far apart, the charge is shielded; in contrast, the interactions occur when the nanoparticles get close. Fresnais and coworkers (2009) described the role of desalting kinetics on the nanoparticle clustering. About the assembly monitoring, different procedures were tested as direct mixing, dialysis, dilution, and quenching. Moreover, the kinetics of assembly of the electrostatic clusters were led by the rate dIS/dt at which the salt was removed from the solution, where IS denotes the ionic strength. Moreover, it was shown how the assembling process was driven by the desorption–adsorption transition of polymers on the nanoparticles surface. An ionic strength decrease caused the polymeric nanoparticles clustering. Furthermore, it was demonstrated that by regulating the desalination kinetics, the size of the clusters varied from 100 nm to over 1 μm. Meyer et al. (2006), in his work studied the influence of charges on the nanoparticles assembling process. It was demonstrated how the continuing accumulation of charges during the clustering process led to an increase in electrostatic repulsion. This phenomenon can be used to regulate the cluster growth process. On the base of the previously described work, Xia and coworkers (2012) showed the spontaneously assembling inorganic nanoparticle with nonuniform distributions in superparticles with core‐shell morphologies. The electrostatic repulsion and the van der Waals attraction led this self‐growth process.
3.3 Clustering Methods
In the last two decades, considerable efforts have been undertaken for the development of approaches for the controlled clustering of multiple magnetic nanoparticles. Several groups have also provided proofs of concepts for the applications of such magnetic clusters as sorting probes for cells/analytes separation, as vehicles for magnetically guided drug delivery, or as contrast agents in magnetic resonance imaging. In the following subsections, some relevant studies on the topic are discussed, focusing on preparation methods, on main properties enhancements, and significant achievements. The studies are grouped according to the clustering approach or the molecular coating class (Scheme 3.1).
Figure Scheme 3.1 Schematic illustration of magnetic nanoparticles clustering and covering some of the principal superstructures reported in the literature.
3.3.1 Synthetic Approach
The nanostructures reported here are not obtained by the assemblies of presynthesized nanoparticles, but by promoting the in situ nucleation of the magnetic core in a molecular or polymeric matrix. Most of the nanostructures were prepared by using a solvothermal approach at high temperature in autoclave, but some additional methods will also be considered in this section.
In 2005, Deng et al. in a pioneering article described for the first time the solvothermal synthesis of magnetic microsphere of different crystal phases (Fe3O4, MnFe2O4, ZnFe2O4 or CoFe2O4) by promoting the nucleation of magnetic nanoparticles in the presence of ethylene glycol, sodium acetate, and polyethylene glycol and obtaining regular magnetic sphere from 200 to 800 nm (Deng et al. 2005).
Ge et al. proposed a one‐pot preparation route based on the synthesis of magnetic nanoparticles followed by their controlled clustering at high temperature in the same reaction flask (Ge et al. 2007). In detail, a mixture of iron chloride, diethylene glycol, and poly(acrylic acid) was heated at high temperature, 220 °C; then the hot‐injection of sodium hydroxide induced first the nucleation of single crystals with a diameter of 6–10 nm, and subsequently, caused their agglomeration in flower‐like nanostructures. The cluster size could be tuned from 30 to 180 nm by simply changing the rate of hydrolysis of NaOH. By performing the synthesis under nitrogen reflux in flask, instead of using the autoclave, Cha et al. investigated the intermediate steps that bring to the cluster formation (Cha et al. 2013). The authors suggested that at the initial stage, the hydrolysis/condensation of FeCl3 occurred to generate an amorphous ferrihydrite. Whereupon, the dehydration process induced a phase transformation to crystalline lepidocrocite, followed by a second phase transformation to magnetite crystals. According to this solid‐state phase transformation model, the final grains of magnetic nanoparticles were generated directly into the resulting cluster matrix that acted as a synthetic environment. A similar approach has been exploited recently by Casula et al. by introducing a doping element in the superparamagnetic nanoparticles to enhance the magnetic properties of the nanostructure. Manganese was chosen as a typical dopant, mainly used for the replacement of Fe(II) in the lattice of magnetic phases. In detail, a mixture of iron(III) chloride and manganese(II) chloride salts was used as metal precursor. Interestingly, the presence of dipolar interactions induced a blocked state at room temperature, with evidence of coercivity at 310 K. The obtained structures were investigated both in MRI, thanks to their considerable performances, by showing a maximum value of r2 of 571 mM−1 s−1 at 0.5 T, and magnetic hyperthermia (Casula et al. 2016).
Also, Lu et al. in 2012 reported a one‐pot approach to synthesize the magnetic microspheres by using a solvothermal method. In detail, iron chloride, sodium polyacrylate, sodium acetate, and ethylene glycol were mixed and autoclaved at 210 °C for 10 hours. During this process, first, some magnetic seeds were formed and subsequently assembled in controlled clusters, protected by a layer of PAA. The microsphere sizes (from 100 to 500 nm) were tuned by simply modifying the water amount in the reaction. These nanosystems exhibited a quasi‐superparamagnetic behavior with an excellent saturation magnetization of 81.6 emu g−1 (Lu et al. 2013). A similar synthetic approach has been proposed recently with the introduction of urea in the precursors mixture, obtaining dense nanoclusters with a size ranging from 250 to 640 nm (Ganesan et al. 2019).
Lin et al. by using a similar method investigated the contribution of ethylenediaminetetraacetic acid disodium salt (EDTA‐2Na) and sodium acetate in the formation of magnetic nanosphere. In this work, a growth–dissolution–regrowth model is reported for the formation of single crystals in the superstructure. The sodium acetate amount governed the size of the magnetic grains, ranging from 5 to 30 nm. In contrast, the EDTA amount and the sonication pretreatment time were considered for controlling the overall size of the nanocluster (Lin et al. 2013).
The synthesis of magnetite nanoclusters by using sodium citrate within a mixed‐solvent system of diethylene glycol and ethylene glycol was evaluated by Wang et al. in 2015. In this solvothermal method, the sodium citrate acted not only as a ligand for the stabilization of the resulting clusters but also as one of the key parameters to control the cluster size, ranging from tens to hundreds of nanometers (Wang et al. 2015). By using a similar approach, recently, another study reported the preparation of a multifunctional nanosystem. In this regard, the citrate‐coated nanoclusters were first covered by a NIR molecule, namely cypate, and then enveloped in a red‐blood‐cell ghost membrane. The as‐obtained biomimetic complex showed a significantly improved physiological stability and an enhanced tumor accumulation after intravenous injection in mice (Wang et al. 2020).
In 2013, Daniele et al. adapted a classic coprecipitation reaction to functionalize magnetic nanoclusters through a modified copolymer by exhibiting an alkyne surface functionality. This structure can be exploited for rapid click chemistry functionalization. In details, poly(acrylic acid‐co‐propargyl acrylate) was used as a model, in which the acrylic acid guarantees for the carboxylate groups that anchor onto the iron oxide surface, whereas propargyl acrylate acts as the functional comonomer due to its general application in click reactions. The polymer was added just after the ammonium hydroxide addition in the nanoparticles synthesis, leading to the formation of the cluster with a hydrodynamic diameter around 150 nm. By AC susceptometry analysis, the authors observed that the relaxation time was dominated by Brownian relaxation, suggesting that the interaction between the nanoparticles and the copolymer arose before the clustering process (Daniele et al. 2013).
Bain