**1. Introduction**

Spintronics open a new scenario for the fabrication of e fficient devices and refer to electronics based on the electron spin, which is typically controlled by magnetic fields or by ferro-/paramagnetic materials. An important prerequisite for realizing operative spintronic devices is to achieve high spin injection coe fficient materials. With this aim, spin filters characterized by the coexistence of one spin conducting and one insulating channel are required. Organic molecules were proposed as a spin filter for "molecular spintronics" due to their relatively longer coherent spin lifetimes and spin transport distances, resulting from weak spin–orbit coupling and weak hyperfine interactions [1,2]. The latter may allow the development of nanoscale devices with improved performance or new functionalities. Porphyrin derivatives, which have already been investigated for molecular electronic devices due to their peculiar electronic and optical properties, have been recently proposed as a class of molecules

that is particularly promising as a building block for spintronic technology [3,4]. One-dimensional chromium porphyrin arrays showing half-metallic behavior [5], manganese porphyrin molecules connected with a p-phenylene-ethynylene group [6], and porphyrin/graphene hybrid materials [7] have been reported so far. Theoretical and experimental investigations allowed demonstrating that metallic substrates are able to induce magnetic ordering and the switching of paramagnetic porphyrins, which is due to a superexchange interaction between Fe atoms in the chromophores and Co or Ni atoms in the substrate [8]. Spin-dependent transport properties in an iron-porphyrin such as a carbon nanotube have been investigated, reporting a magnetoresistance ratio that is strongly dependent on the magnetic configuration of the system [9]. Recently, a new promising effective approach for spintronics has emerged using spin selectivity in electron transport through chiral molecules [10–14]. This effect, defined as chiral-induced spin selectivity (CISS), is due to the special property of chiral symmetry that couples the electron spin and its linear momentum acting as a spin filter depending on the handedness of the molecules [15]. The spin-polarized electron current due to the CISS effect can be used to magnetize ferromagnets, potentially allowing the fabrication of less expensive and high-density devices. A DNA double helix [16,17], as well as chiral molecules with (i.e., oligopeptides or chiral polymers) [12,18–21] or without (helicenes) [22] stereogenic carbon centers have been reported so far. The driving of electrons through chiral layers has been reported for α helix L-polyalanine and CdSe nanocrystals by local light-induced magnetization [23] and for self-assembled monolayers of polyalanine to magnetize a Ni layer [24]. However, the development of this technology is hindered by the fact that solid thin films and nanostructures need a precise control of homogeneity, morphology, and chirality. Spatially uniform chiral films [25,26], supramolecular structures with a programmed helicity [27,28], artificial assemblies [29], and the patterning of chiral nanostructures [30,31] have been developed up to now. These latter showed significant improvements toward technological applications, even if the uniformity and spatial control of the local handedness of chiral self-assembled systems on a surface is still a fundamental open challenge. In this framework, we already investigated the self-assembly of the achiral water soluble 5,10,15,20-tetrakis-(4-sulfonatophenyl)-porphyrin (TPPS) into chiral aggregates on a substrate by combining a wet lithographic method with the local induction of specific chirality imprinted by a chiral templating agen<sup>t</sup> [32]. It is well known that the diacid form of this porphyrin in solution under the opportune experimental conditions is able to self-organize into highly ordered chiral J-aggregates with or without the assistance of chiral species [33–35]. Indeed, in the absence of a chiral bias, TPPS could be assembled into chiral aggregates showing a dichroic signal characterized by a positive bisignate Cotton effect [34,36–41] whose shape and magnitude is strictly related to the experimental conditions used for the preparation of the aggregates [34,37,42–46]. The formation of chiral supramolecular assemblies from achiral building blocks is a particularly interesting phenomenon [47–51] for the possible correlation with homochirality observed in nature. In many bioprocesses, the interactions between molecules induce a redistribution of the electronic charge accompanied in a chiral system by spin polarization. It has been experimentally demonstrated that this spin polarization adds an enantioselective term to the forces, thus leading to homochiral interaction energies different from heterochiral ones [12]. Herein, we report on the role of organic capped superparamagnetic iron oxide nanoparticles (SPIONs), after their incorporation in the hydrophobic core of polyethylene glycol (PEG)-modified phospholipid micelles (SPIONs@micelles) on the kinetic and spectroscopic behavior of porphyrin J-aggregates, pointing to the formation of a hybrid system. We anticipate that the SPION-loaded micelles are able to efficiently trigger the growth of TPPS J-aggregates under proper experimental conditions and reagen<sup>t</sup> mixing order protocol. Moreover, this assembling strategy allows conjugating the magnetic properties of the SPIONs@micelles with the chiral optical properties of the J-aggregates, thus obtaining hybrid architectures in solution and on solid state by means of an applied magnetic field. By the application of an external magnetic field, SPIONs show magnetization values comparable and magnetic susceptibility much higher than bulk paramagnetic materials [52]. However, they exhibit high coercivity or residual magnetization value with zero magnetic field [53], anisotropy, and large magnetic saturation values [54]. For these magnetic

properties and their good biocompatibility, iron oxide nanoparticles (NPs) have been largely reported for biomedical applications [55,56] and in particular as contrast agents for magnetic resonance imaging (MRI) [52,57], in drug delivery [58,59], magnetic hyperthermia [60], magnetically assisted genetic transfection [61], and/or in combined therapeutic and diagnostic use (theranostics) [62]. To the best of our knowledge, no examples of SPIONs involved in the self-assembly of porphyrins in solution and on the solid state of chiral porphyrin aggregates have been reported so far. Besides, iron oxides thin films with an Fe<sup>3</sup>+/Fe2<sup>+</sup> ratio of 2 have been reported as a promising candidates for spintronic applications, showing high saturation magnetization and magnetic susceptibility [63]. The merging of the NPs magnetic properties and chiroptical properties of the porphyrin aggregates makes the hybrid SPIONs@micelle J-agg assemblies interesting candidates for a variety of potential applications, ranging from optics to electronics and spintronics.

#### **2. Materials and Methods**

*Chemicals.* Oleic acid (90%), dodecan-1,2-diol (90%), iron pentacarbonyl (98%), oleyl amine (70%), and 1-octadecene (90%) were purchased from Sigma-Aldrich. 1,2-dipalmitoyl-sn-glycero-3- phosphoethanolamine-N-[methoxy (poly(ethylene glycol))-2000] (16:0 PEG-2-PE, ammonium salt) and amine 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[amino(polyethylene glycol)-2000] (DSPE-PEG-amine, ammonium salt) were from Avanti Polar Lipids. The 5,10,15,20-tetrakis-(4 sulfonatophenyl)porphyrin (TPPS) was purchased from Aldrich Chemicals, and its solutions of known concentration were prepared using the extinction coefficient at the Soret maximum (ε = 5.33 × 10<sup>5</sup> M−1cm−<sup>1</sup> at λ = 414 nm). All the reagents were used without further purification, and the solutions were prepared in dust-free Milli-Q water.

Preparation of PEG-modified phospholipid micelles loaded with SPION. Organic capped SPIONs were synthesized according to the experimental protocols reported in the literature [64]. For the preparation of the micelles loaded with SPIONs and composed only of PEG-2-PE (SPIONs@PEG-micelles), 150 μL of PEG-2-PE in chloroform (3.5 × 10−<sup>2</sup> M) were mixed with 240 μL of a SPION stock chloroform dispersion (0.08 M); while for the preparation of amine functionalized superparamagnetic micelles (SPIONs@NH2-PEG-micelles), 120 μL of PEG-2-PE (3.5 × 10−<sup>2</sup> M) and 30 μL of DSPE-PEG-amine (3.5 × 10−<sup>2</sup> M) were used. After the complete evaporation of organic solvent, the SPION/PEG-modified lipid film was treated with 2 mL of phosphate buffer (PBS, 10 mM, pH 7.4). Three consecutive cycles of heating–cooling, at 80 ◦C and room temperature respectively, were carried out in order to obtain SPION-loaded micelles. The excess of organic capped SPIONs, not eventually encapsulated in the micelles, was removed by mild centrifugation at 5000× *g* for 1 min; the empty micelles were subsequently removed by ultracentrifugation (200,000× *g*) for 16 h. The SPIONs@micelles that were recovered as pellets were resuspended in water, filtered by using 0.2 μm filters (Anotop, Whatman, Merck, Italy), and lyophilized. Water or PBS was used to reconstitute the dried micelles before their characterization or application [65,66].

Aggregation and deposition procedure. TPPS@SPIONs@micelle J-aggregates were prepared in 0.3 M HCl following two different mixing protocol procedures: (i) porphyrin-first protocol (PF), consisting of the addition of a proper volume of acid stock solution to a diluted solution of porphyrin and SPION-loaded micelles, and (ii) porphyrin-last protocol (PL), in which a known amount of porphyrin stock solution is added to diluted magnetic micelles in acidic solution. The concentration of magnetic micelles is reported in terms of the phospholipid monomer concentration. This has been calculated taking into account a weight percentage of about 74% for the phospholipid into the micelles [66]. For the amine-terminated phospholipid, the PEG-PE/DSPE-PEG-amine mixture is in a 4:1 ratio. Glass slides, carefully cleaned with Piranha acidic solution, were immersed into 3 mL of solution containing TPPS@SPIONs@micelle J-aggregates. We used a neodymium magne<sup>t</sup> cube (1 cm × 1 cm × 1 cm) placed below the glass slides to deposit aggregates onto glass surface from the solution. After an overnight aging time at room temperature, the slides were washed by quick immersion in aqueous acidic solution and dried under a gentle nitrogen flow.

Spectroscopic and morphological characterization. UV-Vis spectra were collected on a diode-array spectrophotometer Agilent model 8452. Kinetic experiments were carried out in the thermostated compartment of the spectrophotometer, with a temperature accuracy of ± 0.1 K. The analysis of the kinetic profiles has been performed by a non-linear fit of the absorption data according to Equation (1):

$$E\_{\rm xt} = E\_{\rm xt\,co} + (E\_{\rm xt\,0} - E\_{\rm xt\,co})\left(1 + (m - 1)|k\_0 t + (n + 1)^{-1}\left(k\_\circ t\right)^{n+1}\right)^{-1/(m-1)}\tag{1}$$

with *Ext*0, *Ext* ∞, *k*0, *k*c, *m* and *n* as the parameters to be optimized or Equation (2):

$$E\_{\rm xt} = E\_{\rm xt0} + (E\_{\rm xtso} - E\_{\rm xt0}) \left( \exp(-(kt)^{\rm r}) \right) \tag{2}$$

with *Ext*0, *Ext* ∞, *k* and *n* as the parameters to be optimized (*Ext*, *Ext*0 and *Ext* are the extinction at time t, at starting time, and at the end of aggregation, respectively). The circular (CD) and linear (LD) dichroism spectra were recorded on a JASCO J-720 spectropolarimeter equipped with a 450 W xenon lamp. The LD spectra under an applied magnetic field have been recorded by setting a couple of neodymium magnets (1 cm × 1 cm × 1 cm) close to the cuvette walls in a perpendicular direction with respect to the light beam. CD and LD spectra were corrected both for the cell and solvent contributions.

Fluorescence emission and resonance light scattering (RLS) experiments were performed on a Jasco mod. FP-750 spectrofluorimeter. A synchronous scan protocol with a right angle geometry was adopted for collecting RLS spectra [67], which were not corrected for the absorption of the samples. Time-resolved fluorescence emission measurements were performed on a Jobin Yvon-Spex Fluoromax 4 spectrofluorimeter using the time-correlated single-photon counting technique. A NanoLED (λ = 390 nm) has been used as the excitation source.

For the transmission electron microscopy (TEM) investigation, a Jeol JEM-1011 microscope, working at an accelerating voltage of 100 kV, was used. TEM micrographs were acquired by an Olympus Quemesa Camera (11 Mpx). A total of 400 mesh amorphous carbon-coated Cu grids were dipped in SPION chloroform dispersion to achieve the sample deposition.

Hydrodynamic particle sizes and size distributions were measured by Dynamic Light Scattering (DLS)and carried out at 25 ◦C by a Zetasizer Nano-ZS (Malvern Instruments) equipped with a 633 nm He−Ne laser using backscattering detection. Each DLS sample was measured several times, and the results were averaged.

Atomic force microscopy (AFM) and magnetic force microscopy (MFM) measurements were performed using a NT-MDT Smena head working in tapping mode and equipped with a CoCr coated tip (mod. MFM01). To collect the MFM data, the instrument was configured to work in the "double pass AC Magnetic Force" mode. In this working mode, the system produces two images; the first consists of the bare morphology collected line by line in a standard non-contact scanning mode. In the second step, the tip is uplifted at some tenths of a nanometer from the surface, and the phase shift due to the magnetic interaction between the tip and the sample is recorded. During the second step, the profile data collected in the first step are used to assure that the distance between the tip and the sample is constant. To improve the sensitivity, before the measurement, the tip was exposed to the field of a neodymium magne<sup>t</sup> for a few hours.
