Materials used in this study include ferrous sulphate-heptahydrate (FeSO₄∙7H₂O), copper (II) acetate Cu(CH₃COO)₂, ammonium hydroxide 30% (NH₄OH), sodium hydroxide (NH₄OH), ferric chloride (FeCl₃∙6H₂O), ethylene glycol (CH₂OH)₂, hydrochloric acid (HCl), and AYR dye. All chemicals and materials used in this study were acquired from Sigma Aldrich Company.

Prepared ferrite nanoparticles (FNPs) with different molar ratios of (CuFNPs). FeSO₄·7H₂O, FeCl₃·H₂O, and Cu (CH₃COO)₂ were used in varying molar ratios of Cu²⁺ so that the FNPs could be represented as (Cu_xFe_1−x)O∙Fe₂O₃, where divalent Copper replaces Fe²⁺. The co-precipitation process involved adding a 30% ammonium hydroxide solution dropwise while continuously stirring to reach a pH of 10. The reaction was conducted at 50˚C, after which the mixture was allowed to cool to ambient temperature. Through centrifugation, the particulates were separated, rinsed multiple times with deionized water, and subsequently treated with ethyl alcohol. In the final stage, they were desiccated in a hot air oven at 80˚C for 12 hours. The obtained CuFNPs samples were denoted as CuFNPs (0), CuFNPs (1), CuFNPs (2), CuFNPs (3), and CuFNPs (4), corresponding to Fe₃O₄, (Cu_0.25Fe_0.75)O∙Fe₂O₃, (Cu_0.5Fe_0.5)O∙Fe₂O₃, (Cu_0.75Fe_0.25)O∙Fe₂O₃, and CuFe₂O₄, respectively.

Copper ferrite nanoparticles were synthesized using a microwave technique. In this method, 5.4 g of ferric chloride (FeCl₃·H₂O) and 1.81 g of copper (II) acetate Cu(CH₃COO)₂ were mixed with ethylene glycol at a solid-to-liquid ratio of 1:10. An enclosed Teflon tube contained the precursor mixture, which was then exposed to microwave radiation for 2 minutes at 700 W in a 2.45 GHz multimode cavity. This design took advantage of ethylene glycol’s high dielectric loss, allowing it to heat rapidly during microwave irradiation. The resulting solid was filtered, rinsed with ethanol, and desiccated. The obtained samples are denoted as CuFNPs (5).

The synthesized nanoparticles were analyzed using an X-ray diffractometer (XRD) (Shimadzu-7000, USA) for phase analysis and crystallite size measurement. Dye concentrations during the photocatalytic process were monitored using a UV-visible spectrophotometer (UV-1800 Shimadzu) with a resolution of 1 nm. The morphology and structure of the samples were examined using a scanning electron microscope (SEM; JEOL JSM-7600F). A Lake Shore Model 7410 (USA) Vibrating Sample Magnetometer (VSM) was used to explore the magnetic features.

The concentration of a stock solution was 1000 mg/L, achieved by dissolving 1 g of AYR dye in 1 L of distilled water. This solution was kept in a volumetric flask for preparing diluted AYR dye solutions. The Shimadzu UV-1800 UV-visible spectrophotometer was used to determine dye concentrations before and after adsorption. The concentrations were measured by absorbance within a wavelength range of 200 to 500 nm for AYR dye. Calibration curves were constructed using solutions of 25, 50, 75, and 100 mg/L for equilibrium and kinetic investigations.

In a typical batch adsorption experiment, 10 g of adsorbent was combined with 20 mL of 50 mg/L dye. The mixture was stirred at 25˚C throughout the experiment. The pH was initially adjusted to the desired value using small amounts of NaOH and HCl (0.1 mol/L). Experiments used 20 mL of 50 mg/L dye at varying pH levels (2 to 10). Different time intervals (15 to 150 minutes) were used to study the impact of contact time on adsorption. The effect of the initial concentration of AYR dye was examined using dye solutions of 25, 50, 75, and 100 mg/L with 10 mg of CuFNPs. Additionally, the effect of adsorbent mass (10, 15, 20, and 40 mg) on dye removal was studied by interacting with 20 mL of 50 mg/L dye at ambient temperature. The following equation was used to calculate the removal percentage (R%) of AYR dye (1) [21] :

$R\left(\%\right)=\frac{C_0-C}{C_0}\times 100\%$, (1)

where C₀ and C denote the dye concentration before and after a catalyst degrades it.

Using nutrient agar plates, qualitative evaluations were conducted [22] . A spore suspension of the pathogenic isolates was generated and calibrated to a concentration of approximately 0.5 McFarland standard (1.5 × 10⁸ cells/ml). Each plate contained 20.0 mL of sterile nutrient agar medium, and 25.0 µL of yeast and bacterial preparations were inoculated into it. After the medium cooled and solidified, 100 µL of the prepared samples were inserted into pre-formed wells on the agar plates, which had a diameter of 0.9 mm. Incubation was conducted at 37˚C for 24 hours after the inoculated dishes were refrigerated for one hour. The samples’ efficacy against the targeted bacteria was assessed by measuring their zones of inhibition (ZI) in millimeters [23] [24] .

XRD was used to determine the crystallite phase and calculate the crystallite size. The XRD pattern for the samples from CuFNPs (0) to CuFNPs (5) is shown in Figure 1 . According to the XRD figure, the pattern for the CuFNPs (0) sample aligns with JCPDS (79-0417) indicating a pure magnetite phase with 2θ = 29.75˚, 32.15˚, 35.75˚, 43.3˚, 57.1˚, 60.95˚, and 62.95˚. The XRD pattern of the CuFNPs (1) sample shows an amorphous nature and a pattern corresponding to magnetite, making it difficult to identify any peaks belonging to copper ferrite. This could be due to the small amount of ferrite and the amorphous nature of the pattern. The diffractogram for the CuFNPs (2) sample shows a pattern representing Fe₃O₄ and CuFe₂O₄, in agreement with JCPDS (79-0417) and (34.0425), respectively. Peaks for Fe₃O₄ are at 2θ = 29.6˚, 32.8˚, 34.4˚, and 61.1˚, while those for Cu Fe₂O₄ are at 2θ = 43.7˚, 35.15˚, 58.15˚, and 70.05˚. The CuFNPs (3) sample shows an amorphous nature with a pattern mainly assigned to copper ferrite at 2θ = 35.4˚, 43.25˚, 56.9˚, and 69.5˚. The few peaks belong to magnetite at 2θ = 32.8˚ and 61.7˚. On the other hand, the CuFNPs (4) sample represents the XRD pattern in agreement with JCPDS card number 34-0425. The assigned pattern includes peaks at 2θ = 16.25˚, 26.7˚, 35.15˚, 38.75˚, 43˚, 55.9˚ and 48.5˚. The CuFNPs (5) sample showed a well-crystalline pattern with peaks assigned at 2θ = 16.75˚, 22˚, 26.7˚, 28.9˚, 30.65˚, 32.6˚, 34.05˚, 35.45˚, 40.85˚, 43.05˚, 44.8˚, 49.1˚, 54.45˚, 57.4˚, and 93.9˚. Remarkably, the sample prepared using microwave assistance exhibited a crystalline structure, with its crystallite size remaining relatively unchanged and comparable to that of the traditionally prepared sample. Comparing the XRD results of the as-prepared non-stoichiometric copper ferrite reveals that, despite the smaller crystallite size of the microwave-assisted sample compared to traditionally prepared samples, it exhibits a more defined crystallite pattern. This may be attributed to the shorter reaction time in the microwave process compared to the traditional method. However, the thermal energy involved in the microwave preparation method is much greater than that in the traditional co-precipitation method. Therefore, the microwave-prepared sample possesses sufficient thermal energy for good crystallinity, but the crystallization time is too brief to allow for an increase in crystallite size.

Scherer’s equation was applied to analyze the crystallite size of the samples:

$L=\frac{0.89\lambda }{\beta \cos \theta }$, (2)

where the diffraction peak’s full width at half maximum (FWHM) is designated by β, the crystallite size by L, the X-ray wavelength is 0.154 nm, and the Bragg peak position by θ.

The morphology of samples was investigated using SEM. The SEM of synthesized CuFNPs (0) Figure 2(a) showed agglomerated nanoparticles with some degree of homogeneity. The particle size in CuFNPs (1) appeared smaller than that of sample CuFNPs (0) ( Figure 2(b) ). This is likely due to the initial formation of the ferrite phase, which reduces agglomeration. Samples CuFNPs (2) and CuFNPs (3) ( Figure 2(c) , Figure 2(d) ) exhibited a similar trend of non-magnetic particle aggregation. Sample CuFNPs (4) Figure 2(e) showed pure CuFe₂O₄ agglomeration from nano-sized particles and had a lower degree of homogeneity than the other samples. Sample CuFNPs (5) is shown in Figure 2(f) . From the image, it is clear that the CuFe₂O₄ particle size is much smaller than in samples prepared by the traditional co-precipitation method. This reduction in particle size may be due to the shorter preparation time, which limits the growth of crystallites and particles. Additionally, the solvent used (ethylene glycol) reduces particle aggregation during precipitation. The high thermal energy from the microwave-assisted technique also enhances the crystallinity of the produced CuFe₂O₄.

The assessment of magnetic materials requires examining their magnetic characteristics. This study investigated the relationship between the intensity of the applied magnetic field (H) and magnetization (M) for various samples. The hysteresis loop of the sample CuFNPs (0), shown in Figure 3(a) , which contains no copper and is essentially Fe₃O₄, exhibited a saturation magnetization (Ms) of 58.28 emu/g and a coercivity of 15.26 Hci. The Ms value aligns with existing literature, although the coercivity was notably higher than reported in the literature [25] . Given that iron oxide behaves as a superparamagnetic material at room temperature with particle sizes below 10 nm, the larger crystallite size of Fe₃O₄ in this study resulted in non-superparamagnetic behavior [26] . As copper ions were introduced into the ferrite structure, significant changes were observed. For CuFNPs (1) with 25% copper, as shown in Figure 3(b) , the Ms dropped to 6.8 emu/g due to the introduction of Cu²⁺ ions, leading to surface inhomogeneity and an anisotropic layer, while coercivity decreased to 7.24 Hci [27] . For CuFNPs (2) with 50% copper, Figure 3(c) shows the hysteresis loop indicating superparamagnetic properties, with a decrease in saturation magnetization to 0.75 Hci, attributed to reduced particle size. This trend continued with CuFNPs (3) Figure 3(d) , containing 75% copper, which maintained superparamagnetic characteristics, showing an Ms of 28.9 emu/g and a coercivity of 0.74 Hci. The pure CuFe₂O₄ sample, CuFNPs (4), exhibited similar linear behavior in Figure 3(e) , with an Ms of 41.8 emu/g, which is smaller than that cited in previous literature (33.3 emu/g) [28] [29] . Also, it showed increased coercivity of 0.52 Hci, linked to reduced particle size. The microwave-assisted preparation of pure CuFe₂O₄ (CuFNPs (5), Figure 3(f) produced an Ms of 53.75 emu/g and a coercivity of 0.91 Hci, slightly lower than conventionally prepared samples, possibly due to the formation of small metallic clusters of Cu and Fe from rapid reactions [30] . Overall, the variations in saturation magnetization across different samples were primarily due to changes in particle size, except for the microwave-assisted sample, which exhibited unique properties. Coercivity generally increased with decreasing particle size. The magnetic parameters measured were close to those reported for bulk CuFe₂O₄, attributed to crystallite sizes near the single-domain range (CuFe₂O₄ 24 m²) [31] .

To evaluate the adsorption efficiency of the CuFNPs adsorbent samples, preliminary experiments were conducted under varying parameters, including contact time, adsorbent mass, initial adsorbate concentration, and pH, to determine the optimal conditions for eliminating AYR dye from water-based solutions.

The contact time between pollutants and sorbent material is an important factor affecting the efficiency of dye removal. Therefore, a series of experiments was conducted at various contact times of the dye solution. The experimental parameters were set at a pH of 6.7, an adsorbent dosage of 10 mg, and 20 mL of AYR dye at a concentration of 50 ppm, all conducted at ambient temperature. During the initial phase of the adsorption process (0 - 15 minutes), the apparent rate constant of the reaction increased with extended contact time, attributed to the abundance of active sites on the surfaces of the nanoparticles available for AYR dye adsorption (see Figure 4 ). For comparison, a test for AYR dye adsorption on different molar ratios of nanocrystalline copper ferrite was conducted under the same conditions at different contact times. The results indicate that the removal percentage increases with a higher molar ratio of copper ferrite and longer contact time. The optimum contact time after 150 minutes for maximum removal for the six samples from CuFNPs (0) to CuFNPs (5) was 37.28%, 43.44%, 63.41%, 80.91%, 94.03%, 98.89%, respectively. On the other hand, the increase in removal efficiency for AYR can be attributed to the higher molar ratios of copper ferrite. This is because Cu²⁺ ions play a key role in enhancing the performance of the samples in removing the dye. By contrast, although the two samples, CuFNPs (4) and CuFNPs (5), contain the same molar ratios of copper ferrite, we noticed a higher removal efficiency for sample CuFNPs (5), which was prepared by microwave. The rapid transformation by microwaves helps form nanoparticles with smaller sizes and a high surface area. Hence, many active sites are created on the surface of CuFNPs (5).

A series of experiments were conducted using various adsorbent masses added to the dye solution, as illustrated in Figure 5. The experimental settings included a pH of 6.7, a dye concentration of 50 ppm in 20 mL, and an initial temperature of room temperature for 90 minutes. The adsorbent mass ranged from 10 to 40 mg. The findings indicated that increasing the adsorbent mass significantly improved removal efficiency by providing more active sites for AYR particle binding. The data showed that the adsorption percentage of AYR dye increased with higher quantities of CuFNPs. The maximum removal efficiencies for the six samples, from CuFNPs (0) to CuFNPs (5) at a dosage of 40 mg, were about 73.2%, 86.09%, 87.65%, 89.45%, 95.84%, and 98.85%, respectively. An adsorbent weight of 10 mg was chosen as the optimum weight for studying other factors due to the minor increase in removal efficiency with doses above 10 mg.

AYR dye solutions with concentrations between 25 and 100 mg/L were prepared to study the impact of initial dye amounts on removal efficiency. The experiments were conducted at a pH of 6.7, with an adsorbent mass of 10 mg and a dye concentration of 50 ppm in 20 mL, at room temperature for 90 minutes. Figure 6 shows that the optimal contact concentration was 25 ppm at equilibrium time for maximum removal. However, the removal percentage declined as the initial dye concentration increased. For example, the removal efficiency for CuFNPs (5) decreased gradually from 99% to 95.17% as the initial dye concentration increased from 25 to 100 ppm. It is evident that the quantity of dye adsorbed decreases as the initial dye content increases. Higher initial dye concentrations result in lower percentages of dye being adsorbed due to the increased competition for active sites on the CuFNPs surface, eventually leading to saturation of these sites.

The impact of pH on AYR removal was examined, with values ranging from 2 to 10, as illustrated in Figure 7 . The removal efficiency increases in acidic conditions, with a pH decrease from 6.8 to 2. The maximum removal at pH 2 for the six samples from CuFNPs (0) to CuFNPs (5) was approximately 79.30%, 81.35%, 86.15%, 88.96%, 93.42%, and 98.90%, respectively. This is because the ionized CuFNPs functional groups form strong electrostatic contacts with the negatively charged anionic dye molecules and the positively charged CuFNPs surface. On the other hand, removal effectiveness dropped for the six samples from CuFNPs (0) to CuFNPs (5) to about 20.24%, 31.31%, 49.89%, 59.86%, 80.03%, and 96.20%, respectively, at pH 10. This decrease at basic pH values occurs because anions compete with excess OH ions for adsorption sites on CuFNPs, destabilizing the anionic dyes.

The pseudo-first-order (PFO) and pseudo-second-order (PSO) chemisorption models are frequently used to analyze the kinetics of aqueous pollutant removal with various adsorbents [32] . Adsorption kinetics are important as they provide insights into the adsorption mechanism, which is crucial for enhancing process efficiency [33] . In this study, adsorption data for CuFNPs at different time intervals were examined using the PFO and PSO kinetic models. The adsorption rate is a critical factor when selecting an adsorbent, as it needs to exhibit both high adsorption capacity and rapid adsorption rate [34] .

The principle of the PFO kinetic approach is that the rate of variation in solute absorption over time is directly correlated to the difference between the concentration at equilibrium and the quantity of solute adsorbed at that time [35] .

The first-order kinetics, as postulated by Lagergren, is denoted by the subsequent equation (3) [36] :

$\ln \left(q_e-q_t\right)=\ln q_e-K_{1}t$. (3)

The values of q_e and K₁ can be obtained using linear regression of experimental data shown as ln(q_e − q_t) vs time t. In this context, q_e and q_t denote the adsorption amounts per unit mass at equilibrium and at any specific time, respectively, where K₁ signifies a PFO rate constant. The equilibrium adsorption capacity q_e and the PFO rate constant K₁ can be experimentally measured from the intercept and slope derived by graphing ln(q_e − q_t) versus time. Figure 8(a) demonstrates that the PFO model does not align well with the kinetic data, as shown by the low regression coefficient R² presented in Table 1 .

The PSO model principle takes into account the different phases during adsorption processes, effectively explaining the majority of sorption systems. PSO characterizes the adsorption process as chemisorption and can be represented as [37] :

$\frac{\text{d}{\left(AC\right)}_t}{\text{d}t}=K{\left[{\left(AC\right)}_0-{\left(AC\right)}_t\right]}^2$, (4)

where (AC)_t represents the number of active sites occupied on the CuFNPs at any time, and (AC)₀ is the number of active sites available on the adsorbent surface.

The kinetic rate equations can thus be rewritten as follows:

$\frac{\text{d}q}{\text{d}t}=K_2{\left(q_e-q\right)}^2$, (5)

In this context, k₂ represents the rate of sorption (g/mg min), while q_e (mg/g) and q (mg/g) denote the amounts of dye adsorbed at equilibrium and at any time, respectively. Separating Equation (5) gives:

$\frac{\text{d}q}{{\left(q_e-q\right)}^2}=K_2\text{d}t$. (6)

Integrating Equation (6) for the boundary conditions t = 0 to t = t and q = 0 to q = q gives:

$\frac{1}{q_e-q}=\frac{1}{q_e}+K_{2}t$, (7)

which is the integrated rate law for pseudo-second reaction.

These values can be obtained by plotting t/q_t against t, as shown in Figure 8(b) . The linear relationship of t/q_t versus t, based on the PSO kinetic model, fits the kinetic data well, as indicated in Table 1 , with a very high regression coefficient close to unity (R² = 1) for sample CuFNPs (4) and equal to (R² = 1) for sample CuFNPs (5). This suggests that dye adsorption onto the adsorbent involves chemical bonding between the adsorbate and the functional groups on the adsorbent’s surface. This material’s adsorption capacity is substantially enhanced by these chemical interactions. At equilibrium, the equilibrium adsorption is impacted by the quantity of adsorbate on the adsorbent’s surface and the quantity adsorbed.

Adsorption isotherms are used to evaluate the efficacy of an adsorbent in removing specific contaminants and to better understand the adsorption mechanism. They demonstrate the equilibrium relationship between the adsorbent and the adsorbate [38] . Langmuir and Freundlich models are frequently implemented to approximate equilibrium adsorption data among the numerous adsorption isotherm models.

The Langmuir model suggests that the adsorbate molecules do not interact intermolecularly, and the pigment is adsorbed in a monolayer on the adsorbent’s surface. Equation (8) illustrates this in its linear form. The linear plot of the Langmuir isotherm is depicted in Figure 9(a) .

$q_e=\frac{q_{\max }{k}_L{C}_e}{1+k_L{C}_e}$, (8)

where k_L is the Langmuir constant, q_m represents the maximum adsorption capacity, and n denotes the adsorption intensity.

The relevant Langmuir parameters and correlation coefficients are provided in Table 2 . The slope of (1/q_max) and the intercept of (1/k_L) are obtained from Figure 9(a) . To predict the efficiency of adsorption, the key features of the Langmuir isotherm are described using a dimensionless constant known as the separation factor, or equilibrium parameter R_L. This parameter sheds light on the attraction between the adsorbate and the adsorbent and is expressed by the following equation (9):

$R_L=\frac{1}{1+k_L{C}_0}$, (9)

where C₀ (mg/g) represents the amount of dye adsorbed per unit mass at equilibrium, and k_L is the Langmuir constant.

The following Table 2 shows the values of R_L. The R_L values for AYR dye were more than zero and less than unity, indicating good adsorption. Furthermore, it was found that R_L values decrease as the initial concentration increases, suggesting that AYR dye adsorption becomes more favorable. The adsorption process follows the Langmuir adsorption isotherm. This means that the surface is homogeneous, all sites are energetically equivalent, and the energy of adsorption is equal for all sites. The process involves monolayer coverage only, without interactions between adsorbate molecules on neighboring sites. The Freundlich adsorption isotherm equation was applied to AYR dye. The linear form of the Freundlich isotherm can be represented by an equation (10).

$\log q_e=\log k_f+\frac{1}{n}\log C_e$, (10)

where k_f represents the Freundlich constants, q_m is the maximum adsorption capacity, and n indicates the adsorption intensity.

The plot of logq_e against logC_egives the constants $\frac{1}{n}$ and k_f, corresponding to

the slope and intercept, respectively, as shown in Figure 9(b) . The corresponding Freundlich parameters and correlation coefficient are reported in Table 2 . The Langmuir model provides a significantly better fit for the experimental equilibrium adsorption data of AYR dye on the surface compared to the Freundlich model, as indicated by a higher R² value. The kinetic study shows that the process is controlled by the chemisorption step. This is evident because AYR dye, an azo dye containing N = N with an electron cloud, has functional groups that acquire a negative charge. These characteristics allow the addition of ferrite molecules, providing the composite with an active center, due to the presence of Fe²⁺ and Fe³⁺, which can serve as centers for chemical bonding between the dye and the composite.

The antibacterial activities of CuFNPs samples were evaluated against four bacterial strains, including both gram-positive and gram-negative bacteria, as well as pathogenic yeast. The antibacterial findings are presented in Figure 10 and Table 3 . In this experiment, CuFNPs (5) demonstrated exceptional antimicrobial efficacy against B. cereus and S. aureus, with inhibition zones of 39.0 to 41.0 mm, and were also effective against P. aeruginosa and E. coli, showing inhibition zones of 43.0 mm and 37.0 mm, respectively. Additionally, it exhibited broad-spectrum activity against the yeast Candida albicans, with a 37.0 mm inhibition zone. This effectiveness is attributed to the disruption of the cell membrane structure and the inhibition of the budding process [39] . On the other hand, based on SEM results, the nanosize of CuFNPs contributes to providing a higher surface area, leading to better contact. In contrast, CuFNPs samples (1) to (4) showed lower activity against B. cereus and S. aureus, with inhibition zones ranging from 16.0 to 26.0 mm, and lacked efficacy against the pathogenic yeast C. albicans. Furthermore, CuFNPs (1) and (2) exhibited antibacterial activity primarily against gram-positive and gram-negative bacteria. Overall, CuFNPs (5) is identified as the most promising antimicrobial agent among the tested samples. The mode of action for CuFNPs (5) involves inhibiting the bacterial cell wall’s mucopolysaccharide component, peptidoglycan, and includes various actions contributing to its bactericidal effect [40] . These actions include inhibition of cell wall biosynthesis, protein biosynthesis, membrane disruption, nucleic acid synthesis inhibition, mycobacterial ATP synthase inhibition, direct interaction with PBPs, inhibition of transpeptidase activity, binding to the peptide chain of peptidoglycan subunits, blocking transglycosylation and transpeptidation, and blocking transport of peptidoglycan subunits across the cytoplasmic membrane.

NiL: No antimicrobial activity recoded; *CN: Gentamycin standard antibiotic disc (10.0 mcg).