Green Synthesis of Co-Zn Spinel Ferrite Nanoparticles: Magnetic and Intrinsic Antimicrobial Properties

Spinel ferrite magnetic nanoparticles have attracted considerable attention because of their high and flexible magnetic properties and biocompatibility. In this work, a set of magnetic nanoparticles of cobalt ferrite doped with zinc was synthesized via the eco-friendly sol-gel auto-combustion method. Obtained particles displayed a room-temperature ferromagnetic behavior with tuned by chemical composition values of saturation magnetization and coercivity. The maximal values of saturation magnetization ~74 Am2/kg were found in cobalt ferrite nanoparticles with a 15–35% molar fraction of cobalt replaced by zinc ions. At the same time, the coercivity exhibited a gradually diminishing trend from ~140 to ~5 mT whereas the concentration of zinc was increased from 0 to 100%. Consequently, nanoparticles produced by the proposed method possess highly adjustable magnetic properties to satisfy the requirement of a wide range of possible applications. Further prepared nanoparticles were tested with bacterial culture to display the influence of chemical composition and magnetic structure on nanoparticles-bacterial cell interaction.


Structural and Magnetic Properties
The X-ray diffraction (XRD) studies were performed with a PW 1830 powder diffractometer (Philips, Eindhoven, The Netherlands) using Co Kα source (λ = 1.78919 Å) in the 2θ geometry in the range of 30-80 degrees. The morphology of the samples was investigated with a JEM-2100 transmission electron microscope (TEM, JEOL, Tokyo, Japan) operating at 30 kV. The magnetic properties were measured with a vibrating sample magnetometer (7400 System, Lake Shore Cryotronics Inc., Weterville, OH, USA) in the field range up to 1.0 T at room temperature (~300 K). The powder samples were fixed with diamagnetic glue in plastic holders to prevent any movement of the powder during the measurements.

Antimicrobial Activity
All experiments were done using E.coli K-12 MG1655 bacterial cells (taken from laboratory collection of IKBFU, Kaliningrad, Russia). First, antimicrobial activity was tested with the disk diffusion method in order to qualitatively define the parameters of the experiment (see Figure S1 and description below). Then quantitative analysis was performed by the optical density (OD 600 ) measurement method on a UV/vis spectrophotometer (SmartSpec Plus, Bio-Rad, Hercules, CA, USA) detecting absorbance at 600 nm wavelength every half hour. Bacterial cells were grown aerobically at 37 • C under constant shaking (~120 rpm) in 5 mL of liquid (without adding of agar) hand-made Lysogeny Broth (LB) medium, containing 1% tryptone, 0.5% yeast extract and 1% sodium chloride. In all experimental samples was added 0.5 mg of nanoparticles. To assess the changes in the optical properties of the LB medium after the addition of nanoparticles, its optical density was detected at a wavelength of 600 nm. It was found that the increase of optical density occurred no more than 10% compared with the control. Obtained OD 600 data values were analyzed using a two-way analysis of variance (ANOVA, GraphPad Prism 7.04 software, Graph Pad Software Inc., San Diego, CA, USA). All OD 600 data discussed in the Biology section below are statistically significant (p-value meanings were ranked by asterisks **** (p ≤ 0.0001), *** (p ≤ 0.001), ** (p ≤ 0.01).

Structural and Morphological Properties
The spinel ferrite structure was confirmed by XRD patterns (Figure 1a) showing for all the samples the presence of spinel cubic structure (card № 00-154-0973 for cobalt ferrite or card № 00-154-0936 for zinc ferrite). Any other phases have been not detected, in contrast to nickel-doped cobalt ferrite MNPs prepared with the same SGAC method where at a high nickel content (x = 0.75 and 1) a small amount of metallic nickel and hematite (α-Fe 2 O 3 ) has been observed [37]. The size of crystallites (D XRD ) was calculated by using the Scherrer formula [43]: where K is a crystallite-shape factor (0.94 for spherical particles), B is the full width at the half maximum estimated after fitting of peaks with the Voigt function and θ is the position of corresponding peaks. The D XRD value is slowly decreased over the concentration of zinc ( Figure 1b, Table 1). A similar decrease was observed in ref. [27] in zinc-cobalt ferrite MNPs in the range of x = 0.08−0.56. It is worth to underline, that in the case of cobalt-nickel ferrite MNPs obtained with the same method, the smaller size of crystallites was in pure cobalt and nickel ferrite but remarkably higher for mixed Ni-Co ferrite nanoparticles [37,44]. Similarly, the position of main reflections is continuously increasing with the increase of zinc content despite the similar ionic radii of Zn 2+ (~0.82 Å [45]) and Co 2+ (~0.82 Å [45]) ions: this can be ascribed to the migration of the smaller Fe 3+ (~0.67 Å [45]) ions to octahedral sites since zinc ions prefer to occupy the tetrahedral position [25]. The fact that bigger Zn 2+ occupies the smaller tetrahedral sites is confirmed by the expansion of the lattice parameter a (Figure 1c). The lattice parameter of cubic structure (a) was calculated by the equation a = d hkl h 2 + k 2 + l 2 , where d hkl is the interplanar spacing of a plane with Miller indices h, k and l. The linear dependence of a versus x agrees with Vegard's law predicting that the lattice parameter of a chemically homogenous mixture will be the approximately weighted mean of two constituents [46].
Materials 2020, 13, x 4 of 13 154-0936 for zinc ferrite). Any other phases have been not detected, in contrast to nickel-doped cobalt ferrite MNPs prepared with the same SGAC method where at a high nickel content (x = 0.75 and 1) a small amount of metallic nickel and hematite (α-Fe2O3) has been observed [37]. The size of crystallites (DXRD) was calculated by using the Scherrer formula [43]: where K is a crystallite-shape factor (0.94 for spherical particles), B is the full width at the half maximum estimated after fitting of peaks with the Voigt function and θ is the position of corresponding peaks. The DXRD value is slowly decreased over the concentration of zinc ( Figure 1b, Table 1). A similar decrease was observed in ref. [27] in zinc-cobalt ferrite MNPs in the range of x = 0.08−0.56. It is worth to underline, that in the case of cobalt-nickel ferrite MNPs obtained with the same method, the smaller size of crystallites was in pure cobalt and nickel ferrite but remarkably higher for mixed Ni-Co ferrite nanoparticles [37,44]. Similarly, the position of main reflections is continuously increasing with the increase of zinc content despite the similar ionic radii of Zn 2+ (~0.82 Å [45]) and Co 2+ (~0.82 Å [45]) ions: this can be ascribed to the migration of the smaller Fe 3+ (~0.67 Å [45]) ions to octahedral sites since zinc ions prefer to occupy the tetrahedral position [25]. The fact that bigger Zn 2+ occupies the smaller tetrahedral sites is confirmed by the expansion of the lattice parameter a (Figure 1c). The lattice parameter of cubic structure (a) was calculated by the equation a d √h k l , where dhkl is the interplanar spacing of a plane with Miller indices h, k and l.
The linear dependence of a versus x agrees with Vegard's law predicting that the lattice parameter of a chemically homogenous mixture will be the approximately weighted mean of two constituents [46].    (4) * In parenthesis the systematical error in the last digit is presented; ** M S values are after approximation with LAS, the error is fittings error.
The TEM-images indicate the presence of non-regular shape of particles with relatively broad distribution (i.e., in the range 10-30 nm) and shape edges confirming the crystalline nature of particles. Examples of typical TEM-images for CoFe 2 O 4 are presented in the SI ( Figure S2).

Chemical Control. of Magnetic Properties
Preliminary magnetic characterization has been performed for all the samples at 300 K. All investigated samples exhibited hysteretic behavior (Figure 2a) with the values of saturation magnetization (M S ), coercivity field (H C ) and reduced remanent magnetization (M R /M S ) reported in Table 1. Since almost all M-H curves were not saturated, the Law of Approach to Saturation (LAS) was used to estimate M S values [47,48]: where a and b are free parameters of the fitting. M S first increases with the increase of Zn-content, up to x = 0.25. After this value, M S starts to decrease in good agreement with literature data and may be explained by the cation distribution [5,[25][26][27]. The maximal value of M S of~74 Am 2 /kg was for x = 0.25, which is higher than reported for samples prepared with a similar method, probably because we used a higher reaction temperature, leading to more crystalline particles. For Ni-doped ferrite, we observed maximal value M S of~69 Am 2 /kg for Ni 0.25 Co 0.75 Fe 2 O 4 [37]. The coercivity field (H C ) and reduced remanence (M R /M S ) are decreasing with the increase of zinc concentration ( Figure 2b). M S values calculated with Equation (3) were converted into the magnetic moment per formula unit in Bohr magnetons to better understand processes associated with filling sublattices in the structure of ferrites. The following equation was used to do that converting [27,45]: where M w is the molecular weight, N A is Avogadro constant (6.022 × 10 23 mol −1 ) and µ B is Bohr magneton (9.274 × 10 −24 J/T). Figure 3a shows the values of M S in µ B units extracted from M-H curves recorded at 300 K. In the region of 0 < x < 0.25 increase of magnetization is due to a partial migration of Fe 3+ cations to B-sites. In fact, for low zinc concentration, Zn 2+ push Fe 3+ from A to B sites ( Figure 3b). If exchange interactions between two different sites (exchange integral J AB ) are stronger than intra-lattice interactions (J AA and J BB ), the net magnetization may be expressed as (3 + 7x) × µ B , because one uncompensated magnetic moment of Co 2+ (3µ B ) is replaced by two uncompensated Fe 3+ (5 µ B ) in the octahedral position [4,49,50]. With an increase of concentration of zinc, A-B interaction is weakening. In an extremal case x = 1, J AB and J AA are equal to 0 because the tetrahedral positions are filled with non-magnetic Zn 2+ ions and B-B interaction leads to the establishment of antiferromagnetic ordering among magnetic moment in the octahedrally coordinated lattice. On basis of this simple model, a general rule of zinc dependence of magnetization can be ascribed as m where δ is delta equal to 1 if J AB >> J AA + J BB and equal to −1 if J AB << J AA + J BB , µ I is the initial magnetic moment of undoped ferrite and µ' is a concentration-depended magnetic moment resulting due to recombination of cations between lattices and disruption of ideal ferrimagnetic ordering.
where a and b are free parameters of the fitting. MS first increases with the increase of Zn-content, up to x = 0.25. After this value, MS starts to decrease in good agreement with literature data and may be explained by the cation distribution [5,[25][26][27]. The maximal value of MS of ~74 Am 2 /kg was for x = 0.25, which is higher than reported for samples prepared with a similar method, probably because we used a higher reaction temperature, leading to more crystalline particles. For Ni-doped ferrite, we observed maximal value MS of ~69 Am 2 /kg for Ni0.25Co0.75Fe2O4 [37]. The coercivity field (HC) and reduced remanence (MR/MS) are decreasing with the increase of zinc concentration (Figure 2b). The linear fitting of region 0 < x < 0.25 gives the value (2.8 + 0.9x) × μB. The value of magnetization per formula unit in the extremal case x = 0, pure cobalt ferrite, is very close to the theoretical value of 3 μB (only spin contribution to the magnetic moment it is taken into account) for a totally inverted spinel (γ = 1). In our case coefficient before x is much lower than estimated which indicates the Zn 2+ occupy only partially tetrahedral sites while some amount of Zn 2+ can occupy octahedral positions. This value is also affected due to antiferromagnetic order between magnetic cations in B sites since the B-B super-exchange interactions become more significant because of the breaking of the dominant A-B interactions. Spin canting arises when in B-sites concentration of non-magnetic ions becomes too high and can be explained in the frame of the Yafet-Kittel model [27,32,34]. Gómez-Polo and colleagues observed a divergence of the calculated value of magnetic moment by measured value of inversion degree by neutron diffraction which was enhanced when the concentration of zinc increased [34]. In our case trend becomes more pronounced in the region of 0.25 < x < 1 because of the change of the slope in the linear fitting of this region (4 − 3.6 x) × μB. This non-monotonical trend of magnetization suggests that the filling of the spinel positions is also non-monotonical. The observed trend well agrees with literature data (adopted values from the literature of room The linear fitting of region 0 < x < 0.25 gives the value (2.8 + 0.9x) × µ B . The value of magnetization per formula unit in the extremal case x = 0, pure cobalt ferrite, is very close to the theoretical value of 3 µ B (only spin contribution to the magnetic moment it is taken into account) for a totally inverted spinel (γ = 1). In our case coefficient before x is much lower than estimated which indicates the Zn 2+ occupy only partially tetrahedral sites while some amount of Zn 2+ can occupy octahedral positions. This value is also affected due to antiferromagnetic order between magnetic cations in B sites since the B-B super-exchange interactions become more significant because of the breaking of the dominant A-B interactions. Spin canting arises when in B-sites concentration of non-magnetic ions becomes too high and can be explained in the frame of the Yafet-Kittel model [27,32,34]. Gómez-Polo and colleagues observed a divergence of the calculated value of magnetic moment by measured value of inversion degree by neutron diffraction which was enhanced when the concentration of zinc increased [34]. In our case trend becomes more pronounced in the region of 0.25 < x < 1 because of the change of the slope in the linear fitting of this region (4 − 3.6 x) × µ B . This non-monotonical trend of magnetization suggests that the filling of the spinel positions is also non-monotonical. The observed trend well agrees with literature data (adopted values from the literature of room temperature M S are presented in Figure 3a). Mameli et al. observed maximal value M S for~7-nm particles prepared with thermal decomposition method for concentration of zinc of 0.46 at 5K, however at room temperature, the maxim of M S -x dependence shifted to the lower x values [25]. Ben Tahar et al. observed a maximum for x = 0.4 iñ 5-nm particles prepared with the polyol method [32]. Andersen et al. found the maximum value of M S at x = 0.2 for~14-nm particles prepared with the hydrothermal method [51]. A small difference in the absolute value of magnetic moment and position of maximum related to the fact, that particles were prepared with different methods and have different sizes and cation distribution.
The surface of spinel ferrites is preferentially formed by ions located in the octahedral sites. (Figure 3c). Thus, the chemical, catalytical and biological activity of mixed ferrites will mainly depend on structural properties (level of inversion degree) rather than on their chemical composition [38,39]. While the importance of the magnetic microstructure on the catalytical properties of nanoferrites is well known, their influence on biological activity is still relatively not well understood.
The introduction of zinc in ferrite structure significantly decreases Curie (or Néel) temperature of ferrites MNPs which leads to the decrease of magnetization at room temperature. Previously it was demonstrated that the Curie temperature of cobalt ferrite MNPs was decreasing from 713 to 453 K when the concentration of zinc changed from x = 0 to 0.5 [52]. That makes this material potentially useful in, for example, the self-controlled hyperthermia [26,53] or any other applications where MNPs with high saturation magnetization and tunable Curie temperature are requested [54,55].
In contrast to this non-monotonically behavior of M S , the H C trends to decreasing continuously as presented in Figure 2b as expected for the reduction of Co 2+ . It is interesting to underline the difference in comparison with a similar system of cobalt ferrite MNPs doped with nickel [37]. In the case of zinc, the H C decreases much faster showing the downward trend, while for nickel-cobalt ferrite MNPs, the upward trend was observed.
To better understand the role of magnetic anisotropy the constant of effective magnetic anisotropy (K eff ) was calculated following [56]: where µ 0 H A is anisotropy field can be estimated by the closure field of hysteresis (H irr [58,59]. Equation (4) is valid for non-interacting MNPs with uniaxial anisotropy. Despite, the magnetocrystalline anisotropy of cobalt ferrite is cubic type, in the case of nanoparticle system the effective anisotropy frequently turns on uniaxial type because of the contribution of such factors of shape, surface, and interparticle interactions [47]. K eff calculated using equation 4 was maximal for CoFe 2 O 4 14 × 10 4 J/m 3 (the bulk 30 × 10 4 J/m 3 [60]), and fast drop until reaching a value in three-orders of magnitude lower for ZnFe 2 O 4 . The obtained M R /M S and K eff values are lower than expected probably due to the influence of temperature and interparticle interactions leading to the magnetization relaxation for particles with the lower anisotropy. Also, it is interesting to note a divergence in the behavior of coercivity and closure field, the H irr /H C increased in two times for Zn-doped cobalt ferrites comparing with the pure cobalt ferrite MNPs. That confirms the crucial role of magnetocrystalline (core) anisotropy in cobalt ferrite samples and arising a stronger surface anisotropy in magnetically softer mixed Zn/Co ferrites MNPs.

Biology
Since the nanoparticles are iron-based, their presence should affect the bacteria iron metabolism system. Fe-containing MNPs may affect bacterial cells by reactive oxygen species production and induction of membrane disruption [61,62]. Thus, the data on the different effects of nanoparticles on Materials 2020, 13, 5014 8 of 13 wild-type cells and bacteria transformed with the plasmid can indirectly indicate the possibility of their penetration into the cell. The effect of iron ions on cell growth is a complicated parameter since iron is participating in many important biochemical processes, even if its overdose can cause a strong cytotoxic effect. To protect genomes from damage, E.coli uses the DNA-binding protein of starved cells (Dps), which combines ferroxidase activity and the ability to bind DNA. Both activities depend on the integrity of this multi-subunit protein, which has an inner cavity for iron oxides. It was shown by X-ray absorption near edge structure (XANES) and Mössbauer spectroscopy presence of both trivalent and divalent iron ions in the Dps protein [40,41]. Thus, if iron-based MNPs penetrate the cell, then they should affect the metabolic processes involving Dps.
Measured values of the OD 600 signal are proportional to the cell's population after the incubation time [63,64]. In the presence of CoFe 2 O 4 and ZnFe 2 O 4 MNPs, E.coli demonstrate a decreased OD 600 of the bacterial medium compared with the control (non-treated) samples, resulting in the lag of bacterial growth inhibition in different timepoints (Figure 4a). The values of OD 600 at 330 min, the time selected as moments when bacterial strains reached sub-maximum population, demonstrate that pure CoFe 2 O 4 and ZnFe 2 O 4 MNPs have a more pronounced lag effect than mixed Zn-Co ferrite MNPs with the concentration of zinc x = 0.25 (Figure 4b). In Figure 4, results for selected samples only are presented, more data confirming the same trend are in Figure S3. Cells growth of E.coli K-12 MG1655 was inhibited after MNPs (x = 1) from 240 to 330 min (p-value ≤ 0.001) and MNPs (x = 0) from 120 to 360 min (p-value ≤ 0.01). Therefore, it can be concluded that in the case of different Zn concentrations, a slowdown in the increase in OD is observed in the period from 200 to 400 min. Results cannot be attributed to the inhibition of bacterial growth that occurs during toxic exposure, since the achievement of the stationary growth phase occurs simultaneously for control cells and cells grown in the presence of nanoparticles. This bimodality can be explained by the triggering of special compensatory or protective mechanisms in bacterial cells. Such mechanisms can be implemented by post-translational modification [65], special chaperone proteins [66], or the triggering of special proteins of stresses [67]. in the presence of nanoparticles. This bimodality can be explained by the triggering of special compensatory or protective mechanisms in bacterial cells. Such mechanisms can be implemented by post-translational modification [65], special chaperone proteins [66], or the triggering of special proteins of stresses [67].  the non-monotonical evolution of the magnetic structure. Namely, at Zn concentration of around x = 0.25, Fe 3+ ions partially migrate to octahedral positions lowering the toxic effect of Co 2+ ions at the same positions. However, at high zinc concentration, Zn 2+ ions start also occupy octahedral positions due to the non-thermodynamically equilibrium state of nanoparticles involving an increase of toxicity.

Conclusions
Zn-Co ferrite MNPs synthesized by the green synthesis method revealed tunable magnetic properties, which can be adjusted via their chemical composition. The increase of zinc content reduces magnetocrystalline anisotropy and has a non-monotonical effect on the saturation magnetization value with a maximal of~74 Am 2 /kg for x = 0.25 zinc-concentration. This can be considered in a strong correlation with change of the magnetic structure of spinel structure: at small Zn-concentration (<25%), Zn 2+ ions can occupy the tetrahedral sites pushing the Fe 3+ with a high magnetic moment into the octahedral site. At a higher concentration of zinc, the ferrimagnetic order between two lattices is disturbed that leads to dropping in saturation magnetization and magnetic anisotropy. Here we should stress that magnetic properties correlate with microstructural and magnetic structural properties (first of all, cation distribution) which in their turn are in strong dependence on the synthesis method (in general, with the kinetic of reaction). Zn-Co ferrite MNPs induced the lag in E.coli growth, which was found in correlation with the magnetic and structural properties of spinel ferrites structure.

Conflicts of Interest:
The authors declare no conflict of interest.