Photo-induced electric polarizability of Fe3O4 nanoparticles in weak optical fields

Using a developed co-precipitation method, we synthesized spherical Fe3O4 nanoparticles with a wide nonlinear absorption band of visible radiation. Optical properties of the synthesized nanoparticles dispersed in an optically transparent copolymer of methyl methacrylate with styrene were studied by optical spectroscopy and z-scan techniques. We found that the electric polarizability of Fe3O4 nanoparticles is altered by low-intensity visible radiation (I ≤ 0.2 kW/cm2; λ = 442 and 561 nm) and reaches a value of 107 Å3. The change in polarizability is induced by the intraband phototransition of charge carriers. This optical effect may be employed to improve the drug uptake properties of Fe3O4 nanoparticles. PACS 33.15.Kr 78.67.Bf 42.70.Nq

In fact, Fe 3 O 4 nanoparticles have been examined for the presence of unique magnetic properties because magnetite is a narrow-gap semiconductor [20][21][22] and the optical properties of other semiconductor nanoparticles have been thoroughly studied. Currently, there are several experimental and theoretical works dedicated to studying the optical properties of both bulk magnetite [23][24][25][26] and its nanoparticles [27][28][29]. However, some specific optical properties of Fe 3 O 4 nanoparticles (in particular, the effects of electric polarizability on their biological activity, conductivity, ferroelectricity, and electro-optical properties) as well as the nature of these properties remain virtually unexplored.
In this paper, we demonstrate that Fe 3 O 4 nanoparticles exhibiting a wide nonlinear absorption band of visible radiation (1.7:3.7 eV) are able to significantly change their electric polarizability when exposed to lowintensity visible radiation (I ≤ 0.2 kW/cm 2 ). The observed change in polarizability was induced by the intraband phototransition of nanoparticle charge carriers, and polarizability changes were orders of magnitude greater than those of semiconductor nanoparticles and molecules [30,31].
In the first step (Figure 1a), Fe 3 O 4 nanoparticles were synthesized by co-precipitation of soluble salts of ferrous and ferric ions with an aqueous ammonia solution: Oleic acid (in a mass ratio of 0.7:1 with the formed Fe 3 O 4 ) was added to a 0.5% solution of iron salts (FeSO 4 /FeCl 3 = 1:2.2 molar ratio) in 0.1 M HCl. The aqueous solution of iron salts was heated to 80°C, followed by the addition of concentrated aqueous ammonia (20% excess). The solution was heated and stirred for an hour.
Stabilized nanoparticles were then extracted from the aqueous phase into a nonpolar organic solvent hexane at a ratio of 1:1. The organic layer containing the iron oxide Fe 3 O 4 was separated from the aqueous medium. The sample was centrifuged for 15 min (6,000 rpm) to remove larger particles. Excess acid was removed with ethanol.
The size of the nanoparticles was determined by dynamic light scattering method (Zetasizer Nano ZS, Malvern, UK). Measurements were conducted in hexane with a laser wavelength of 532 nm. The average hydrodynamic diameter of the synthesized nanoparticles was 15 nm, as illustrated in Figure 2.

Composite preparation
The second step (Figure 1b) focused on obtaining a solid composite based on Fe 3 O 4 nanoparticles and MMAS. The organic solvent containing nanoparticles and monomers (methyl methacrylate with styrene) was subjected to stirring and ultrasonic homogenization. To prevent nanoparticle aggregation during the polymerization process, we used the pre-polymerization method at 75°C because the nanoparticles had different affinities to the monomer and polymer.
Finally, the composite was synthesized in situ by radical polymerization. The polymerization of methyl methacrylate with styrene (in the mass ratio of 20:1) proceeded for over 10 h (in a temperature gradient mode that progressed from 55°C to 110°C) in the presence of benzoyl peroxide (10 −3 mol/L).
The obtained solid composites had 0.001%, 0.003%, 0.005%, and 0.01% volume concentrations of Fe 3 O 4 nanoparticles in MMAS. Importantly, the synthesized Fe 3 O 4 nanoparticles generally had a thick layer of acids [36,39] surrounding them to prevent aggregation of the nanoparticle. In our case, the synthesized Fe 3 O 4  nanoparticles had a monolayer of oleic acid that allowed the nanoparticles to exhibit their specific optical properties.

UV-vis spectroscopy
Room-temperature optical absorbance spectra of pure MMAS ( Figure 3, black curve) and of the composites were obtained using a Varian Cary 5000I spectrophotometer (Agilent Technologies, Santa Clara, CA, USA) over the wavelength range of 300 to 1,500 nm. These spectra allowed the derivation of the absorbance spectra for Fe 3 O 4 nanoparticle arrays ( Figure 3, color curves). Figure 3 shows the absorbance values (Abs) and the absorption coefficients (α = (Abs × ln 10)/l, where l = 7.95 mm is the length of the composite) measured at a maximum radiation intensity of 1 μW/cm 2 .

z-Scan experiments
Because they have absorption bands of 380 to 650 nm, Fe 3 O 4 nanoparticles should exhibit an optical response upon external radiation with wavelengths in this band [40]. To detect the optical response of the nanoparticles contained in the composite (0.005% nanoparticle volume concentration), we used the standard z-scan technique [41]. This technique enabled the analysis of changes in the absorption coefficient Δα(I) and refractive index Δn(I) of the composite and pure MMAS, which were induced by weak optical radiation with different intensities 0 to 0.14 kW/cm 2 .
For radiation sources, we used semiconductor lasers of continuous wave (cw) radiation with wavelengths of 442 nm (blue) and 561 nm (yellow) providing maximal intensities of 0.07 and 0.14 kW/cm 2 . Lenses with focal lengths of 75 mm provided the beam waists ω 0 = 102 and 110 μm for blue and yellow radiation (Figure 4b).
The length (L) of experimental samples of the MMAS and the composite was 2.7 mm (inset in Figure 3).
Because the Rayleigh range z 0 = πnω 2 / λ exceeded 10 cm, the calculation of Δα and Δn was performed using the formulae [40,41]: where ΔT(I) (Figure 4a) and ΔT pv (I) (Figure 5b) were the integral transmitted intensity and the normalized  transmittance between the peak and valley at different radiation intensities, respectively; λ and α were the radiation wavelength and absorption coefficient (Figure 3), respectively, and S was the fraction of radiation transmitted by the aperture without the sample, which was 0.184. The experimental curves T(I) and T pv (I), which contain information about ΔT and ΔT pv , showed that only the reverse saturable absorption of yellow radiation occurred in pure MMAS (Figure 4a). In contrast, the composite manifested the expected optical response: the shape of the experimental curves T(I) and T pv (I) indicated the saturable absorption of visible radiation in the composite and a negative change in its refractive index ( Figure 5), and the values of ΔT(I) and ΔT pv (I) increased linearly with increasing intensities of blue ( Figure 5a) and yellow (Figure 5b) radiation.
The approximation of T pv based on the theoretical curves (solid lines in Figure 5) was performed using the equation [42]: where the coupling factor ρ = Δα × λ / 4π × Δn and the phase shift due to nonlinear refraction ΔΦ = 2π × Δn × L eff / λ had the following values: ρ = 0.09 and ΔΦ = −0.23 and −0.5 for blue radiation with intensities of 0.019 and 0.054 kW/cm 2 and ρ = 0.05 and ΔΦ = −0.7 and −1.45 for yellow radiation with intensities of 0.04 and 0.093 kW/cm 2 .

Discussion
The saturable absorption of visible radiation with intensities less than 0.14 kW/cm 2    assumed that Δn was caused by the thermal effect of the radiation. We estimated the contribution of this effect to the changes of the composite refractive index using the equation [43]: where c hc was the MMAS heat capacity (0.7 J/g·K), ρ d was the MMAS density (1.3 g/cm 3 ), dn/dT was the MMAS thermo-optic coefficient (−10 −5 K −1 ), and ΔE was the energy absorbed by the composite per unit volume per second. The thermal effect of cw low-intensity radiation on the change in the refractive index (red dashed lines in Figure 6b) was relatively small (not more than 20% for blue radiation and 8% for yellow radiation). Generally, the possibility of a nonthermal optical response of the composite due to external optical radiation is associated with the polarization of Fe 3 O 4 nanoparticles in the external field E. Nanoparticle polarization occurs at the spatial separation of positive and negative charges, i.e., at the electron transition to higher allowed energy states (quantum number l ≠ 0). These transitions should be accompanied by the absorption of external radiation. In our case, we observed the absorption of radiation with wavelengths of 380 to 650 nm ( Figure 3). This absorption band consisted of three maxima (380, 480, and 650 nm), indicating the broadened quantum-size states for the electrons in Fe 3 O 4 nanoparticles. Because the bandgap of magnetite is rather small (approximately 0.2 eV) [20][21][22], the conduction and valence bands of the nanoparticles should be coupled due to quantum-size effect [44]. Therefore, the transitions of Fe 3 O 4 nanoparticle electrons to higher energy states by the action of photons with energies of 2.3 eV (λ = 561 nm) and 2.6 eV (λ = 442 nm) can be considered intraband transitions. In turn, these transitions result in changes in the refractive index of the media as follows [45][46][47]: where e was the electron charge, c was the speed of light, ε 0 was the electric constant, m e was the electron mass, and N e was the concentration of excited electrons, which depends on the number of photons in the beam or the radiation intensity I.
Using Equation 4 to approximate the experimentally observed behavior of Δn(I) (Figure 6b, blue dashed lines), we estimated that the concentration of optically excited electrons in Fe 3 O 4 nanoparticles was approximately 10 23 m −3 , being the radiation intensity of less than 0.14 kW/cm 2 .
The amplitude of the nanoparticle polarization is determined by |E| of the external field and the nanoparticle susceptibility (χ) or polarizability (α) measured in cubic angstrom. In turn, the change in the refractive index induced by the radiation is associated with the change in nanoparticle polarizability Δα (Å 3 ) by classical relations [48]. Therefore, we could calculate the values of Δα (Å 3 ) for Fe 3 O 4 nanoparticle using the experimental values of Δn(I) and the following equations (SI): where ε was the real part of the dielectric constant, the composite refractive index n(I) = n 0 + Δn(I), and n 0 was the refractive index of pure MMAS (approximately 1.5).
The obtained values for the changes in nanoparticle polarizability are orders of magnitude greater than those for semiconductor nanoparticles and molecules [30,31] in extremely weak optical fields. In addition, the average nanoparticle volume was approximately 2.2 × 10 6 Å 3 , and the maximum value of Δα (Å 3 ) was 9 × 10 6 Å 3 . Thus, we can conclude that the nanoparticle polarization should be formed by several optical intraband transitions of nanoparticle electrons in weak optical fields.

Conclusions
We used the developed co-precipitation method to synthesize spherical Fe 3 O 4 nanoparticles covered with a monolayer of oleic acid that possessed a wide nonlinear absorption band of visible radiation 1.7 to 3.7 eV. The synthesized nanoparticles were dispersed in the optically transparent copolymer methyl methacrylate with styrene, and their optical properties were studied by optical spectroscopy and z-scan techniques. We report that the electric polarizability of Fe 3 O 4 nanoparticles changes due to the effect of low-intensity visible radiation (I ≤ 0.2 kW/ cm 2 ; λ = 442 and 561 nm) and reaches a relatively high value of 10 7 Å 3 . The change in polarizability is induced by the intraband phototransition of charge carriers and can be controlled by the intensity of the visible radiation used. This optical effect observed in magnetic nanoparticles may be employed to significantly improve the drug uptake properties of Fe 3 O 4 nanoparticles.