In order to bind the NPs to the CMC chains, the nanoparticles are treated with 3-aminopropyltrimethoxysilane (APTMS, Figure 1). APTMS hydrolyzes in the presence of water and, in a subsequent reaction, the silanol groups condensate with the hydroxyl groups on the NPs’ surface to form M-O-Si covalent bonds. This method introduces onto the NP surface the amino group essential to create an amide bond with the activated carboxylic group of CMC polymers [12,17].

The success of the nanoparticle silanization can be confirmed by infrared analysis. Figure 2 shows APTMS (bottom) and NP-NH₂ (top) FT-IR spectra. Both spectra show common -NH₂ and Si-O absorptions: the bands at ~1,600 and ~1,410 cm⁻¹ are assigned to the bending mode of the -NH₂ group and to the scissor vibration of -NH, respectively. The two bands appearing at ~1,100 and ~1,040 cm⁻¹, as well as a somewhat weaker band at ~ 920 cm⁻¹, can be attributed to the vibrational modes involving the bridging oxygen atoms in Si-O-Si and Si-O bonds [18]. Moreover, the XPS spectrum of NP-NH₂ showed the presence of signals at 400 eV and 150 eV, ascribable respectively to the presence of N and Si, absent in the spectrum of bare NPs. This evidence clearly confirms the presence of an aminosilane coating on the surface [12].The XPS study provided an estimated thickness of APTMS on the nanoparticle surface of about 1–2 layers, and showed the presence of many amino active groups on the surface for binding the polysaccharides chains.

Other information regarding the quantity of NH₂ groups present on NP was obtained by electrochemical analysis of NP, NP-NH₂ and APTMS in DMSO and CH₂Cl₂ solutions. Although the proposed drug delivery systems will be used in an aqueous environment, DMSO and CH₂Cl₂ solvents were employed instead of water because they offer a wider potential window; in an aqueous solution, the processes under study would have been masked by solvent discharge. XPS and IR spectroscopy have investigated the NP in the solid state, while by means of cyclic voltammetry, we look at their solution properties. In Figure 3, the redox behaviors of sonicated NP (a), sonicated NP-NH₂ (b) and APTMS (c) are compared. The NP in DMSO solution undergoes a not-well-defined reduction process at E_pc = −1.17 V that seems to be an overlapping of several processes; the cathodic potential value and the shape of this peak correspond closely to those found for iron ion reductions in CoFe₂O₄ spinel electrodes [19,20].

APTMS in CH₂Cl₂ solution undergoes an irreversible oxidation process at E_pa = +1.56 V which is assigned as amine centred [21]. The irreversibility of this process means that the oxidation process causes APTMS decomposition and generates a by-product which reduces at E_p = −0.40 V (asterisked peak).

NP-NH₂ cyclic voltammogram in DMSO solution shows both an irreversible cathodic process at E_pc = −1.17 V and an irreversible anodic process at E_pa = +1.02 V that generates a by-product (E_p = −0.61 V) (asterisked peak). The reduction process is identical to that found in NP cyclic voltammetry. As shown in the insert in Figure 3, the oxidation process of NP-NH₂ in DMSO and that of APTMS in CH₂Cl₂ show similar anodic pathways and similar by-product formation, so both are amine oxidation processes, although the anodic potential value of APTMS is 500 mV higher than that of NP-NH₂. The NP-NH₂’s oxidation shift towards less anodic potentials cannot be explained merely in terms of a different diffusion coefficient, but rather depends on APTMS deposits on the NP surface, and may also depend on the conjugation of the redox-active species with an electron withdrawing group such as the nanoparticle core. The electrochemical study not only confirms the presence of an aminosilane coating on the nanoparticle surface, but most importantly demonstrates that the amino groups are free for successive reactions.

In cyclic voltammetry, the current value of a specific process is proportional to the redoxactive species concentration. This consideration suggests a means for estimating the number of APTMS molecules on the NP surface. The current peak height of 2.3 × 10⁻³ mol dm⁻³ APTMS in CH₂Cl₂ solution is 52 ± 6 µA, and that of 0.023 g in 0.020 L of NP-NH₂ in DMSO solution is 32 ± 4 µA, so we can therefore calculate a (1.4 ± 0.3) × 10⁻³ mol dm⁻³ concentration of NH₂ groups in NP-NH₂ solution. With an average particle size of 60 nm measured by AFM (as discussed in the next paragraph) and a cobalt ferrite density of 5.15 g cm⁻³, we can further estimate that the number of APTMS molecules bonded to the surface was about (4 ± 1) × 10⁵ per nanoparticle. (assuming spherical nanoparticles and negligible aminosilane coating contribution to nanoparticle density). Moreover, for a typical Si-O bond length of 0.16 nm, it is possible to calculate a thickness of APTMS on the nanoparticle surface of about 2.5 ± 0.5 layers, very close to XPS and the literature data [22,23,24,25,26].

Synthesis of the hybrid hydrogel requires a uniform distribution of the nanoparticles—The cross-linker agents-in the reaction mixture. Figure 4 shows the dispersions of bare (left) and silanized (right) nanoparticles. The instability of bare nanoparticles, which tend to aggregate and sediment, is clearly evident. The introduction of a coating of silane molecules on the surface prevents the agglomeration and makes for a more stable suspension [27,28]. As previously reported, one factor that may influence the dimension of the aggregates is sonication [29,30]. This effect in NP-NH₂ dispersion is depicted in the AFM images in Figure 5.

Figure 5 shows the morphology of bare nanoparticles (a); unsonicated silanized suspension (b); and silanized suspension sonicated for two different time periods (c = 5 minutes, d = 45 minutes). The images clearly indicate the formation of large aggregates, which explain the instability of the suspension reported in Figure 4. Figure 5b shows silanized nanoparticles without sonication; here, there are fewer aggregates of smaller dimensions than bare nanoparticles. In Figure 5c, the suspension after 5 minutes of sonication shows smaller nanoparticle dimension and a better dispersion. The same trend is shown in Figure 5d (45 minutes of sonication), where the absence of aggregates and more uniform NP-NH₂ distribution is evident. Moreover, particle dimension and distribution remain unchanged even after an entire day, as determined by further AFM analysis. The average diameter of the more dispersed nanoparticles measured by AFM analysis is 60 nm.

The strategy for the synthesis of CMC-NP involves the formation of amide bonds between the carboxylic groups of CMC and the amine groups of the functionalized NPs in the presence of EDC and NHS [12,17]. A schematic diagram of the reaction is shown in Figure 6. EDC reacts with a carboxyl group to form an amine-reactive O-acylisourea intermediate, which allows reactions with the primary amine of the NP-NH₂.

The CMC-NP hydrogel (Figure 7) shows a homogenous coloration, indicating an even distribuition.

The results of the rheological characterizations of CMC-NP were compared with those of the CMC hydrogel cross-linked with 1,3-diaminopropane (CMC-DAP), a bifunctional organic compound used to cross-link polysaccharides without the use of nanoparticles (Table 1) [1]. Stress sweep test analyses were performed to measure the elastic (G') and viscous (G") moduli of the hydrogels samples. The dynamic mechanical analysis indicated a “gel-like” behavior, with the G' greater than the G" within the frequency range analyzed for both the hydrogels. The difference between the elastic modulus for the CMC-NP hydrogel (G' = 3,000 ± 300) and the CMC-DAP hydrogel (G' = 550 ± 100), i.e., the hydrogel obtained crosslinking the CMC chains by 1,3 diamino propane has to be attributed to the presence of NPs in the hydrogel matrix. As for nanocomposite, the enhancement of the above-mentioned properties is strongly dependent on the interactions between filler and polymer. The introduction of an inorganic phase in a hydrogel matrix determines enhanced mechanical proprieties. The increased value of G' can be explained by the good distribution of NP-NH₂ that leads to the formation of an ordinate structure, which is reflected in the increased value of the elastic modulus [31,32]. Instead, the loss modulus shows no significant difference between the samples, it means that their viscosity is almost the same. The viscosity is mainly decided by the continuous phase, which is represented by the CMC hydrogel in both the samples [33].

In order to verify whether the CMC-NP hydrogel is capable of working as a system for controlled drug release by means of the application of AMF, some samples containing magnetic nanoparticles (0.3 mg) into the lyophilized hydrogel (1 mg) were immersed in a solution containing 2 mg/mL of methylene blue as a drug model. The hydrogel thus loaded with large amounts of drug model turned blue. One set of the loaded hydrogel was placed in a cuvette and left with no exposure to the AMF. A second set of loaded hydrogel was placed in a cuvette, put in the center of the solenoid and continuously subjected to AMF for several hours. Cumulative release of the model drug for both samples was quantified by UV spectrophotometry measuring the absorbance of the diagnostic peak at 665 nm at regular time intervals. The results are shown in Figure 8, where cumulative drug release vs. time is plotted for the samples under study. The experimental results show a greater drug release at all time intervals for the sample exposed to AMF than that of the unexposed sample. This improved release is caused by the alignment of the nanoparticles’ magnetic moments with that of the external magnetic field, the continuous alteration of which probably leads to the expansion of the hydrogel networks and allows a greater number of methylene blue molecules to leave the gel matrix and enter the medium. No heating effect was observed in our measurements even if the time of AMF applications is rather long (10 hours). The release must depend on a different mechanism than the thermal effect which is reported in literature as being responsible for the delivery of drug for similar systems with free NP inside [10,13,14].

Extensive analyses are in progress to gain further insight into the effects of the alternating magnetic field on drug release, in order to explain the release mechanism.

The sodium salt carboxymethylcellulose (CMC, molecular weight: 700 kDa, with a degree of carboxymethylation of 0.86) was supplied by Sigma-Aldrich. Commercial water dispersion of cobalt ferrite NPs was provided by Colorobbia (Italy). The NPs have a CoFe₂O₄ stoichiometry. Silane coupling agent 3-aminopropyltrimethoxysilane (APTMS), N-hydroxysuccinimide (NHS), N-(3-dimethylaminopropyl)-N-ethylcarbodiimidehydrochloride (EDC), 1,3-diaminopropane (DAP), and all the reagents were purchased from Sigma-Aldrich.