Fabrication, characterization, and antibacterial activity of ferrite, chromite, and aluminate nanoparticles

Zn0.33Co0.33Mg0.33X2O4 nanoparticles (NPs), where X = Fe, Cr, Al and denoted by F, C, and A, were prepared by the co-precipitation method. X-ray diffraction patterns validated the formation of NPs with cubic spinel structure with the detection of small amounts of impurities in samples C (Cr2O3) and A (MgO). Transmission electron micrographs showed a nearly spherical shape for samples F and A. However, sample C revealed cubic and nearly spherical shapes. Energy-dispersive x-ray analysis ensured the presence of chemical constituents in all samples. The vibrational modes of NPs were confirmed with Fourier transform infrared spectroscopy. The direct bandgap energy values, calculated using ultraviolet-visible spectroscopy, were in the range of 2.355 and 2.967 eV for F, C, and A samples. X-ray photoelectron spectroscopy analysis confirmed the compositions as well as the valence states of all elements. Magnetic hysteresis (M–H) loops revealed a soft ferromagnetic behavior. Sample F exhibited a higher saturation magnetization, remanent magnetization, magnetic moment, and magnetic anisotropy compared to those of samples C and A. The antibacterial activity of the tested samples against four bacteria (Staphylococcus aureus, Stenotrophomonas maltophilia, Escherichia coli, and Enterococcus faecium) was determined using the broth microdilution assay, minimum bactericidal concentration (MBC), and time-kill test. The prepared NPs exhibited varying antibacterial activity due to multiple factors. These results highlighted the potential utility of the ternary ferrite, chromite, and aluminate NPs in the treatment of microbial infections, particularly multidrug-resistant bacteria.


Introduction
Bacterial infections are one of the main reasons for chronic contagion and death [1].Yet, several studies have shown that the extensive use of antibiotics has contributed to the creation of multidrug-resistant bacteria [1].Nowadays, research interests have been focused on cutting-edge NPs that exhibit antibacterial effects [1].Due to their unique properties, spinel oxide NPs are among the most commonly used and synthesized nanomaterials [2].
antibacterial activity was evaluated against two Gram-positive (S. aureus and E. faecium) and two Gram-negative (S. maltophilia and E. coli) bacteria using broth microdilution assay, minimum bactericidal concentration (MBC), and time-kill test.

Materials and methods
2.1.Chemicals Zinc chloride (ZnCl 2 , 98.00%), Magnesium chloride hexahydrate (MgCl 2 .6H 2 O, 99.00%), Iron (III) chloride hexahydrate (FeCl 3 .6H 2 O, 99.00%), and chromium (III) chloride hexahydrate (CrCl 3 .6H 2 O, 98.00%) were supplied from Sigma-Aldrich, Germany.Cobalt (II) chloride hexahydrate (CoCl 2 .6H 2 O, 98.00%) was purchased from Sigma-Aldrich, UK.Aluminum chloride (AlCl 3 , 98.50%) was purchased from Thermo Scientific, Germany.Sodium hydroxide pellets (NaOH, 99.44%) were acquired from Fisher Scientific, UK.Sodium chloride BP (NaCl, 7647-14-5), nutrient broth (CM0001, OXOID), and agar (05039, Fluka) were acquired from Scientific & Technical Supplies company.O, respectively.The stoichiometric amounts needed for each precursor were dissolved in a precise volume of deionized water to achieve 1 M solutions.The salt chlorides were then mixed and the precipitant NaOH (2 M) was added drop by drop to the mixture under constant stirring until pH reached 13.The solution was heated to 80 °C with magnetic stirring at a fixed rate for 2 h and then left to cool down at room temperature.The solution was then frequently washed with deionized water until the pH value was reduced to 7. Subsequently, the precipitates were dried at 90 °C for 16 h, ground, and annealed at 750 °C for 4 h.The synthesis process of Zn 0.33 Co 0.33 Mg 0.33 Cr 2 O 4 NPs, denoted by C, was performed by the co-precipitation method starting with stoichiometric quantities of 1.253 g, 2.187 g, 1.868 g, and 14.692 g for ZnCl 2 , CoCl 2 .6H 2 O, MgCl 2 .6H 2 O, and CrCl 3 .6H 2 O, respectively.According to stoichiometric calculations, 1 M of each precursor was separately dissolved in deionized water and then mixed.The obtained mixture was heated at 65 °C accompanied by magnetic stirring for 2 h.To obtain a basic solution, 1 M NaOH solution was added drop by drop to the solution under continuous stirring, maintaining its temperature at 65 °C until pH was achieved 10.Then, when the temperature of the solution was reduced to room temperature, it was repeatedly washed with deionized water to reach a neutral pH value (PH = 7).Finally, the precipitates were dried at 120 °C for 16 h, ground, and annealed under the same conditions for the ferrite NPs.The co-precipitation method was also adopted to synthesize Zn 0.33 Co 0.33 Mg 0.33 Al 2 O 4 NPs, denoted by A, starting with stoichiometric amounts of 1.979 g, 3.454 g, 2.952 g, and 11.615 g for ZnCl 2 , CoCl 2 .6H 2 O, MgCl 2 .6H 2 O, and AlCl 3 , respectively.The stoichiometric amounts of each starting material were dissolved in a precise volume of deionized water to get 1 M precursor solutions.The homogeneous solutions were mixed and heated at 40 °C with constant stirring for 30 min.Then, a solution of NaOH with a molarity of 3 M was added drop by drop to the solution until pH reached 12, keeping the temperature at 40 °C.The temperature was raised to achieve 60 °C and maintained for 2 h under constant stirring.After cooling down the solution at room temperature, it was frequently washed with deionized water until the pH reached 7. Later, the precipitates were dried at 90 °C for 18 h and ground.The last step was calcining the dried powders with identical calcination conditions of the first preparation.

Characterization techniques
The x-ray powder diffraction (XRD) patterns of the prepared NPs were recorded using a Bruker D8 advance equipped with a radiation source of Cu-K α (λ = 1.54060Å), with 2θ ranging from 10°to 80°.The surface morphology and particle size distribution of the studied samples were investigated via the transmission electron microscope (TEM, Jeol JEM-100CX microscope, 80 kV accelerating voltage).High-resolution TEM (HRTEM) and selected area electron diffraction (SAED), JEM-2100 plus electron microscope with 200 kV accelerating voltage, were used to determine the interplanar distance.Energy-dispersive x-ray (EDX) measurements of the prepared samples were determined by EDAX-ZAF, with an SDD Apollo X detector, applied at a voltage of 20 kV.Fourier transform infrared spectroscopy (FTIR) studies of the prepared samples were performed using FTIR-8400 S combined with a Nicolet iS5 spectrometer in the 4000-400 cm −1 spectral range.The FTIR measurements were carried out by mixing and grinding 2 mg of each sample with 200 mg of potassium bromide (KBr) medium and pressed at 13,790 kPa to obtain a disk shape.The optical properties of the prepared samples were examined using an ultraviolet-visible (UV-vis) spectrophotometer (Jasco V-670), at room temperature, in the range of 300 to 600 nm.The elemental composition and its oxidation states were discovered by x-ray photoelectron spectroscopy (XPS) measurements using K-Alpha (ThermoFisher Scientific) XPS with the following conditions: monochromatic Alk α radiation, pressure of 10 −9 mbar, pass energy of 200 eV and narrow spectrum of 50 eV.The M-H loops of the investigated samples were measured using a Lakeshore 7410 vibrating sample magnetometer (VSM), at room temperature, with magnetic fields ranging between -20 and +20 kG.

Antibacterial assay 2.4.1. Bacterial species
The tested bacterial species in this study were Gram-positive (Staphylococcus aureus and Enterococcus faecium) and Gram-negative (Stenotrophomonas maltophilia and Escherichia coli) bacteria.The tested bacteria were isolated from the waste water obtained from South Lebanon Water Establishment (SLWE) station (Lebanon, Sidon).The antibacterial activity of the prepared NPs was evaluated through broth microdilution assay, minimum bactericidal concentration (MBC), and time-kill test.All experiments were performed in triplicates.

Evaluation of minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC)
The MIC assays were performed against the four bacterial species employing the broth microdilution method.Two-fold serial dilution of the tested materials was performed in nutrient broth to obtain concentrations ranging from 20 to 0.02 mg ml −1 .100 μl of the diluted NPs was dispensed to the wells of the 96-well plates.Then, 100 μl of bacterial inoculum (5 × 10 5 CFU ml −1 ) was transferred to each well.100 μl of broth medium and 100 μl of the diluted sample served as blank.Two controls were considered in this method, the growth control (GC) contains 100 μl nutrient broth and 100 μl diluted bacterial suspensions, and the negative control (NC) includes 200 μl nutrient broth only.The plates were incubated at 37 °C for 24 h and the optical density (OD) was assessed using an ELISA spectrophotometer at a wavelength of 595 nm.The percentage of inhibition was calculated according to this equation [37]: The mixture was represented by 100 μl diluted NPs and 100 μl bacterial inoculum.The lowest concentration that causes inhibition of bacterial growth was regarded as MIC.
For MBC, the nutrient agar plates were streaked with 10 μl of the mixture, the GC, and the NC followed by incubation at 37 °C for 24 h.The minimum concentration needed to kill bacteria was considered as MBC [38].

Time-kill test
The time required by the NPs to exert bacteriostatic or bactericidal activity was examined by the time-kill test assay.100 μl of MIC and MBC of the tested material and 100 μl of bacterial cultures (5 × 10 5 CFU ml −1 ) were pipetted in 96-well microplates.The GC wells were loaded with 100 μl inoculums and 100 μl broth medium, while the NC wells were filled with 200 μl broth medium.100 μl NPs added to 100 μl broth solution were taken as blank.The plates were incubated at 37 °C and analyzed using an ELISA spectrophotometer to measure the OD at 595 nm in defined time intervals t 0 , t 1 , t 2 , t 4 , and t 24 corresponding to 0, 1, 2, 4, and 24 h [39].

Statistical analysis
All data were expressed as the mean ± standard error of the mean (SEM) of independent experiments performed in triplicate.Microsoft Excel software (2016) was employed to analyze data obtained from the study by using one-way analysis of variance (ANOVA) followed by Tukey's post hoc tests, assuming significance levels of p 0.05.Particle size (D TEM ) estimation of the prepared NPs was carried out using ImageJ 1.53e software (National Institute of Health, USA).Material Analysis using Diffraction MAUD software was used for Rietveld refinement and verification of all phases presented in the samples.The de-convolution of the graphs was done with the help of Fityk 1.3.1.Statistical charts and graphs were plotted using Origin2018 (OriginLab Corporation, USA).
The lattice parameter (a) and the crystallite size (D XRD ) values were calculated, for all prepared samples, and are listed in table 1.Note that the average values of a and D XRD were calculated by taking the average values of all the peaks corresponding to the above-mentioned Miller indices of the main phase for each sample.Bragg's law equation was used for the estimation of a [42]: where d hkl is the interplanar distance and h, k, l refer to Miller indices.The average values of the calculated lattice parameter from XRD (a calculated ) by employing equation (2) for F, C, and A samples were 8.397 ± 0.0004, 8.325 ± 0.0009, and 8.076 ± 0.0003 Å, respectively.The obtained values are consistent with those stated in the literature [9,46,47].The a calculated values match well with those obtained from the Rietveld analysis (a MAUD ).
The maximum value of a is obtained in sample F, whereas the lowest value of a is obtained in sample A. This variation in a values can be related to the ionic radius of Fe 3+ (0.645 Å), Cr 3+ (0.615 Å), and Al 3+ (0.535 Å).Pani et al [13] identified a comparable trend of variation.It is obvious that samples C and A exhibit a lower lattice parameter than that of sample F. According to Bitar et al [48], this variation in the lattice parameter can be attributed to the formation of a secondary phase.Subsequently, the formation of Cr 2 O 3 in sample C and MgO in sample A may cause the lattices to contract.Moreover, Debye-Scherrer's formula was employed to calculate D XRD values [47]: where k is a dimensionless (shape) factor equal to 0.9, λ is the wavelength of the Cu-K α radiation (1.5406 Å), β hkl is the full-width at half maximum (FWHM) of the XRD peaks (in rad), and 2θ is the diffraction angle (in rad).The mean D XRD values are 27.155 ± 0.0007, 39.666 ± 0.0008, and 17.850 ± 0.0008 nm for F, C, and A samples, respectively.The D XRD values in the present study confirm well with those obtained in literatures [42,47,49].It is clear that the maximum D XRD value is attained in sample C, while the minimum value is shown in sample A.
The smallest size of sample A may depend on the lowest heating growth temperature applied throughout the synthesis procedure [50].This can be demonstrated by the XRD peaks of sample A that are broader and less sharp than those of samples F and C. A similar phenomenon was obtained by Sripriya et al [41].Moreover, this size reduction reflects a decrement in crystallinity [51].Franco et al [52] reported that the small-sized NPs improved the antibacterial effect.In addition, Bitar et al [48] suggested that the presence of a secondary phase caused an enhancement in the inhibitory effect.

TEM, HRTEM, and SAED analysis
TEM was employed to investigate the morphology and particle size distribution of the prepared samples, as represented in figures 2(a)-(c).The mean particle size histograms were estimated from the TEM images performed by ImageJ software at 100 nm scale, and their corresponding particle size (D TEM ) values are listed in table 1. Figure 2(a) shows that sample F exhibits a nearly spherical shape with a small degree of agglomeration.Kanagesan et al [53] identified the same shape for ZnFe 2 O 4 NPs.In addition, figure 2(c) reveals that sample A is nearly spherical with non-agglomeration.Likewise, Jamal et al [54] noticed that ZnAl 2 O 4 NPs were spherical in shape.However, the TEM micrograph of sample C appears with a non-homogeneous morphology, with a low degree of agglomeration, indicating the coexistence of both phases: cubic equiaxial NPs of different sizes and small faceted nearly spherical NPs (figure 2(b)).Jafarnejad et al [55] also obtained a similar duplex morphology for MgCr 2 O 4 NPs.Dumitru et al [56] noticed two shapes of ZnCr 2 O 4 NPs calcined at 700 °C, while those calcined at 450 °C showed a uniform morphology.They studied the morphology of ZnCr 2 O 4 NPs thermally treated at different temperatures, and it could be observed that as the calcination temperatures increased, more variety in shapes emerged [56].This demonstrated that the morphology of chromite NPs could be affected by the temperature of calcination.The cubic morphology can be attributed to the directed self-assembly growth of NPs which improves the antibacterial property and is then considered as the preferable shape for applications in hyperthermia treatment [57][58][59].Moreover, it is obvious that the magnetic F and C samples tend to agglomerate due to the magnetic attractions between NPs [60,61].However, the particles in sample A are well distributed and not agglomerated, which can be related to the presence of MgO secondary phase, as detected by XRD analysis.
According to Du et al [45], MgO acted as an isolation medium and then inhibited the agglomeration of NPs   The minor variation in particle size and shape, as shown in figure 2, indicates that the particles are nearly monodispersed [67,68].Dheyab et al [67] stated that the monodisperse particles had a narrow FWHM, whereas polydisperse particles possessed a large FWHM.According to figure 1, the peaks possess a nearly narrow FWHM, indicating that the particles are almost monodispersed.The ability to synthesis monodisperse particles can be related to the selection of the synthesis method [69].For instance, Yang et al [69] developed a straightforward approach, which is the co-precipitation method, to produce monodispersed iron oxide (Fe 3 O 4 ) NPs without the use of surfactants, organic salts, or organic solvents.Moreover, the appropriate selection of the synthesis process conditions is among the factors that influence particles monodispersing [70].These conditions comprise transition metal ions, reaction temperature, and reaction time [70].Potbhare et al [68] reported other synthesis conditions that contributed to the formation of NPs with nearly monodisperse nature involving constant magnetic stirring and rpm of centrifugation.
The morphology and crystalline nature of the prepared samples were further investigated by HRTEM and SAED (figure 3).The HRTEM micrographs clearly illustrate the grains of NPs with an ordered pattern of distinct lattice fringes.Using image J software, the microstructure was performed to determine the d hkl obtained from the HRTEM micrographs.The distances between the layers of atomic planes of 2.11 and 1.62 Å are in good agreement with the diffraction peaks indexed to the (400) and (511) planes of sample F (figure 3

EDX analysis
The elemental composition of the studied samples was conducted via EDX, represented in figure 4    divalent metals in sample C, which resulted in a smaller amount of Zn, Co, and Mg than those in samples F and A [74].Meanwhile, for sample A, it can be noticed that in addition to the prominent peaks denoted to Zn, Co, Mg, Al, and O elements represented in the EDX curve, as depicted in figure 4(c), there is a smaller extent of chlorine (Cl).As mentioned above, the chemicals used during the synthesis contain chloride ions; therefore, the precipitate was washed with deionized water to remove Cl.Consequently, the detection of an additional signal related to Cl reveals that this element was not completely removed, as reported by Tariq et al [75].Moreover, it is obvious that the amount of Mg in sample A is higher than that of Zn and Co, confirming the formation of MgO secondary phase.This can be related to the evaporation of other metal ions in the compound [74].This evaporation arises from divalent or trivalent metals, although it most probably originates from Al, resulting in a significant decrease in its content more than that of Fe and Cr in samples F and C, respectively.For all samples, the registered element stoichiometry was found to be close to the theoretically predicted amounts (XRD results).
The EDX studies confirm the formation of the prepared NPs.Similar findings were stated in literatures [20,47,76].

FTIR spectra analysis
The FTIR analysis of the prepared samples was performed at room temperature in the spectral range of 4000-400 cm −1 with an enlarged view within 700-400 cm  [80,82,83].In all FTIR spectra, the band in the range of 2361.40-2365.57cm −1 is assigned to the presence of carbon dioxide molecules (O=C=O) absorbed from the atmosphere [84].Moreover, the peaks at 2849.90 cm −1 in sample A and 2928 cm −1 in samples F and A are ascribed to the asymmetric and symmetric C-H stretching vibration, respectively [47].The O-H stretching vibration of H 2 O molecules is also detected in sample F at 2978.51 cm −1 [85].Also, in all samples, traces of water detected in the range of 1635.74-1637.75cm −1 , and 3420.61-3444.24cm −1 can be attributed to the stretching vibration of O-H [86,87].The hydroxyl groups originate from the adsorbed H 2 O, indicating that the humidity is evident in the tested samples [80].Further, the

UV-vis spectroscopy
The UV-vis absorbance spectra were recorded at room temperature from 300 to 600 nm to determine the optical energy gap for F, C, and A samples, as illustrated in figure 6(a).The absorbance spectra show an intense peak for sample F centered at 336 nm, while a weak absorption peaks are observed at 339 and 370 nm for samples C and A, respectively.This weakness in the absorption edge can be attributed to the formation of the secondary phase, as reported by Munawar et al [92].Moreover, the intensity of the absorbance peak is highest for sample A and lowest for sample F. The higher absorbance peak intensity indicates the more free excitons are created, which is beneficial for bacterial growth inhibition [93,94].Furthermore, the energy gap (E g ) of the tested samples can be calculated using Tauc's equation [93]: where α = 2.303(A/t) is the absorption coefficient (A is the measured absorbance and t is the path length), B is a material-dependent constant, hυ is the incident photon energy (h is Planck's constant and υ is the frequency of the photon), and n is a constant related to a mode of transition, with values of 2 for direct transition and ½ for indirect transition.By plotting (αhυ) 2 versus hυ, the direct E g values were determined and listed in table 3.This was found by extrapolating the linear portion of the plots to hυ as shown in figure 6(b), noting that the intersection represents E g .Samples F, C, and A exhibit direct E g values of 2.967, 2.429, and 2.355 eV, respectively.The obtained values confirm with those reported in literatures [56,93,95].Several factors can affect E g , for instance, crystalline size, presence of impurities, and structural considerations [96].It is noticeable that samples C and A exhibit lower E g values as compared to sample F. The presence of secondary phases in samples C and A, as revealed by XRD examination, may contribute to reduce their E g .Similar findings were obtained by Aridi et al [51].As a result, the formation of secondary phase improved the NP's inhibitory effect against bacteria [48].Furthermore, the smallest D XRD of sample A may also contribute to reduce its E g more than that of samples F and C, as obtained by Akhiruddin et al [97].It was stated by Akbari et al [94] that the narrow E g improved the antibacterial potential of the sample.Moreover, the presence of disorder in the NP's structure is indicated by the appearance of a tail, known as the Urbach tail, in the absorption spectrum [98].Consequently, the Urbach energy is given according to the following relation [98]: where a 0 is constant and E u is the Urbach energy used to quantify disorder (or defect) in the band structure.Figure 6(c) shows the variation of lnα as a function of hυ.The E u values for F, C, and A samples are obtained from the reciprocal of the slope of the linear part at the tail region which are 0.674, 0.816, and 0.856 eV, respectively (table 3).These values are comparable to those reported in literatures [48,51].It is obvious that the lowest E u value is achieved in sample F, while the highest E u value is obtained in sample A. This reveals that the structure of sample F has the least disorder, while the structure of sample A has the most disorder.The higher E u of samples C and A compared to that of sample F can be caused by the formation of secondary phase, which produces disorders in the bandgap.Also, it is crucial to point out that the inverse relation between E g and E u suggests that the disorders induced in the structure of the presented samples have increased [51].

XPS study
The identification of the oxidation states of the investigated NPs was achieved with the use of XPS measurements.Figure 7 displays the survey spectra for F, C, and A samples.The XPS spectra of the prepared samples demonstrate the presence of Zn, Co, Mg, O, and C. In addition, peaks of Fe, Cr, and Al are detected in the measured spectra for F, C, and A samples, respectively.Moreover, two peaks of Cl (Cl-2 s and Cl-2p) are detected near C-1 s in the spectrum of sample A, coming from the starting chemicals that do not completely interact, which coincides with the EDX results.Additionally, the appearance of Cl can also originate from the thermal decomposition of AlCl 3 in which Al cation (Al 3+ ) and Cl anion (Cl -) are formed after calcination at 750 2 ) [99].This can be explained according to the study of Iaiche and Djelloul [99], which studied zinc oxide/zinc aluminate (ZnO/ZnAl 2 O 4 ) nanocomposite by using aluminum sulfate hydrate (Al 2 O(SO 4 ) 3 .18H 2 O) salt as a precursor in synthesis and two peaks of sulfur (S-2 s and S-2p) appeared close to C-1 s in the survey spectrum.Carbon peaks appear in all spectra as a result of atmospheric exposure before measuring the samples [100].
The XPS spectra of all the prepared samples identify the core levels of the main elements, which are deconvoluted using the Voigt function.The binding energies of the de-convoluted peaks are tabulated in table 4. First, the binding energies of Zn-2p spectra are found to be around 1021 and 1044 eV in samples F and C, and around 1023 and 1046 eV in sample A, as depicted in figure 8(a).These two major peaks are attributed to Zn-2p 3/2 and Zn-2p 1/2 , confirming the existence of the oxidation state of Zn 2+ in the samples [101].The deconvoluted peaks of Co-2p contain the spin-orbit doublets of Co-2p 3/2 and Co-2p 1/2 associated by satellite peaks, as shown in figure 8(b).The two de-convoluted peaks of Co-2p 3/2 attribute to Co 2+ in the octahedral (O h ) and tetrahedral (T d ) sites [102].Moreover, the core spectra of Mg-1 s in figure 8(c) indicate the presence of a single peak at around 1303 eV in samples F and C, and around 1305 eV in sample A, which is related to Mg 2+ ions [103,104].Figure 8(d) displays the core spectra of O-1 s, which consist of two peaks and reflect the −2 site occupancy of O.The former peak is attributed to lattice oxygen (O L ) contribution in the M-O bond, while the latter peak is attributed to the oxygen defect indicating the presence of some vacant oxygen (O V ) [105,106].Now regarding the different elements in the samples, the spectra illustrated in figures 9(a)-(c) show six peaks in Fe-2p, four peaks in Cr-2p, and two peaks in Al-2p.The spectra of Fe-2p contain the spin-orbit doublets of Fe (Fe-2p 3/2 and Fe-2p 1/2 ).These doublets are composed of two peaks, each arising from Fe 3+ located at the O h and T d sites and accompanied by their satellites [100,107].They are in agreement with the study achieved by Murugesan et al [108], which reported the structural, spectroscopic, and electrical properties of Zn substituted manganese ferrite (MnFe 2 O 4 ).However, in the Cr-2p spectra, the spin-orbit doublets of Cr are characteristics of   Cr-2p 3/2 and Cr-2p 1/2 .Each doublet consists of two peaks; the first is associated with Cr 3+ ions and the second is associated with Cr 6+ ions [109].According to Attia et al [110], the +3 valence state of Cr corresponded to Cr 2 O 3 , which is well correlated with the XRD results.As announced in the study of Zhou et al [111], the surface of Cr was found to be rich in Cr 6+ ions, which only appear at the surface.Finally, the Al-2p spectra exhibit two dominant peaks at 74.11 and 75.76 eV, corresponding to the distribution of Al 3+ ions at the O h and T d sites due to M-O and M-OH, respectively [112,113].These results revealed the mixed spinel structure of the prepared samples, due to the cation distribution at both O h and T d sites [114].

VSM measurements
The M-H curves of the prepared samples were examined at room temperature in the field ± 20 kG using VSM, as depicted in figures 10(a)-(c).In all the prepared samples, a small hysteresis loop is observed at low fields, as shown in the inset of figure 10.The magnetization increases as the applied field increases with attaining saturation for sample F and absence of saturation in samples C and A at high fields.These reveal that a higher field is needed to magnetize samples C and A compared to sample F. VSM measurements reveal that the prepared samples exhibit a weak ferromagnetic (FM) behavior, similar to those found in literatures [18,21,115].The magnetic behavior can be influenced by the grain/crystallite size, the distribution of cations, and the creation of a secondary phase [116].The FM behavior in ferrite NPs may originate from the disordering of cations coming from the powerful interaction of Fe 3+ ions between O h and T d sites [93].Similarly, the FM phase of chromite NPs can be created from the cationic disorder [16].On the other hand, in the case of aluminate NPs, the FM phase mainly depends on three causes such as the raise of free electrons and holes (as a result of doping), the presence of secondary phase, and the existence of defects like Zn and O interstitials or vacancies [106].
The measured values of saturation magnetization (M s ), remanent magnetization (M r ), coercivity (H c ), magnetic moment (μ m ), magnetic anisotropy (K ), and squareness ratio (S) were estimated and listed in table 5.The obtained M s , M r , and H c values are in agreement with those stated in literatures [16,21,96].By comparing the M s values of all samples, the highest value is found in sample F due to the strong exchange interaction of Fe 3+ between O h and T d sites which improves the magnetization [93].The M s of spinel ferrites is strongly dependent on the cation distribution [76].In addition, the surface spins in ferrite NPs do not obey the core anisotropy.However, the surface spins anisotropy in chromite NPs are responsible to the lowest value of M s [42].According to Sayed et al [117], the high M s provided notable antibacterial performance.Furthermore, the narrow M-H loops of the prepared samples provide small M r values, which positively affect biomedical applications [118].It is clear that sample F attains the maximum value, while sample A achieves the minimum value.The variation in H c is affected by size-dependent different contributions that act as coercivity sources.These factors refer to the effective anisotropy energy, including surface anisotropy, shape anisotropy, magneto-static anisotropy, magneto-crystalline anisotropy, and magneto-elastic anisotropy energy [119,120].An increase in the these anisotropy energies leads to a rise in the H c values [119,120].The surface anisotropy is associated with the NPs'  size [119].The data show that the relationship between size (table 1) and H c (table 5) is straightforward.Bakeer et al [21] noticed a comparable trend of variation.In addition to the influence of NPs' size, the shape anisotropy, which is related to the NPs' morphology, also affects H c .Slimani et al [119] reported that the NMs synthesized via the hydrothermal and sonochemical methods possessed a nearly spherical morphology with negligible shape anisotropy energy.On the other hand, Liu et al [121] stated that the cubic-shaped NPs had a larger contact area and exchange coupling effect than the sphere-shaped NPs, increasing their H c values.In the current study, the cubic shape contribution in sample C, as evidenced by TEM data, is one of the factors that strengthened its H c more than those of the spherically shaped F and A samples.Issa et al [120] reported that the effect of magnetostatic anisotropy was surpassed by the core-shell (core-surface) interface exchange coupling, increasing the coercive and exchange fields.The opposite phenomenon was obtained by a large shell and small core, resulting in a decrease in the coercive and exchange fields [120].The magneto-crystalline anisotropy is connected to the atomic structure of the host crystal and exhibits a preferred magnetization directions [119].A decrease in magneto-crystalline anisotropy reduces the H c values [122].Moreover, the magneto-elastic anisotropy corresponds to the stress inside the crystal lattice [119].For instance, dopant ion inclusion induces stresses inside the crystal lattice and affects the magnetic susceptibility of NPs [119].Furthermore, alongside the effective anisotropy energy, porosity is one of the factors influencing H c .El-Ghazzawy et al [123] stated that the H c value of cobalt-manganese-chromium ferrite (Co 0.7 Mn 0.3 Cr x Fe 2−x O 4 , 0 x 1) NPs increased with porosity.Hence, the highest H c value in sample C and the lowest H c value in sample A can be attributed to the above-mentioned factors.Moreover, the largest H c in sample C can also be related to mass-produced NPs, which may have contributed to the weakness of its FM nature [16,124].The small M r and H c values of the prepared samples indicate that they are soft FM materials, making them suitable for high density data storage and applicable in the fields of biomedicine and magnetic hyperthermia [118,125].Soft FM materials are characterized by narrow hysteresis curves, low H c , low magneto-crystalline anisotropy, and can be easily magnetized and demagnetized [118].The extracted M s values from the M-H curves were used to calculate the magnetic moment (μ m ), per formula unit in Bohr magneton (μ B ), according to the following equation [21]: where M w is the molecular weight of the NPs and 5585 is a magnetic factor.Moreover, the following equation is used to estimate the magnetic anisotropy [51]: It is clear that sample F, which reached saturation, has a relatively higher μ m and K values than those of samples C and A, which did not reach saturation.Subsequently, M s values vary proportionally with μ m and K values, as seen in table 5. Similar variation behaviors have been identified in literatures [21,126].The μ m values of samples F and A confirm with those reported in previous literatures [21,100].On the other hand, the K value of sample F is higher than that of the pure ZnFe 2 O 4 NPs prepared by Aridi et al [51].
The squareness ratio (S = M r /M s ) is defined as the ratio of remanent magnetization to saturation magnetization.If S has a value below 0.5, magneto-static interactions take place between the particles, whereas a value of 0.5 implies that the particles are non-interacting single domain [98,127].According to table 5, the S values of the investigated samples lie below this limit, indicating the creation of a multi-domain structure with uniaxial anisotropy [127].Subsequently, this demonstrates that the particles interact through magneto-static interactions.

Antibacterial activity 3.8.1. MIC determination
The MIC of NPs was recorded against Gram-positive (S. aureus and E. faecium) and Gram-negative (S. maltophilia and E. coli) bacteria.The findings of this investigation are presented in table 6.The minimum inhibitory effect of ferrite NPs was found to be the same against all the tested microorganisms with a concentration of 10 mg ml −1 .However, the MIC values extended from 5 to 10 mg ml −1 for chromite NPs and were in the range of 1.25 and 5 mg ml −1 for aluminate NPs.The MIC values of sample F confirmed with the study of Al-Jameel et al [128] in which comparable values were obtained against S. aureus and E. coli.However, the MIC values of samples C and A, against S. aureus and E. coli, conflicted with those reported by Sayed et al [117] and Subhan et al [129], respectively.It is noticed that the highest concentration needed to suppress bacterial growth was registered by sample F against all of the tested microorganisms, as well as sample C against E. faecium.On the other hand, the lowest concentration needed to suppress bacterial growth was recorded by sample A against S. aureus, E. coli, and E. faecium.It is very important to highlight that the inhibitory action could be influenced by several factors: physicochemical characteristics of NPs, environmental factors, exposure period, and bacterial species [1].The physicochemical conditions included (1) size, (2) shape, (3) roughness, (4) doping with various concentrations, and (5) zeta potential of NPs [1].In previous study, Wypij et al [130] reported that the NPs were considered effective against microorganisms if their sizes were less than 50 nm.This coincided well with the size of the prepared samples (table 1) in the present study, which did not exceed this limit and making them all effective.As stated by Bitar et al [33], particle size and inhibitory effect were inversely proportional, which agrees with our findings.The study of Franco et al [52] pointed out the influence of particle's size on the antibacterial effect and provided a good comparison of small and large NPs.They found that the small-sized NPs had a stronger antibacterial effect due to their capacity to enter cell and prevent bacterial growth [52].This enhancement was attributed to the fact that the small NPs had a higher surface-area-tovolume ratio (SA:V) than the larger NPs, which stimulated the generation of reactive oxygen species (ROS).Subsequently, the smallest size of sample A (table 1) contributed to its antibacterial activity (table 6).The observed differences in the NPs' inhibitory effects were also caused by differences in the NPs' morphology [131].Previously, a study by Slavin et al [58] related the efficacy of cubic-shaped NPs to their exposed facets and oxidation states.It is worth noting that when the particle had corners, the SA:V was higher, hence improving the antibacterial effect [58].The TEM results (figure 2) supported this statement, in which the cubic-shape contribution in sample C may have decreased their MIC values more than those of the spherically shaped sample F against S. aureus, S. maltophilia, and E. coli.However, the MIC values of sample C were higher than those of the spherically shaped sample A against S. aureus, E. coli, and E. faecium, implying that other factors influenced their antibacterial capacity.In addition to size and morphology, it was found that the presence of secondary phase increased bacterial growth inhibition [48].This is consistent with samples C and A in which the formation of secondary phases, as supported by the XRD data (figure 1), displayed a better antibacterial capacity than sample F with no secondary phase.Additionally, the appearance of secondary phase induced defects in the bandgap, according to UV analysis, which played the same role as corners in increasing SA:V and improving the antibacterial property [51,58].Moreover, the optical properties of NPs could influence their antibacterial activity.For instance, the higher absorbance peak intensity led to more production of excitons, more ROS generation, strengthening the inhibitory effect [93,94].Sagayaraj et al [132] revealed an inverse relationship between E g and antibacterial activity, which is consistent with our findings.According to UV data, sample A had the highest absorbance peak intensity (figure 6(a)) and the lowest E g value (table 3), which may have contributed to its antibacterial potential (table 6).Regarding the magnetic properties, the high M s improved the antibacterial activity, as mentioned in earlier research [117].According to VSM results, sample F with the highest M s value (table 5) gives noticeable antibacterial potential (table 6).Moreover, the low M r and H c values were preferred for biomedical applications [118,122].These values were the lowest for sample A (table 5) which resulting in a greater reduction in bacterial growth as compared to samples F and C (table 6).Another factor affecting the antibacterial activity was co-doping, as reported by Oves et al [133].In the present study, doping the prepared NPs with Zn, Co, and Mg induced fascinating antibacterial potential which may due to the adhesion/ inactivation of these divalent metal ions to the wall surface of bacteria.
All of the aforementioned factors had an impact on the inhibitory effect of the examined samples.However, the mechanism by which NPs limited bacterial growth remains unknown.The antibacterial potential of metal oxide NPs can be attributed to a variety of mechanisms.Metal donor atom selectivity mechanism was founded on the principle that bacterial membranes included macromolecules with high electronegativity that acted as adsorption sites for metal ions carrying the positive charge [134].Furthermore, metal ions were typically connected to some donor ligand atoms (such as S, N, and O) through powerful and specific interactions.Consequently, exogenous metal ions, or complexes of them, could substitute the original ions that exist in biomolecules, causing cellular malfunction [134].Another reason was the generation of ROS, which can initiate oxidative stress and peroxidation of lipids, resulting in oxidative damage of DNA within microorganisms.Subsequently, the biological interaction between NPs and macromolecules results in bacterial growth inhibition [134].

MBC determination
All samples could effectively inhibit the growth of all tested bacteria, but not all caused killing.The activity of NPs toward bacteria was related to the electrostatic attraction that occurs between the NPs with positive charge and the bacteria wall with negative charge [135].This triggered bacteria to die immediately owing to oxidation [134].The metal reduction potential was another mechanism of action of metal oxide NPs toward bacteria [134].This mechanism entailed the production of ROS, which resulted in oxidative stress and the destruction of DNA, proteins, and lipids, ultimately leading to bacterial death [33].Regarding the MBCs data listed in table 6, a visible bacterial growth was obtained for samples F and C against all the bacterial species and for sample A against S. maltophilia.Similarly, Almessiere et al [136] revealed no MBC for dysprosium-yttrium co-substituted manganese-zinc ferrite, Mn 0.5 Zn 0.5 Fe 2-2x (Dy x Y x )O 4 , for x = 0.00, 0.01, 0.02, and 0.03 against E. coli.In contrast, no visible bacterial growth was obtained for sample A at 10 mg ml −1 against E. faecium and 20 mg ml −1 against S. aureus and E. coli.These concentrations that caused bacterial death were found to be greater than those caused bacterial inhibition.The recorded amounts of sample A were found to be greater than those of Mg 1−x Zn x Al 2 O 4 NPs prepared by the combustion method with E-tirucalli latex as fuel, as demonstrated by Pratibha et al [137].
The mechanism of interaction between NPs and bacterial cell wall that caused bacterial death made some bacteria more resistant than others [1,138].This mechanism depended on the intracellular components, cell composition, and thickness of bacteria which differs from bacteria to other.Another assumption to be supposed for the resistant bacteria is an alteration in genes which occurred in two manners, either mutations in the DNA or horizontal gene transfer (HGT) [139].

Time-kill test
Time-kill test for MIC and MBC of NPs was studied against the tested bacteria as shown in figures 11(a)-(d).Log 10 CFU ml −1 was plotted versus time of incubation (h).The GC demonstrated an increase in bacterial cells over time.On the other hand, a reduction in bacterial cells was observed for all samples.In a similar way, the growth of S. aureus, E. coli, B. subtilis, and P. aeruginosa was significantly reduced after treatment with Cu-doped NiFe 2 O 4 NPs, as reported by Rajivgandhi et al [140].It is noticeable that, none of the prepared NPs attained the bactericidal endpoint at their MIC values (figure 11), proving that they are unable to kill bacteria at these concentrations, which are consistent with the MIC and MBC results (table 6).In addition, the bactericidal endpoint for the treated S. aureus and E. coli with sample A at their MBC were attained at 24 h after incubation as shown in figures 11(a) and (c), respectively.However, the time-kill endpoint for E. faecium treated with sample A was reached at 2 h after incubation (figure 11(d)).This revealed that sample A at its bactericidal concentration killed E. faecium faster than S. aureus and E. coli.In addition, S. aureus and E. coli showed the same sensitivity to sample A due to the similar MIC, MBC, and bactericidal endpoint of sample A against these two bacteria.On the other hand, sample A showed a lower time-kill endpoint against E. faecium than that of S. aureus and E. coli.This demonstrated that E. faecium was highly sensitive to sample A compared to S. aureus and E. coli.These findings revealed that the antibacterial activity of the prepared NPs depend on their concentrations and the tested microorganisms.The response was found to vary from NP to other, also from bacteria to other.As mentioned previously, various factors have been proposed for the antibacterial potential of NPs as well as the sensitivity of bacteria.Adnan et al [141] stated that the smaller particle size resulted in faster interaction between NPs and the bacterial cell wall, which reduced the time required to suppress bacterial growth.This caused the metabolism to slow down, leading to cell death [141].Moreover, the narrow E g induced a facile and effective migration of electrons from the valence band to conduction band [132].As a result, the production of ROS and the interaction of NPs with bacteria in a short period of time increased, inducing more bacteria damage [132].

Conclusion
A successful synthesis of Zn

Figure 1 .
Figure 1.XRD patterns with their corresponding Rietveld refinement performed on MAUD for (a) F, (b) C, and (c) A samples.The experimental data is represented by the black circles and the calculated patterns for F, C, and A samples are represented by the solid blue, green, and red lines, respectively.
throughout calcination.The D TEM values estimated from the particle size distribution histograms are found to be 27.763,40.723, and 15.399 nm for F, C, and A samples, respectively.These values are consistent with those obtained in literatures[55,62,63].The sizes determined by TEM for the prepared samples are in accordance with those calculated in XRD (table1).Šutka et al[64] and Shmait et al[65] stated that D TEM matched with D XRD of ZnFe 2 O 4 and ZnAl 2 O 4 NPs, respectively.Likewise, Manjunatha et al[66] reported that the D TEM and D XRD values of scandium doped CoCr 2 O 4 NPs (Co 1−x Sc x Cr 2 O 4 ) were closely compatible.Furthermore, it is noticed that sample C has a higher D TEM value than that of samples F and A, with an average side of 23.258 nm for the cubic particles.The average size estimated from TEM for sample C is done by taking into account both spherical and cubic particles.Consequently, the appearance of various forms in sample C is responsible for the high value of D TEM .According to previous research, a better antibacterial activity is obtained for the small-sized NPs[52].

Figure 2 .
Figure 2. TEM micrographs and their corresponding size distribution histograms for (a) F, (b) C, and (c) A samples.
(a)), respectively.The observed lattice fringes of figure 3(b) have d hkl of 4.79 and 2.48 Å, originating from the (111) and (311) planes of sample C, respectively.The lattice image of figure 3(c) with d hkl of 4.68 and 2.85 Å is indexed to the (111) and (220) planes of sample A, respectively.The d hkl values estimated from the HRTEM micrographs match well with those obtained from the XRD experiment.Similar d hkl values were reported by Yadav et al [71] and Siragam et al [63].The SAED patterns show small spots making up bright concentric rings, revealing the polycrystalline structure of all of the prepared samples.The diffraction rings observed from inside to outside in the SAED patterns are identified as the (hkl) planes, as depicted in figures 3(d)-(f).The (400), (422), (440), and (444) planes corresponding to d hkl of 2.06, 1.75, 1.46, and 1.20 Å, respectively, are indexed for sample F (figure 3(d)).However, the (400), (422), and (440) planes corresponding to d hkl of 2.02, 1.72, and 1.44 Å, respectively, are indexed for sample C (figure 3(e)).Additionally, the (400), (422), and (440) planes corresponding to d hkl of 1.97, 1.68, and 1.40 Å, respectively, are indexed for sample A (figure 3(f)).The d hkl values obtained from the SAED diffraction patterns concur with those estimated from the XRD measurements.According to the literatures, the obtained diffraction rings with their corresponding d hkl values are compatible with those of the cubic spinel ferrite, chromite, and aluminate phases [45, 72, 73].

Figure 3 .
Figure 3. HRTEM images for (a) F, (b) C, and (c) A samples with their corresponding SAED patterns for (d) F, (e) C, and (f) A samples.
and table 2. Sharp peaks referred to Zn, Co, Mg, Fe, and O constituents are detected in the EDX pattern of sample F without any impurity, as shown in figure 4(a).Additionally, figure 4(b) shows the main peaks of Zn, Co, Mg, Cr, and O species for sample C. The presence of Cr 2 O 3 impurity phase in sample C is verified by the larger amount of Cr in comparison with the other trivalent metals in samples F and A. This can be attributed to the evaporation of the

− 1 ,
as shown in figures 5(a) and (b), respectively.In the spectrum of sample F, two sharp bands at 424.89 and 569.86 cm −1 are attributed to the stretching vibration mode of Fe-O in octahedral and tetrahedral groups, respectively [77, 78].For sample C, three peaks located at 402.08, 535.02, and 632.05 cm −1 correspond to the metal-oxygen M-O (Cr(III)-O) bond [79, 80].Two intense bands assigned to the vibration of the Mg-O and Al-O bonds are recorded at 566.48 and 686.53 cm −1 , respectively, which are the characteristic peaks of sample A [22, 81].Three absorption peaks are shown only in chromite at 900.59, 942.05, and 1652.69 cm −1 which are ascribed to the stretching of the bond Cr-O, the vibration of the bond Cr(VI)-O and the O-H stretching vibration of water, respectively

Figure 5 .
Figure 5. FTIR spectra for F, C, and A samples (a) in the range of 4000-400 cm −1 and (b) in the range of 700-400 cm −1 .

Figure 7 .
Figure 7. XPS survey spectra for F, C, and A samples.

Figure 10 .
Figure 10.M-H curves with inset displaying the magnetization at lower fields for (a) F, (b) C, and (c) A samples.

Figure 11 .
Figure 11.Time-kill curve of tested bacteria (a) S. aureus, (b) S. maltophilia, (c) E. coli, and (d) E. faecium, after treatment with F, C, and A samples.Each point represents the relative value of viable bacteria at a specific time.

Table 1 .
Structural parameters for F, C, and A samples.

Table 2 .
Atomic percentage of the various constituent elements for F, C, and A samples.

Table 3 .
The direct bandgap energy (E g ) and Urbach energy (E u ) for F, C, and A samples.

Table 4 .
The binding energy of the de-convoluted main core level spectra for F, C, and A samples.Binding energy (eV)

Table 5 .
Magnetic parameters for F, C, and A samples.

Table 6 .
The average MIC and MBC values (mg ml −1 ) for F, C, and A samples against some Gram-positive and Gram-negative bacteria.
0.33 Co 0.33 Mg 0.33 X 2 O 4 NPs (X = Fe, Cr, or Al) was performed using the chemical coprecipitation method.XRD results asserted the formation of a cubic structure of the prepared samples with a weight % of 100% for sample F, 94.56% for sample C, and 97.73% for sample A. TEM micrographs of samples F and A displayed a nearly spherical shape with D TEM of 27.763 and 15.399 nm, respectively.On the contrary, sample C possessed two morphologies, cubic and nearly spherical shapes, with a D TEM of 40.723 nm and an average side of 23.258 nm.The presence of Zn, Co, Mg, Fe, Cr, Al, and O elements was proved by EDX patterns, and the experimental elemental amounts were consistent with the theoretical ones.FTIR analysis confirmed the identification of functional groups contained in the samples.Samples C and A showed lower E g of 2.429 and 2.355 eV and higher E u of 0.816 and 0.856 eV, respectively, compared to sample F which had E g of 2.967 eV and E u of 0.674 eV, due to the existence of secondary phase.XPS analysis confirmed the elemental composition, the valence states (Fe 3+ , Cr 3+ , Cr 6+ , and Al 3+ ), and the mixed spinel structure of the prepared samples.VSM analysis displayed a soft FM nature, for all the prepared NPs.Samples F and C recorded the highest M s (46.5025 emu g −1 ) and H c(52.1939 G)values, respectively, while sample A had the lowest M r (0.5769 × 10 -2 emu g −1 ), H c (6.2262 G), μ m (0.0127 μ B ), and K (1.3178 emu G g −1 ) values.Sample A inhibited the growth of S. aureus, E. coli, and E. faecium more effectively than samples F and C due to its smallest size, presence of secondary phase, highest absorbance peak intensity, and lowest E g , M r , and H c parameters.Only sample A was capable of killing S. aureus, E. coli, and E. faecium.The time-kill test demonstrated that sample A at MBC values killed S. aureus and E. coli after 24 h of incubation, and E. faecium after 2 h of incubation.The time-kill test results agreed the MIC and MBC results.In this regard, ternary ferrite, chromite, and aluminate NPs are promising alternatives to antibiotics against multidrug resistant bacteria.