**3. Results**

The TPPS porphyrin is present in neutral aqueous solution as free base characterized by the presence of a Soret band centered at 414 nm and four Q-bands between 500 and 700 nm in the UV/Vis spectrum. Under acid conditions, the protonation of the pyrrole nitrogen atoms of the central core occurs (pKa: approximately 4.9), with the formation of a protonated species showing a Soret band at 434 nm and two Q-bands. This diacid specie is able to self-arrange in J-aggregates characterized by a linear arrangemen<sup>t</sup> of the chromophores and stabilized by electrostatic interactions between the negatively charged benzenesulfonate groups and the positively charged nitrogen atoms of the pyrrole rings, as well as by hydrogen bonds and stacking interactions [41,68–74]. The formation in solution of these aggregates leads in the absorption spectrum to the formation of a new band, bathochromically shifted (Δλ: approximately 50 nm) with respect to the monomeric species. Their aggregation kinetics can be influenced by different experimental parameters such as the nature of the acid [45], the ionic strength [41,74], or the reagen<sup>t</sup> mixing order protocol [34,72,73]. In particular, this latter influences the dynamics of the growth and eventually both the morphology and size of the assemblies [73]. Here, we investigated on the ability of SPION-loaded micelles to efficiently promote the formation of hybrid TPPS@SPIONs@micelle J-aggregates. In particular, two different magnetic micelles were prepared to investigate the self-assembly process. The aqueous dispersibility of the hydrophobic SPIONs was guaranteed by the formation of SPIONs encapsulating micelles composed of 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy (poly(ethylene glycol))-2000] (PEG-2-PE), or a mixture of PEG-2-PE and amine 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[amino(polyethylene glycol)-2000] (DSPE-PEG-amine], bearing PEG chains and amine terminated PEG chains, respectively (Figure 1) [66,75]. Indeed, magnetic micelles composed only of PEG-2-PE (SPIONs@PEG-micelles, Figure 1A) and amine-functionalized magnetic micelles based on a PEG-2-PE/ DSPE-PEG-amine mixture (SPIONs@NH2-PEG-micelles, Figure 1B) were obtained by the encapsulation of a certain number of organic-capped SPIONs having an average diameter of about 9 nm (Figure 1C) clustered in the hydrophobic core into the single micelle.

**Figure 1.** Schematic representation of superparamagnetic iron oxide nanoparticles, after their incorporation in the hydrophobic core of polyethylene glycol (PEG)-modified phospholipid micelles (SPIONs@PEG-micelles) (**A**)- and amine functionalized superparamagnetic micelles (SPIONs@NH2-PEG-micelles) (**B**), along with the corresponding legend. TEM micrograph of organic capped SPIONs (**C**).

DLS measurement was carried out on SPIONs@PEG-micelles to test their stability. From the analysis of the data, the micelles both in neutral and acidic aqueous solution show an average RH of 150 (±10) nm. When SPIONs@PEG@micelles (15 μM) were added to a solution of TPPS at neutral or mild acidic pH, no evidence of modification in the spectroscopic behavior of the chromophore, in terms of absorption (inset of Figure 2), fluorescence emission, fluorescence lifetimes, and time-resolved fluorescence anisotropy has been observed (data not shown). The experimental evidence excludes a preinteraction among porphyrins and magnetic micelles before aggregation. In order to promote the formation of supramolecular assemblies among porphyrin aggregates and magnetic micelles, we lowered the pH (0.3 M HCl) following the two previously mentioned mixing order protocols. Using a PF protocol, although the J-aggregates electronic spectrum remains unchanged (Figure 2), their formation kinetics are conversely deeply influenced by the presence of SPIONs@PEG-micelles. A kinetic analysis of the extinction/time traces on the aggregate band (491 nm) has been performed by using an autocatalytic model already reported in the literature and largely reported for kinetic investigations on porphyrin aggregation [34,39,44,45,76].

**Figure 2.** UV-vis spectra of SPIONs@PEG-micelles (black line) and chiral J-aggregates formed by 5,10,15,20-tetrakis-(4-sulfonatophenyl)-porphyrin (TPPS) embedded into the magnetic micelles (TPPS@SPIONs@PEG-micelles) in aqueous solution, porphyrin first (PF) (red line) and porphyrin last (PL) mixing protocol (green line). Experimental conditions: [SPIONs@PEG-micelles] = 15 μM, [TPPS] = 5 μM, [HCl] = 0.3 M. For comparison, TPPS@SPIONs@PEG-micelles at pH 3 is reported in the inset.

The kinetic profiles exhibit a sigmoidal behavior characterized by the presence of an initial induction period that was shortened in the presence of magnetic micelles, (Figure 3A) with an observed rate constant *k*c that is about an order of magnitude larger than that observed for the pure system. All the kinetic parameters are summarized in Table 1.

**Figure 3.** Extinction kinetic traces at 491 nm with (black line) and without (red line) SPIONs@PEG-micelles. The solid lines are the best-fitting of the experimental data (λ491 nm) to Equation (1) (**A**) and Equation (2) (**B**). The best-fitting parameters are collected in Table 1. Experimental conditions: [SPIONs@PEG-micelles] = 15 μM, [TPPS] = 5 μM, [HCl] = 0.3 M PF, *T* = 298 K mixing order protocol (**A**), PL mixing order protocol (**B**).

An increase of the aggregation kinetic rate has been observed also for the PL mixing order protocol. According to the literature, the PL protocol induces a dramatic difference in the kinetic profiles, which now obey a stretched exponential form (Equation (2)). The presence of magnetic micelles further increases the rate constants of the aggregation process (k), whose value goes from 1.8 × 10−<sup>2</sup> s<sup>−</sup><sup>1</sup> to a 2.7 × 10−<sup>2</sup> s<sup>−</sup><sup>1</sup> (Figure 3B). It is noteworthy that unlike the previous experimental protocol, in this case, the electronic spectrum of the final aggregates shows a slight bathochromic shifted and widened J-absorption (Figure 2). DLS experiments point to the presence in solution of particles having a *RH* value of approximately 1 μm. Since it is known that TPPS aggregates in solution in analogous experimental conditions are nanorods sizing hundreds of nanometers [41], our experimental findings sugges<sup>t</sup> the formation of a hybrid system formed by porphyrin aggregates and magnetic micelles. As TPPS J-aggregates can undergo spontaneously to symmetry breaking, circular dichroism spectra have also been collected. At the end of aggregation kinetics for the hybrid system obtained by PL protocol, the CD spectrum shows a profile with a positive bisignate Cotton effect centered at the aggregate absorption band that is much higher with respect to the neat sample (Figure 4).

**Figure 4.** Circular dichorism (CD) spectra of J-aggregates in solution (**A**) and on solid state (**B**) in the presence (black line) or in the absence (red line) of SPIONs@PEG-micelles. Experimental conditions: [SPIONs@PEG-micelles] = 15 μM, [TPPS] = 5 μM, [HCl] = 0.3 M, PL mixing order protocol.

It is interesting to note that according to the literature, for "pure" TPPS J-aggregates, whatever the pretreatment of the samples, the dissymmetry g-factor generally decreases on increasing the value of the kinetic rate constant [44]. On the contrary, for the hybrid system here reported, we observe an increase on the CD intensity signal, despite the increase of the aggregation rate constant. Taking advantage of the magnetic properties of the SPIONs@PEG-micelles, we used a magne<sup>t</sup> field below the sample to try to deposit aggregates obtained by PF and PL protocols, respectively, onto glass surface from the solutions. After aging overnight at room temperature, the formation of a green film from the PL samples and of a dark orange precipitate from the PF protocol samples respectively, is evident. As a blank experiment, the magne<sup>t</sup> field has been applied also for the aggregate solutions in the absence of magnetic micelles and, as expected, no deposition has been observed. The extinction spectrum of the solid sample shows unequivocally the presence of the characteristic J-aggregate band (Figure 5), whereas in solution, only the presence of porphyrin in its diacid monomeric form has been observed (Figure 5, inset). Interestingly, the CD spectrum of the film shows the chiroptical properties observed in solution with an induced bisignate Cotton effect in the aggregate absorption region (Figure 4B). This dichroic signal cannot be ascribed to linear dichroism, as no variation due to the sample orientation with respect to incident light has been observed.

**Figure 5.** Extinction spectrum of TPPS@SPIONs@PEG-Micelle J-aggregates deposited on the glass surface by applying a magnetic field through a neodymium magne<sup>t</sup> cube 1 cm × 1 cm × 1 cm and residual specie in solution (inset). Experimental conditions: [SPIONs@PEG-Micelles] = 15 μM, [TPPS] = 5 μM, [HCl] = 0.3 M, PL mixing order protocol.

These experimental findings prove the formation of a hybrid system only for the PL protocol. So, in our case, the reagen<sup>t</sup> mixing order affects not only the kinetic rates of the self-assembly processes but also the formation of the hybrid supramolecular system. This sort of "YES/NO" effect has already been reported for TPPS J-aggregates obtained in the presence of polyamines containing less than three protonable nitrogen atoms [73]. In particular, the nucleation step is the critical parameter affecting both the kinetic and mesoscopic structure of the resulting aggregates [34,77]. In the present case, we hypothesize that the PL protocol, due to a concentration effect, induces a very quick formation of a larger number of porphyrin seeds that can be entrapped in peripheral hydrophilic chains of the SPIONs@PEG-micelles, leading to hybrid aggregates. On the contrary, in the PF protocol, a longer nucleation period allows the organization of porphyrins in much larger aggregates, thus preventing their entrapment in the magnetic micelles.

In order to characterize the morphology of the deposited sample on a glass surface, we performed atomic force microscopy (AFM). Figure 6 shows the sample topography and the line profile acquired along the blue line, as depicted on the picture. The morphology consists of an almost homogeneous layer of compactly arranged nanostructures, whose dimension ranges between 100 and 150 nm. It is worth noting that a preferential direction is present; this alignment is probably due to the magnetic field used during the deposition method. Even if no clear detection of the TPPS aggregates is possible, their embedding into the film is confirmed by the absorption of the sample (Figure 5).

**Figure 6.** Atomic force microscopy (AFM) topography images (**a**) and relative profile (**b**) of the TPPS@SPIONs@PEG-micelle J-aggregates deposit on a glass substrate after applying a magnetic field.

As the interaction among TPPS porphyrins and functional molecules bearing protonable nitrogen atoms such as polyamines or PAMAM dendrimers has been already reported in aqueous solution [72,73,78–80], we exploited amine functionalized magnetic micelles based on a PEG-2-PE/ DSPE-PEG-amine mixture (SPIONs@NH2-PEG-micelles) to trigger the aggregation process. The presence of protonable nitrogen atoms in the periphery of the micelles should be useful for interaction with the negatively charged groups present on the porphyrin ring. Accordingly, we tested the preinteraction between the chromophoric unit in its monomeric neutral and diacid forms with SPIONs@NH2-PEG-micelles at different stoichiometric ratios. In all the investigated samples, no changes on the spectroscopic behavior of the porphyrin have been observed so excluding the presence of the chromophoric units embedded in the micelles or entangled in their peripheral chains, even if some interaction with the peripheral counter-cations cannot be ruled out. As for the previous system, the stability of the SPIONs@NH2-PEG-micelles under the investigated experimental conditions has been tested by UV/Vis and DLS measurements. This latter technique evidences the presence in solutions of objects with RH of about 140 (±10) nm. The aggregation process in the presence of TPPS has been induced by employing the two different mixing order protocols, PF and PL, respectively. When PL protocol is adopted on SPIONs@NH2-PEG-micelles under acidic conditions (HCl 0.3 M), J-aggregates form almost instantaneously. The resulting solutions are highly unstable, and precipitation occurs in a very short time. DLS investigations confirm the presence of micrometric objects, thus justifying the experimental observations. Due to the scarce reproducibility of these samples, we decided to not further investigate them. On the contrary, when PF protocol is used, the aggregation is fostered by the addition of acidic solution of SPIONs@NH2-PEG-micelles at different concentrations; the extinction and kinetic features are reported in Figure 7. In particular, the inset shows a set of kinetic traces obtained from the extinction increase at 491 nm (Figure 7). These curves display a sigmoidal profile characterized by a nucleation early stage, which shortens upon increasing the concentration of magnetic micelles in solution (inset Figure 7). The values of the parameter *m* (~3), that is the critical size for the nuclei, and n (~3−4), the time exponent, are comparable to those reported in the literature for similar systems in aqueous solutions [34]. Differently, the values of the rate constants for the uncatalyzed pathway, *k*0, and those for the catalyzed pathway, *kc*, in the presence of magnetic micelles are higher with respect to the pure self-assembled system and increase, exhibiting an exponential dependence on SPIONs@NH2-PEG-micelles concentration. All the kinetic parameters are collected in Table 1. After equilibration, the UV–Vis spectra provide evidence that the amount of the final aggregate increases with the concentration of SPIONs@NH2-PEG-micelles.

**Figure 7.** Extinction spectra for TPPS@SPIONs@NH2-PEG-micelles J-aggregates formed at different SPIONs@NH2-PEG-micelles concentrations. [SPIONs@NH2-PEG-micelles] = 15 μM (red line), [SPIONs@NH2-PEG-micelles] = 22 μM (blue line), [SPIONs@NH2-PEG-micelles] = 45 μM (green line). In the inset, the extinction kinetic traces at 491 nm and the best fitting experimental data according to Equation (1). Experimental conditions: [TPPS] = 5 μM, [HCl] = 0.3 M PF mixing order protocol, *T* = 298 K.

The aggregated samples are stable in solution, and DLS measurements show the presence of objects with R H values of about 700 (±50) nm. The CD spectra recorded at the end of the aggregation process show in the absence of magnetic nanoparticles the typical positive bisignate spectrum, whereas unusual profiles can be observed in the presence of SPIONs@NH2-PEG-micelles at di fferent porphyrin/magnetic micelles concentration ratios (Figure 8a). A generally inverted and consistently bathochromically shifted band is present at the absorption of the J-component, while the H-band shows the usual positive bisignate feature, even if it is slightly red shifted. Since linear dichroism (LD) could be responsible for the observed e ffects in the case of alignment of the aggregates, we performed LD measurements on the same TPPS@SPIONs@NH2-PEG-micelles J-aggregates (Figure 8b). The LD spectra show for all the samples the presence of a positive band at 491 nm and a less intense and negative one at 420 nm, similarly to those reported for analogous systems, where the J-aggregates were aligned by flow. The presence of di fferent signs for the two LD bands has been attributed to the di fferent polarization within the porphyrin aggregate of the J and the H electronic transitions [81].

**Figure 8.** CD (**a**) and linear dichorism (LD) (**b**) spectra for TPPS@SPIONs@NH2-PEG-micelles J-aggregates at di fferent SPIONs@NH2-PEG-Micelles concentrations. In the absence of SPION-NH2 (black line) [SPIONs@NH2-PEG-Micelles] = 15 μM (red line), [SPIONs@NH2-PEG-Micelles] = 22 μM (blue line), [SPIONs@NH2-PEG-Micelles] = 45 μM (green line), and under an applied magnetic field perpendicularly to the light path (green dashed line). Experimental conditions: [TPPS] = 5 μM, [HCl] = 0.3 M, PF mixing order protocol.


**Table 1.** Kinetic parameters *k*0, *kc*, *m,* and *n* for the aggregation of TPPS@SPIONs@PEG-micelles and TPPS@SPIONs@NH2-PEG-micelles as a function of the mixing order protocol and/or magnetic micelles concentration (*T* = 298 K). [a] Data obtained according to Equation (2); [b] Data obtained according to Equation (1).

This experimental evidence suggests that the obtained aggregates being large enough are somehow aligned, probably by e ffect of barodi ffusion. Taking advantage of the magnetic property of TPPS@SPIONs@NH2-PEG-micelles J-aggregates, a further LD spectrum was taken on the solution containing the larger amount of NPs by applying an external magnetic field perpendicularly to the direction of the optical beam during the measurement. In comparison with the LD spectrum recorded at zero field (Figure 8b, full green line), the profile in the presence of the magnetic field exhibits a twofold increase in the intensity of the linear contributions with no alteration of the shape (Figure 8b, dashed green line), thus suggesting an enhanced alignment effect. All our experimental findings point to the formation of a hybrid system, evidencing the role of the magnetic NPs on their kinetics of growth and final structure. In this case, also deposition on a glass surface has been achieved by applying a magnetic field, and the electronic spectrum of the film shows the presence of the J-aggregates band (Figure 9).

**Figure 9.** Extinction spectrum of TPPS@SPIONs@NH2-PEG-micelles J-aggregates deposited on a glass surface by applying a magnetic field under the solid substrate. Experimental conditions: [SPIONs@NH2-PEG-micelles] = 45 μM, [TPPS] = 5 μM, [HCl] = 0.3 M, PF mixing order protocol.

Figure 10 displays the morphology (a) and the magnetic response (b) of this sample obtained through MFM. The morphology consists of an almost homogeneous compact layer of nanoparticles. The surface is also populated by a small number of aggregates with larger sizes. The MFM map reveals a general magnetic activity, which is highlighted by the signal modulation distributed over almost all the sample surface, as the intrinsically limited resolution of the technique is not able to map the magnetic response of the smallest nanoparticles. The bigger aggregates clearly show a much stronger magnetic signal. Interestingly, some particles show a bright contrast, while others highlight a dark contrast. This behavior is due to the different interaction with the magnetic tip that operates in tapping mode at its resonance frequency: a repulsive magnetic force gradient will cause the resonance curve to shift to a higher frequency, an increase in phase shift, and a bright contrast. Conversely, an attractive magnetic force gradient results in the resonance curve shifting to a lower frequency, and a decrease in phase shift resulting in a dark contrast. Some particles with no magnetic activity are also observed. Therefore, all the spectroscopic and microscopy evidence suggests that the deposited film is constituted by the hybrid TPPS@SPIONs@NH2-PEG-micelles J-aggregates.

**Figure 10.** AFM images for TPPS@SPIONs@NH2-PEG-micelles J-aggregates deposited on glass by applying a magnetic field under the solid substrate. Topographic (**a**) and magnetic (**b**) images. Experimental conditions: [SPIONs@NH2-PEG-Micelles] = 45 μM, [TPPS] = 5 μM, [HCl] = 0.3 M, PF mixing order protocol.
