Introduction

Nanoparticles (NPs) have become very attractive for their applications in different fields, comprising biology, medicine, drug delivery systems, engineering, electronics, etc. [1, 2]. Several previous studies have explored the possibility to combine NPs with dendrimers and other polymers to build hybrid nanoscale structures for various purposes [3–6]. The dendrimer–nanoparticle conjugation can be obtained via two approaches: the physical encapsulation of particles in the internal cavity of a dendrimer, and the chemical formation of dendrimer branches around the inorganic core [7–16]. For example, dendrimer-grafted magnetite (MAG) nanoparticles were synthesized in order to improve the efficiency of coating formation on the surface of MAG nanoparticles. In particular, polyamidoamine (PAMAM) dendrimers were built on the surface of amine-functionalized MAG nanoparticles [7].

An interesting characteristic of dendrimers is that they can be covalently linked with several ligands, dyes, and drugs, thus providing a platform for the specific targeting, imaging, and treatment of cancer. In this context, dendrimer-based organic/inorganic hybrids have been considered for targeting and imaging tumour in animal models of human cancer [7, 17–19]. For example, magnetic NPs (in particular ferric oxide, Fe2O3) were modified with different generations of PAMAM dendrimers and mixed with antisense surviving oligodeoxynucleotides for applications in cancer therapy and MRI diagnosis [8–10, 20]. In a recent study, iron (II, III) oxide, Fe3O4, has been functionalized with polyelectrolyte (polystyrene sulfonate sodium salt, PSS) and PAMAM generation 5 (G5) dendrimers [21]. In detail, dendrimers were pre-functionalized with folic acid (FA) and fluorescein isothiocyanate (FI) moieties (G5.NH2-FI-FA) on the surface of iron oxide NPs using the LbL self-assembly technique. The formed FI- and FA-functionalized iron oxide NPs displayed very high specific binding affinity to cancer cells. Iron oxide NPs have been also synthesized in the presence of carboxylated PAMAM dendrimers G4.5 [12]. The electrostatic interaction of negatively charged carboxylated PAMAM dendrimers with positively charged iron oxide NPs is considered to play an important role for the stabilization of the NPs [9], whereas PAMAM dendrimers with other different functionalities (–NH2, –OH) might not be able to stabilize iron oxide NPs, indicating the role of electrostatic interaction for the NP stabilization [22].

Taking into account the current state of the art briefly described above, the present work is focused on hybrid nanosystems composed of MAG and PAMAM dendrimers functionalized with succinamic acid groups on its surface, prepared in a single step in high-pressure conditions (hydrothermal procedure).

The aim of this paper is to investigate the nature of interactions between MAG NPs and PAMAM dendrimers, in hydrothermal conditions, using computational and experimental techniques.

Computational simulations to predict the interactions between organic and inorganic components at high pressures and low temperatures were validated by experimental approaches in hydrothermal conditions.

Methodology

In this work, molecular modelling has been employed to investigate the interaction mechanism between PAMAM dendrimers and MAG particles, at high pressure values (100 atm). Simulation outcomes served as a starting point for identifying dendrimer types with the highest affinity for MAG nanoparticles. Those dendrimers have been then chosen for hydrothermal synthesis of hybrid nanostructured compound. Commercially available PAMAM dendrimers decorated with succinamic acid groups were specially selected as they are soluble in water and may easily interact with Fe3+ ions through carboxylic groups of succinamic acid.

Simulation models were validated by hydrothermal synthesis experiments at 100 atm and 40 °C. MAG/dendrimer hybrids’ structural properties have been also investigated by high-resolution transmission electron microscopy (HRTEM) and electron paramagnetic resonance (EPR) experimental techniques. A detailed description of materials and methods employed in this paper is reported in the following paragraphs.

Computational method

MD is powerful as a virtual microscope to investigate structural characteristics and interaction dynamics at molecular level [23–26], and it has been extensively used to model dendrimer systems [19–22, 27–31].

The aim of the molecular dynamics (MD) studies was to predict the interaction between dendrimer and MAG particles at specific temperature, pH, and pressure conditions, using atomistic modelling. To this aim, we compared two types of dendrimer differing in terminal groups exposed to the solvent: (i) the 1,4-diaminobutane core poly(amidoamine) (PAMAM) with NH2 terminal groups, which is soluble in methanol and insoluble in water, and (ii) 1,4-diaminobutane core poly(amidoamine) decorated with succinamic acid (PAMAM-SAHs), which has anionic surface groups, and typical carboxylic acid reactivity and is water soluble. For each dendrimer type, four generations (from G1 to G4) were considered. Linear formulas of each type of dendrimer are presented in Table 1.

Table 1 Linear formulas of modelled PAMAM dendrimers
Full size table

Dendrimer models were built, parameterized, and refined according to [13–16]. A refined three dimensional atoms’ arrangement was obtained by employing as a working platform the Dendrimer Building Toolkit (DBT) [27]. Each PAMAM dendrimer was considered as made of three main hyperbranched residue types: AAA (dendrimer core), BBB (dendrimer branch), and CCC (dendrimer terminal). A second terminal type, namely SUC, constituted by a CCC (the PAMAM amine terminal) attached to a succinic group HOOC–(CH2)2–COOH was defined. A dedicated in-house code has been employed for building the decorated dendrimer starting from PAMAM obtained by DBT.

The general Amber force field (GAFF) [32, 33] has been employed for force field parameters.

Force fields define a set of parameters for different types of atoms, chemical bonds, dihedral angles, etc. The typical parameter set includes values for atomic mass, van der Waals radius, and partial charge for individual atoms, and equilibrium values of bond lengths, bond angles, and dihedral angles for pairs, triplets, and quadruplets of bonded atoms, and values corresponding to the effective spring constant for each potential.

Partial charges were calculated by the restrained electrostatic potential (RESP) fitting method at the HF/6-31G* level of theory using Gaussian09 via the RESP ESP charge Derive Server (R.E.D.Server) [34–37]. Simulations have been set up on the basis of the parameters that will be considered in experimental approach: high pressure (100 atm) and pH 10. Hence, dendrimer models and relative force field were built by considering completely deprotonated structures (e.g. non-protonated amine terminals NH2 and carboxylate terminals COO–).

The MAG atomic coordinates were obtained by Crystallography Open Database (COD ID: 1011084). The unit cell had a size of 8.39 Å, composed of 28 atoms, 12 irons, and 16 oxygens, resulting in an oxygen/iron ratio of 4:3. A spherical MAG particle, with a diameter of roughly 4 nm, was created starting from COD data of a MAG super-cell of 5 × 5 × 5 nm. The clay force field [38] have been employed for defining MAG particle atom types and non-bonded topology (vdW and Coulomb). Partial charge optimization was carried out by GULP suite [39–41] following the Rappe and Goddard’s charge equilibration method [42].

Bonded topology for the MAG particle has been defined as an elastic network connected by harmonic potentials. Elastic network parameters have been refined, starting from the clay force field, by an iterative Boltzmann inversion (IBI) procedure which allowed us to define a set of bond parameters (force constant k, and reference length r 0) able to reproduce a reasonable distribution of MAG atomic fluctuations under thermal motion during the simulated conditions. The application of the IBI procedure, implemented following the procedure described in [43–46], allowed to avoid position restraints on MAG particles during MD simulations, taking into account molecule/surface interactions [47, 48]. Dendrimer models have been employed to set up eight molecular systems (four generations for PAMAM and PAMAM-SAHs, respectively), each constituted by one dendrimer and one MAG particle positioned at an initial distance of about 1 nm. Dendrimer- and MAG-based systems were then immersed in a box of TIP3P water molecules [49]. The dimension of each simulation triclinic box was chosen in order to ensure at least 1 nm solvation shell around the solute (dendrimer + MAG). Each system was first minimized by steepest descent energy minimization algorithm followed by a preliminary position-restrained MD of about 1 ns in isothermal–isobaric ensemble (310 K and 100 atm). A further production MD (none restraint applied) in the NVT ensemble at 310 K was carried out for 50 ns leaving the dendrimer and the MAG particle free to fluctuate under thermal motion and eventually interact. All simulations discussed in this work were carried out by GROMACS [50–52]. The visual molecular dynamics (VMD) [53] package was employed for the visual inspection of the simulated systems. Dedicated GROMACS tools were used for a quantitative analysis in terms of root-mean-square deviation (RMSD), root-mean-square fluctuation (RMSF), solvent-accessible surface (SAS) [54], contact surface, and radius of gyration (R g). On a time interval characterized by a stable contact surface between dendrimer and MAG (i.e. the last 10 ns of each simulation), we have estimated the binding energy by the molecular mechanics Poisson–Boltzmann surface area (MM/PBSA), widely employed to analyse biomolecular interactions [55]. MM/PBSA data have been used to obtain insights into the nature of the dendrimer–MAG interaction, in particular in terms of comparison between PAMAM and PAMAM-SAHs.

Experimental details

PAMAM dendrimers with succinamic acid surface groups (generations G4 and G2, respectively) were purchased as solution 10 wt% in H2O from Sigma-Aldrich (www.sigmaaldrich.com) and used as received. Iron (III) chloride hexahydrate, p.a. (Merck) solid salt, was dissolved in water and used as solution 20 wt% in H2O. Ammonia solution 25 % (Chimreactiv S.R.L.) was used as a mineralizing agent. Based on MD results and dendrimers’ solubility in water, PAMAM-SAH G4 dendrimer was selected for hydrothermal synthesis and characterization of hybrid organic–inorganic nanoparticles. Hence, investigated hybrid nanostructures were constituted by MAG NPs and PAMAM-SAH G4. PAMAM-SAH G2 was also considered for hybrid synthesis, as a comparison term. Linear formulas of these dendrimers are presented in Table 1. Hydrothermal synthesis parameters are presented in Table 2.

Table 2 Hydrothermal synthesis parameters
Full size table

Hydrothermal synthesis of hybrid nanostructures

Iron (III) chloride hexahydrate and ammonia solutions were vigorously mixed until a strong alkaline suspension (pH 10) was obtained. PAMAM-SAH dendrimer (solution 10 wt% in H2O) was further added to inorganic suspension, and the mixture was transferred in Teflon vessel of a closed autoclave (SAM, Romania) and endorsed with cooling system, for hydrothermal reaction at 40 °C and high pressure of 40 and 100 atm, respectively. Pressure was created inside the reaction system using argon gas. The resulted suspensions based on MAG–PAMAM nanostructures were lyophilized at −50 °C using a Martin Christ Alpha 1–2 LD Plus freeze dryer.

For comparative reasons, MAG nanoparticles were also prepared in aqueous solution starting from iron (III) chloride hexahydrate and ammonia, in the same conditions as nanostructured organic–inorganic hybrid. MAG nanoparticles were obtained after freeze drying at −50 °C of the resulted suspension. Both nanostructured MAG–PAMAM hybrid systems and MAG nanoparticles were characterized using HRTEM and EPR techniques.

Characterization

Structural properties, thermal behaviour, and morphology of iron oxide–PAMAM hybrid nanostructures were investigated using Fourier-transformed Infrared spectroscopy (FT-IR), EPR, thermal gravimetry (TGA), differential scanning calorimetry (DSC), and HRTEM analyses. FT-IR, TGA, and DSC results were presented elsewhere [3].

  1. (a)

    HRTEM characterization Samples were dispersed in ethylic alcohol, and a drop of the as-resulted suspension was deposited on a TEM copper grid coated with a thin amorphous carbon film with holes. Morphostructural characterization of the samples presented in Table 2 was investigated using high-resolution transmission electron microscope Tecnai F30, G2S Twin (1 Å line resolution)—FEI Company. Examination of the nanoparticles’ morphology was performed at 300 kV.

  2. (b)

    EPR characterization The EPR spectra were recorded for the samples presented in Table 2, in the temperature range between 100 and 350 K, by means of an EMX-Bruker spectrometer operating at X band (9.5 GHz) and interfaced with a PC (software from Bruker for handling and analysis of the EPR spectra). The temperature was controlled with a Bruker ST3000 variable-temperature assembly cooled with liquid nitrogen. The reproducibility of the results was controlled by repeating the EPR analysis three times in the same experimental conditions for each sample. All the spectra were performed in the same instrumental conditions to permit a comparison of the absolute intensity among the different samples. These conditions, termed “standard”, are detailed in the following: (i) receiver gain 6.32 × 102, (ii) modulation amplitude 3 G, (iii) time constant 10.24 ms, (iv) conversion time 40.96 ms, (v) resolution 2048 points, (vi) number of scans 10, and (vii) EPR tube size 2 mm internal.

Results and discussion

Computational analysis

Along the overall MD simulation, the dendrimer–MAG binding dynamics can be reasonably divided into three phases by observing the contact surface plots shown in Fig. 1: (1) a first phase (0–20 ns) characterized by the dendrimer’s nearest position and non-covalent binding of the MAG nanoparticle, (2) a transition phase (20–40 ns) in which the contact surface may still increase due to conformational changes of the dendrimer, and (3) a third phase in which the contact surface is stabilized (40–50 ns). In this phase, also dendrimer conformational properties, such as the radius of gyration (R g), reached a reasonable stability in all cases, as shown in Supporting Information section S1.1. As expected, the average PAMAM-SAH R g value was always higher than the PAMAM R g (S1.1—Figs. S.1, S.2) if comparing the same generation. In detail, as the number of succinic terminals increases, the dendrimer structure loses compactness due to the repulsion of negative succinic terminals on the surface, similarly to the PAMAM dendrimer at a neutral pH [31]. In Fig. 1, the relationship between dendrimer generation and area of the contact surface is presented. In the case of PAMAM dendrimer, the area of the contact surface depends directly on the dendrimer generation, while for PAMAM-SAH no dependence is observed. In particular, the PAMAM-SAH G2 showed the highest contact surface.

Fig. 1
figure 1

Dendrimer–MAG contact surface throughout the MD simulation. A reasonably stable contact surface is observed in the last 10 ns for all simulations

Full size image

A visual inspection through VMD helped in examination of the contact between dendrimer and NPs (Fig. 2, section S1.2). In particular, from the indicative picture shown in the left panel of Fig. 2, it can be noticed that the mode of binding between PAMAM and MAG NP is different from the one between PAMAM-SAHs and MAG NP. A detailed explanation of the binding modes characterizing PAMAM and PAMAM-SAH dendrimers is provided in Supporting Information section S1.2.

Fig. 2
figure 2

Normalized radial distribution function versus dendrimer atoms’ distance from MAG surface (nm). In PAMAM-SAHs G3 and G4, a wide part of the dendrimer remains far from the MAG surface and free to eventually bind other MAG nanoparticles. On the left, an MD snapshot taken in the last 10 ns of the MD simulation for G4 PAMAM and PAMAM-SAHs. Water molecules are not shown in the picture

Full size image

This binding mode may be also quantified by the calculation of the radial distribution function (RDF) of dendrimer atoms with respect to the MAG surface (Fig. 2). Atomic distribution for all generations of PAMAM dendrimers is picked at a distance of about 0.5 nm from the MAG surface, decreasing quickly for higher distances. A slightly wider distribution is found for PAMAM G3 and G4. RDF curves calculated for PAMAM-SAH dendrimers show an interesting dependence on dendrimer size. G3 and G4 RDF curves for PAMAM-SAH clearly indicate that a large part of the dendrimer does not interact with MAG NPs (the contact surface between MAG and dendrimer is lower), as clearly shown in Fig. 2.

The different binding mode emphasized for the decorated (PAMAM-SAHs) and not decorated PAMAM dendrimers reflects the different nature of dendrimer/MAG interaction. Using the MM/PBSA approach, we have calculated the binding energy during the last 10 ns of each simulation (Fig. 3). The binding energy has been calculated by considering van der Waals, electrostatic, non-polar, and polar contributions. In all cases, given a defined size of the NP (~4 nm in this study), the dendrimer/MAG affinity rises proportionally with the dendrimer generation. Moreover, the PAMAM-SAH dendrimers have shown a much higher increase of the binding energy resulting in a value of about −2.0 MJ/mol for G4. While, for non-decorated PAMAM, it is possible to relate the slight increase of affinity to the increase of the contact area, this relationship may not be inferred for PAMAM-SAHs.

Fig. 3
figure 3

MM/PBSA binding energy calculated for different generations of PAMAM and PAMAM-SAH dendrimers. Affinity of the PAMAM dendrimers for MAG increases slightly moving from G1 to G4. Binding energy is mainly related to the increase of the contact surface. PAMAM-SAH affinity is mainly driven by the electrostatic contribution. The number of succinic groups exposed in the dendrimer outer surface is responsible for the affinity increase much more than the contact surface

Full size image

Data coming from MM/PBSA calculations have indicated that dendrimer/MAG affinity may be dependent, as expected, on the dendrimer size. Moreover, our data have also highlighted how the interaction between PAMAM-SAHs and MAG is mainly driven by the electrostatic contribution, whereas, in case of non-decorated PAMAM, the vdW contribution drives the interaction. PAMAM-SAHs, characterized by negatively charged succinamic acid surface groups at high pH, have shown a much higher affinity for MAG NPs. Binding energy decomposition (Fig. 4) clearly highlighted a change in main players contributing the binding energy. While the binding energy is mainly dominated by the vdW contribution for PAMAM, electrostatics is almost totally responsible in the case of PAMAM-SAHs, in particular for higher generations where the number of terminal succinamic groups overcome a certain threshold (as in the case of G3 and G4 PAMAM-SAHs).

Fig. 4
figure 4

Dendrimer/MAG binding energy decomposition over the interaction energy contributions

Full size image

Experimental

Hydrothermal synthesis in high-pressure conditions

Inorganic–organic nanohybrids have been prepared by hydrothermal method at high pressure, with several advantages briefly mentioned as follows:

  1. (i)

    low energy developed by applying pressure (for a liquid phase, the same energy is involved on five units for the temperature scale than on 4000 units for the pressure scale);

  2. (ii)

    negative ΔV value [ΔV = Σ(V/Z)(j)-Σ(V/Z)(i)], where i indicates the precursor and j the product;

  3. (iii)

    improvement of the chemical reactivity. When pressure is imposed, the distance between inorganic nanoparticles and PAMAM decreases, and weak physical bonds appear between terminal groups of dendrimer and Fe3+ ions.

However, little is known about the reaction mechanism of MAG formation in hydrothermal conditions, starting from iron (III) chloride hexahydrate.

The lowering of the surface activity of adsorbed gas molecules is dependent on pressure increase and is more evident in the case of non-polar gases like Ar or O2 [56].

In order to have a combination of Fe3+ and Fe2+ ions present, MAG requires at least a moderate fugacity value of oxygen (fO2) [57]. At a working pressure of 100 atm, water dissociation determined by pressure and solutes (amorphous iron hydroxide and PAMAM dendrimer) may appear [58]:

$$ {\text{H}}_{2} {\text{O}} \leftrightarrow {\text{H}}_{\text{aq}}^{ + } + {\text{HO}}_{\text{aq}}^{ - }. $$

H+(aq) could determine the reduction of Fe3+ to Fe2+ and the formation of MAG [59].

Moreover, in high-pressure conditions (40–100 atm), reaction kinetics are controlled by diffusion of gases to and from haematite–magnetite [60].

A possible explanation of the MAG formation could be that the pressure inside the system was created by Ar gas in aqueous medium, as presented in Scheme 1. Solubility of Ar in water is very low at working pressure [61]. Thus, at the liquid–gas surface, the probability to form hydrogen bonds with the gas is reduced. The surface tension is increased due to the fact that water molecules from the surface are attracted inside the water volume by the remaining stronger hydrogen bonds [56].

Scheme 1
scheme 1

Possible mechanism of hybrid organic–inorganic nanostructure formation

Full size image
$$ {\text{FeCl}}_{3} + 3{\text{NH}}_{3} + 3{\text{H}}_{2} {\text{O}} = {\text{Fe}}\left( {\text{OH}} \right)_{3} + 3{\text{NH}}_{4} {\text{Cl}} $$

Characterization

  1. (a)

    HRTEM characterization Different morphologies of MAG NP alone and nanohybrids MAG/PAMAM-SAHs are revealed in Fig. 5.

    1. (a)

      HRTEM micrograph of MAG nanoparticles alone (Fig. 5a) shows typical morphology of hydrothermally prepared nanostructures with crystallite size ranging between 2 and 5 nm. Round-shaped nanoparticles are formed.

    2. (b)

      In the case of dendrimer G2, small crystallites of 6–15 nm, as well as aggregates consisting of crystallites with 3 nm in size, can be observed (Fig. 5b).

    3. (c)

      The presence of MAG nanoparticles in hybrid structures with G4 PAMAM-SAH was proved by calculated interplanar distance (2.53 Å) and associated Miller index (311). Crystallite size of MAG in this type of hybrid is about 3 nm (Fig. 5c).

Fig. 5
figure 5

HRTEM images of a magnetite nanoparticles, b hybrid nanostructures with PAMAM-SAH G2, and c hybrid nanostructures with PAMAM-SAH G4

Full size image

Unfortunately, HRTEM images cannot indicate if MAG NPs are entrapped or not in dendrimer’s cage. For a better understanding of this aspect, EPR analysis was performed.

  1. (b)

    EPR characterization EPR spectra were recorded in order to study the nature of interactions between MAG and PAMAM dendrimer.

It is well known that MAG is a ferrous-ferric oxide, containing Fe2+ and Fe3+ in its structure. However, Fe2+ ions are not involved in EPR absorption, but their interaction with Fe3+ ions influences the characteristics of the absorption lines, according to Castner and coworkers [62]. Thus, the EPR investigation discussed in this paper refers to Fe3+ and their interaction with PAMAM-SAH dendrimer.

Some examples of the room-temperature (298 K) EPR spectra (normalized in height) obtained for Fe3O4 NPs alone (MAG) and MAG–PAMAM-SAH hybrid structures (samples presented in Table 2) are shown in Fig. 6. The room-temperature spectra of the MAG NPs interacting with the dendrimer structure are constituted by a single line at approximately g = 2.03 (measured using as a reference the 2,2-diphenyl-1-picrylhydrazyl, DPPH radical with g = 2.0036). At g around 2, signal is usually attributed to those ions which interact by super-exchange interaction. The g = 2.03 value is in good agreement with the one reported in the literature for α-Fe2O3 NPs [63].

Fig. 6
figure 6

Examples of room-temperature (298 K) EPR spectra (normalized in height) obtained for MAG NPs alone and embedded into PAMAM dendrimer structure (1:1 ratio) at different generations (G2, G4). The preparation conditions are also indicated

Full size image

EPR parameters which characterize EPR spectra are gyromagnetic factor (g), absorption intensity (I), linewidth (ΔH pp), splitting type, and coupling constant. Linewidth enables to evaluate the type of interactions from the studied hybrids, due to its dependence on spin–spin relaxation time and spin–lattice interactions.

EPR parameters such as absorption line intensity, estimated from double integration of the spectra, and the linewidth (ΔH pp) of the EPR absorption profiles of the MAG NPs and MAG–PAMAM-SAH hybrid structures (samples presented in Table 2) are depicted in Fig. 7.

Fig. 7
figure 7

a Intensity of the g = 2.03 line, estimated from double integration of the EPR spectra for MAG NPs into PAMAM-SAH dendrimers. b Linewidth (ΔH pp) of the g = 2.03 line of the EPR spectra for MAG NPs into the PAMAM-SAH dendrimer

Full size image

The dendrimer structure affects the EPR parameters of the hybrid nanostructures. The decrease of the intensity with the increase of pressure and dendrimer generation (Fig. 7a) could be explained by the interactions with vicinal SAH groups (mainly electrostatic, as shown by computational simulations) which suppressed the super-exchange interaction between Fe3+ and Fe2+ ions.

The linewidth (ΔH pp) increases with synthesis pressure of nanostructured hybrid (Fig. 7b).

The distance between Fe3+ ions and dendrimer surface decreases with increasing pressure.

Two interaction mechanisms are possible: spin–lattice relaxation and spin–spin relaxation. Spin–lattice relaxation implies interaction between Fe3+ ions and dendrimer network. It is characterized by a relaxation time, T1. Spin–spin relaxation or cross relaxation refers to Fe3+–Fe3+ interactions and is characterized by relaxation time T2. When both spin–lattice and spin–spin relaxations contribute to the EPR signal, the resonance linewidth (ΔH pp) can be written as

$$ \Delta H_{\text{pp}} \propto \frac{1}{{T_{1} }} + \frac{1}{{T_{2} }}. $$
(1)

From Eq. 1, we can tell that when T 1 > T 2, ΔH pp depends primarily on spin–spin interactions.

The increase in ΔH pp of the g = 2.03 absorption line could be explained by the increased dominance of clustering mechanism (spin–lattice interactions). The interactions of Fe3+ with the dendrimer surface may on one side separate the ions and diminish the spin–spin (Fe3+–Fe3+) interactions. Decreasing the spin–spin distance, which is the spin concentration, T 1 will become very short, and thus the spin–lattice relaxation will have a larger influence on the linewidth than spin–spin relaxation. On the other side, if the binding sites at the dendrimer surface (Fe3+ ions) are close to each other, the spin–spin interactions increase (T 2 will become very short), with a consequent decrease in the EPR intensity, while the linewidth increases [64].

Based on EPR results, it can be concluded that MAG is not entrapped in PAMAM structure and the interactions between organic and inorganic components take place at dendrimer’s surface.

EPR results validated the MD simulation. Thus, the dendrimer generation is (i) proportionally related to the increase in the linewidth of EPR and (ii) inversely related to EPR intensity (Fig. 7a). A possible explanation based on the observations suggested by in silico investigation could be that at high pH, as the dendrimer generation increases, the dendrimer/MAG affinity raises consequently, due to the higher density of exposed succinic groups on the dendrimer outer surface. Hence, the probability of finding dendrimer/MAG bound states increases, and consequently the exchange narrowing due to the interactions between Fe3+ and Fe2+ is suppressed by the dendrimer coating over the MAG NPs.

Conclusions

In this work, a combined computational/experimental approach was employed to investigate structural characteristics of dendrimer/MAG hybrid nanostructures. Several generations of PAMAM and PAMAM-SAH were investigated by MD simulations in order to find a possible mechanism for the formation of hybrid structures. Computational results indicated that the interaction between PAMAM-SAHs and MAG is mainly driven by the electrostatic contribution, and the affinity of these dendrimers for MAG NPs is higher due to negatively charged succinic terminals. Morphostructural analysis of MAG NPs revealed the formation of crystallites with size range between 2 and 5 nm. In the presence of G2 PAMAM-SAH dendrimers, small crystallites of 6–15 nm were observed. In the case of G4 PAMAM-SAH dendrimers, the presence of MAG nanoparticles was proved by calculated interplanar distance (2.53 Å) and associated Miller index (311). In all types of hybrids, MAG crystallites with 3 nm in size have been highlighted. EPR results have shown that MAG is not entrapped in PAMAM structure and the interactions between organic and inorganic components take place at dendrimer’s surface. This evidence has been also confirmed by calculating the dendrimer void volume from MD trajectories (Supporting Information section S1.3) following the protocol described in [65].

Further investigations aimed at precisely identifying the macromolecular organization of dendrimer/MAG clusters which may be considered as a fruitful future development of this work. Moreover, additional computational and experimental studies on dendrimer/MAG hybrids will extend the presented investigation to other dendrimer types and functionalization. From a computational point of view, also varying the size of the MAG NP may be an interesting extension of this study in order to understand how the binding mode is influenced by the ratio between dendrimer and NP size.