A certain mass of each ferrite, suspended in sodium citrate solution 0.50 mol/L, was stirred with an ultrasonicator for 2 h at 60 °C to disperse any cluster of particles. By this treatment, the surface of the particles was modified to improve zinc hydroxide (Zn(OH)₂) coating. The two samples of modified ferrite particles were centrifuged and washed with acetone.

10 mL of a zinc acetate dihydrate (Zn(CH₃CO₂)₂·2H₂O) 0.55 mol/L solution and 10 mL of ammonium carbonate ((NH₄)₂CO₃) 0.80 mol/L were slowly and simultaneously added dropwise over 20 mL of surface-modified ferrites suspensions at 25 °C. This technique was adapted from a procedure previously published by Hong et al. to obtain a similar material [31].

Since magnetic stirring would cause aggregation of ferrite particles, the suspensions were mixed with nitrogen gas flow during the precipitation. The CO₂ released in the reaction also contributed to the stirring.

Due to the ammonium carbonate addition, the change in the pH caused the hydrated zinc oxide (ZnO∙xH₂O) to precipitate over the modified surface of the ferrite particles. The composites were centrifuged and washed with distilled water, diluted ammonia and absolute ethanol, and dried in a vacuum furnace. Then, they were heated for 4 h at 350 °C, to dehydrate the zinc oxide and gently milled.

These precipitations were done with different mass ratios between ferrite particles (either M00 or M20) and zinc acetate solution to prepare various samples. The names and nominal compositions of these samples are presented in Table 1. Finally, only two composites, labelled Z00 (SrFe₁₁O₁₉/ZnO) and Z20 (Sr_0.8Nd_0.2Co_0.2Fe_10.8O₁₉/ZnO), were chosen to carry out the kinetics experiments. All these samples were characterized by X-ray diffraction (XRD), scanning electron microscopy (SEM), and vibrating sample magnetometry (VSM).

Orange II (Figure 1) was prepared as its sodium salt by a coupling reaction between beta-naphtol (β-C₁₀H₇OH) and sulfanilic acid (para-aminebenzenesulfonic acid) in a sodium nitrite (NaNO₂) solution. Dye was recrystallized in ethanol-water, and 100 mg/L solution was made using distilled water.

Kinetics measurements were carried out with composites or ZnO suspensions. Control tests were performed on single ferrite particles or without light exposition. The results of the experiments are shown in Table 2. Samples for discoloration were prepared by mixing 4.00 mL of catalysts suspensions in distilled water (final conc. 1000 mg/L) and 1.00 mL dye solution (final conc. 20 mg/L).

Hydrated zinc oxide reacts either with hydronium or hydroxide ions [32]. Therefore, the pH is stabilized by combined acid-base/solubility equilibria of ZnO in the suspensions [33], which are presented here in equations (1)-(3) as follows:

$ZnO\cdot x{H}_2{O}_{(s)}+\left(x-1\right)H_2{O}_{(l)}$ (1)

$Zn{\left(OH\right)}_{2(s)}+2H^{+}{}_{(aq)}\rightleftharpoons Z{n}^{2+}{}_{(aq)}+2H_2{O}_{(l)}$ (2)

$$Zn{\left(OH\right)}_{2(s)}+2O{H}^{-}{}_{(aq)}\rightleftharpoons {\left[Zn{\left(OH\right)}_4\right]}^{2-}{}_{(aq)}$$ (3)

Figure 1

The azo dye Orange II as its sodium salt [34]

Samples were placed into glass test tubes (1.5 cm diameter) of high optic quality and attached around the external surface of a borosilicate glass flask (15.0 cm diameter) in a carrousel-like array. A 160 W mixing lamp Philips ML was placed at the center of the flask, and the whole set was put into a thermostat water bath which could be mechanically stirred. The array of reactor tubes can be seen in Figure 2.

Figure 2

Images of the carrousel array of tubes around the flask and the mixing lamp (a-d)

Sunlight was emulated with this mixing lamp. Its emission combines a continuous spectrum of an incandescent tungsten filament and a line spectrum from a high-pressure mercury source. During the kinetics experiments, the irradiance calculated value at the center of each test tube was 1900 W/m² approximately.

All experiments were performed at 25 °C. Samples were homogenized in darkness by mechanical stirring and, immediately, all the test tubes were attached to the flask. At t = 0, the lamp was switched on; the first tube content was extracted with a syringe and centrifuged for 1 min at approx. 800 g (8000 m/s²). The supernatant solution was put into a quartz cell; its absorbance was measured in a Shimadzu UV-2401 spectrophotometer at its maximum absorbance value within the visible spectrum, λ_max = 485 nm (reference: distilled water). This procedure was repeated with the next test tube after fixed intervals up to 240 min. Dye concentration values in each tube were calculated as a function of time using Lambert-Beer’s law [35].

The XRD patterns of all the composites, measured in a Philips diffractometer (using CuKα radiation), are shown in Figure 3. The profiles confirm the hexagonal phase of Sr ferrite and the main peaks assigned to ZnO.

Figure 3

X-ray diffractograms of M00 (Sr hexaferrite) and all the composites

Strontium hexaferrite, Z00 composite (SrFe₁₁O₁₉/ZnO), and zinc oxide nanoparticles were structurally characterized by SEM in a Zeiss DSM 982 GEMINI microscope. Figure 4 shows the SEM images. The hexagonal ferrite looks like platelets (average particle size 230 nm), while the ZnO nanoparticles are spherical and well dispersed (average particle size 60 nm).

Figure 4

SEM images, magnified × 90000 M00 Sr hexaferrite (a); Z00 composite (b); and ZnO nanoparticles (c)

Magnetization versus magnetic field (M-H) loops of the M00/ZnO composites were measured with a Lakeshore 7300 VSM, and their main magnetic properties were calculated (Table 3). Magnetic results of M20/ZnO composite (Z20) are not presented.

Since magnetization of a two-component system depends on component masses [36] and zinc oxide is a diamagnetic material, it is possible to ponderate mass composition from M_S results. In this way, for the Z00 composite, a ZnO mass fraction of 0.926 was calculated. The enhancement in H_C for the Z0L composite is related to the high dilution of the ferrite particles into a non-magnetic material. These composites show good magnetic properties after the one-step preparation method.

Figure 6 shows dye absorbance at 485 nm in the centrifuged Z00 composite supernatants divided by the initial value, Abs(t)/Abs₀ vs. time. By Lambert-Beer’s law, these relative absorbances should be equal to the relation C(t)/C₀, between the concentration at any time and the initial concentration at the beginning of the light exposition. Orange II solution behaviour without any catalyst is included for comparison (Experiment #1 - Table 2).

It can be seen in Figure 6 that dye concentration slightly varies in the absence of catalyst (orange symbol), whereas suspended Z00 catalyst particles (blue symbols) cause an exponential decay in the azo dye absorbance. This type of curve would match with a first-order kinetics reaction.

Figure 6

Orange II / Z00 composite discoloration with light exposition

Figure 7 shows Abs(t)/Abs₀ values as a function of time for the kinetics experiments carried out with 1000 mg/L of Z00, Z20, or ZnO nanoparticles suspended in the Orange II 20 mg/L solution.

Figure 7

Orange II / Z00 or Z20 composites and zinc oxide discoloration with light exposition (for comparison Orange II is included)

It can be seen in Figure 7 that Z00 composite and single zinc oxide suspended particles cause similar absorbance exponential decays. The Z20 composite (Nd-Co substituted strontium ferrite) curve presents a similar profile, although its initial slope is less negative than the others.

Kinetics of all photocatalytic reactions agreed with a first-order law, as Marto and other authors had found using only ZnO [8]. The first-order reaction rate constant (k₁) and half-life (t_½) parameters were calculated from the initial slopes of relative absorbance logarithmic graphs. These values are shown in Table 4.

The t_½ values for the 1000 mg/L catalyst suspensions were 49 min for Z00 composite and 90 min for Z20 composite. Then, the rate of the reaction catalyzed by composites decreased compared with that of the ZnO suspension with the same nominal concentration (t_½ = 43 min).

However, it should be considered that suspensions of Z00 composite have a real mass of zinc oxide 7.4% less than their nominal composition (ZnO mass fraction = 0.926). Therefore, it should be concluded that the decrease in the reaction rate using the composite instead of single ZnO should be lesser than the previously calculated value.

It can be seen in Figure 6 that the initial rate is related to the catalyst concentration up to a value of 250 mg/L. For a higher catalyst concentration (1000 mg/L), the discoloration rates behave similarly for both hexaferrite composites after 180 min, although their initial rates are somewhat different (Figure 7). It is noticeable that the Sr ferrite supported catalyst (blue symbol) presents a similar behaviour as single ZnO nanoparticles (red symbol).

No other analytical technique was used in this work to measure any degradation products from the Orange II. However, as control experiments in darkness showed no significant azo dye adsorption, we proposed that the main discoloration process should be the catalyzed oxidation of the azo group with light exposition. It can also be inferred that photocatalyzed azo dye reaction should be the same with all the tested catalysts, magnetic composites, or suspended ZnO nanopowders.

Figure 8 shows reaction rate constant k₁ values obtained from the experiments with ZnO powder and Z00 composite vs. catalyst loading. Within the tested range of catalyst concentration, the value of k₁ for the ZnO powder measurements increases linearly with the catalyst load. However, the rate-constant increment in Z00 experiments becomes less steep for amounts of suspended composite greater than 250 mg/L. This behaviour has already been reported and analyzed by Sakthivel et al. [37].

At first, a linear rise in the rate constant with the amount of catalyst is observed. This has been attributed to the increase of active sites on the photocatalyst surface, which results in a greater generation of •OH radicals, that would oxidize more azo dye molecules. But, as Nishio et al. found, for ZnO concentrations between 1000 and 1500 mg/L, the slope substantially decreases, and the rate constant becomes nearly independent of the catalyst load [3].

It has been reported by Kaur et al. that the increment of solution turbidity would hinder light transmission due to the significant scattering, and there would be no longer increment of hydroxide radical generation [38]. Therefore, there is an optimal level of zinc oxide concentration, and a further increase of the catalyst load will not be effective. This optimal value would seem smaller for the Z00 composite (between 250 and 1000 mg/L), probably due to a more intensive scattering of the suspended ferrite particles.

Figure 8

Variation of rate constant with catalyst loading