough to see features of the nanoparticle dimensions, and single-particle measurement can be performed in principle if the magnetic stray field generated by the particle is sufficiently strong. [26]. The images were processed using the Gwyddion modular program for SPM. [27].

Magnetic force microscopy (MFM)
The MFM provides the response of the magnetic stray field above the sample. This is detected by a magnetic tip which is oriented perpendicular or normal with respect to the analysis surface. The MFM is a dynamic mode, which traces the topography and phase detection line by line and known as Lift Mode [28].
The force gradient exerted between the magnetic tip and a magnetic sample allows the deflection of the cantilever and magnetic domain mapping. Because the magnetic forces are in longer range than the atomic forces in AFM, the lift mode decreases the lateral resolution by a factor comparable to the average tip separation (lift distance, Δz). The spatial resolution resolved at z is less than 30 nm by MFM, whereas AFM technically has atomic resolution. In addition, the spatial resolution dependent on the tip diameter [29].
The fact that in MFM, the tip is placed several nanometers above Δz implies that the polarised superparamagnetic nanoparticles in powder below 10-12 nm would be difficult to analyse because of the topological differences. The agglomerates produce elongated aggregates, and the magnetic signals depend on the exchange, magnetostatic and anisotropy energies with respect to Δz [30,31], figure 1.
In practice, working MFM distances between a few nanometers and a few tens of nanometers are used. Because of strength force gradient rapidly decays with distance from the surface. Fast scans with tip surface distances as small as 10-20 nm can be achieved using this method. The magnetisation within the sample reaches its equilibrium configuration as a result of exchange interactions, anisotropy and demagnetizing field  distribution [32]. The MFM images show MDs defined by the local magnetic structure of superparamagnetic nanoparticles.   inverse cubic spinel structure with crystallographic planes (220), (311), (400), (422), (511) and (440), respectively, figure 2(b). The lattice parameters (a 0 ), FWHM values and sizes were obtained through a modified Scherrer analysis using a pseudo-Voigt function [25].  [35]. DTPA is derived from the linear amine, diethylenediamine, by the addition of five acetate groups. The resulting ligand was octadentate with three nitrogen donor atoms and five carboxylate oxygen donor atoms [36], as shown in figure 2(a). The coating mechanism of (Fe 3 O 4 /γ-Fe 2 O 3 )@Gd-DTPA NPs is related to the chelating agent DTPA, which can be stable in a solid or liquid state by two valences, auxiliary and principal. Its functional group contains donor atoms such as O and N, which can be coordinated to Gd 3+ , Fe 2+ and Fe 3+ . DTPA has affinity for the metal cations to form stable complexes and reverse the metal binding after chemical treatment for encapsulation of Gd 3+ and ferrites. The order of affinity of the contrast agent chelator endogenous ions is Gd 3+ >Fe 3+ >Fe 2+ [37]. The  nanoparticles prepared by coprecipitation of ferrous and ferric ions in aqueous solution produced NPs agglomeration effect. This method allows the elimination of the organic bases by washing. The Fe ions at the NPs surface can be linked to -OH, -COOH, or -NH 2 to increase their stability. The coprecipitation route allowed to define the contrast of the Gd-DTPA coating by TEM. It is possible to use other synthesis methods as colloidal or solvothermal. However, the oleic acid, n-octylamine, or oleylamine used in these routes are not eliminated from the NPs. A better option to reduce the agglomeration is the dispersion in citrate acid. However, the application of organic dispersants reduces the contrast of amorphous Gd-DTPA layer using TEM [38,39].

Results and discussion
The AFM-MFM topography analysis was performed by Lift mode in a scanned area of 300×300 nm. The

(Fe 3 O 4 /γ-Fe 2 O 3 )@Gd-DTPA NPs bonding and composition
The FTIR spectra of GdCl 3 ·6H 2 O, DTPA and synthesised Gd-DTPA are shown in the figure 5(a). In the GdCl 3 ·6H 2 O spectrum, the bands at 3372, 1160 and 640 cm −1 are assigned to asymmetric and symmetric stretching and bending vibrations of H-O-H [40]. The Gd-Cl interaction was observed in a broad transmittance band at 1402 cm −1 . The band at 1628 cm −1 is a consequence of the chemical interaction between Gd 3+ and H 2 O [41]. In the case of DTPA, the bands at 1736, 1696 and 1633 cm −1 correspond to C=O bending in -COOH [42]. Comparing the FTIR spectra of GdCl 3 ·6H 2 O and DTPA with Gd-DTPA synthesised, the broad band that contains the two wavenumbers, 1626 and 1595 cm −1 , belonging to asymmetric and symmetric stretching vibrations of -COO − of DTPA, respectively, in Gd 3+ coordination [43]. The peak at 1736 cm −1 disappeared in the Gd-DTPA spectrum, indicating that the carboxyl proton dissociates, and the oxygen atom was coordinated to the metal, indicating the successful attachment of Gd 3+ and DTPA [44]. . This band was shifted to 3408 cm −1 in the nanostructure spectra of MG-1 and MG-2. This shift was attributed to the N-H stretching vibration, due to DTPA contribution [46]. In addition, DTPA has a characteristic peak at 1682 cm −1 from the C=O stretching vibration. The band at 1626 cm −1 , confirmed the (Fe 3 O 4 /γ-Fe 2 O 3 )@Gd-DTPA NPs as a result of carboxyl bonds in the MG-1 and MG-2 spectra, figure 5(b).
The XPS spectra showed photoelectron lines at bonding energies of 55, 285, 530 and 710 eV, corresponding to Fe 3p, C 1s, O 1s and Fe 2p, respectively [47]. The elemental surface compositions are shown in table 2. The carbon are shown increased from 37.98 at % in MG-1 and MG-2, and small amounts of nitrogen were   associated with the DTPA coating. Therefore, the Gd 3+ ion was located by energy photoelectron lines at 1186 and 1219 eV for Gd3d 5/2 and Gd3d 3/2 , respectively and a small peak in 143 eV for Gd4d [16,48], figure 5(c). A quantitative composition analysis was performed using EDS-SEM. The EDS spectra showed the atomic compositions of Fe, C, O and Gd. Gd 3+ atoms were not detected in MG-1, because of their low concentration over the (Fe 3 O 4 /γ-Fe 2 O 3 ) NPs surface below the detection limit (<1% minimum detectable mass fraction), figure 6. The results matched well with the atomic porcentages observed in the XPS analysis.  [33,50].

Magnetic analysis
MFM measurements of SPIONs in powder are a real challenging. These NPs can be attracted to the magnetic tip. In these cases, the tips can be cleaned by argon or nitrogen flow at a low pressure and recovered. The main requirement for AFM-MFM is that this sample must be flat. The powders show differences between topological features owing to pronounced hills and valleys. In addition, the orientation of MDs implies the assembly and agglomeration of nanoparticles due to the exchange between neighbouring MDs, and magnetostatic interactions. The scanning of areas below 500×500 nm define the topography and the MDs mapping in SPIONs. Therefore, the MDs local magnetic structure is defined by their alignment in the direction of applied magnetic field and collinear magnetic field lines in the lift mode. The MDs were evaluated by scanning areas of 300×300 nm as attractive and repulsive shifting interactions by the magnetic tip over the NPs surface [51].  MDs. The response of the magnetic stray field above sample showed a random distribution of MDs at initial conditions (H=0) in the three cases. The MDs were aligned at (−12 kOe<H<12 kOe) in the direction of magnetic field lines. These results showed uniaxial anisotropy. The MDs returned to the initial conditions (disorder), under demagnetisation conditions in all cases [52]. The MDs reorientation showed faster relaxation of MG-1 and MG-2 by Gd-DTPA coating than naked (Fe 3 O 4 /γ-Fe 2 O 3 ) NPs (see the sequence in the supplementary information, S1 (available online at stacks.iop.org/NANOX/2/020019/mmedia)). The naked (Fe 3 O 4 /γ-Fe 2 O 3 ) NPs require more relaxation time to recover the initial conditions. The saturation conditions indicated that (Fe 3 O 4 /γ-Fe 2 O 3 )@Gd-DTPA NPs were oriented owing to their superparamagnetic behaviour, because of their low coercivity and remanence. The MFM showed that the local magnetic structure is defined by particle size, shape, and composition in coherence with the magnetisation. The agglomeration showed the same MDs orientation at magnetization conditions [53].
The temperature influences the MDs orientation of (Fe 3 O 4 /γ-Fe 2 O 3 )@Gd-DTPA NPs because of the thermal barrier is reduced as the particle size becomes in the superparamagnetic diameter. MDs can randomly change their direction with respect to temperature. The diffusion in the magnetisation magnitude increases rapidly with temperature when T <T c and saturates at T c . Beyond T c , SPIONs diffusion becomes isotropic and independent of temperature [54,55].
The MFM-zoom shows the MDs alignment (H↑) in the 3D images, Local Gd 3+ dispersion over (Fe 3 O 4 /γ-Fe 2 O 3 ) NPs surface introduced small differences; however, the MDs alignment and coherence were achieved owing to the shape, size, defects and amorphous Gd-DTPA layer. The main control parameter for SPIONs as carriers and diagnosis is the MDs orientation. Anisotropy plays an important role in magnetic stimulation. The shape related to the semispherical shapes of NPs introduces strong magnetic field over a larger magnetic volume [56]. This anisotropy is a key parameter for the RF tuning and can be modified from the NPs shape. The (Fe 3 O 4 /γ-Fe 2 O 3 )@Gd-DTPA NPs can be considered to induce proton interactions along the z-axis for spin-lattice relaxation time, T 1 at 90°flipping and the alignment in the xy-plane