1 Introduction

The aquatic environment polluted by dyes and pesticides is conventionally treated by physical and/or biological techniques which reduce the pollution level, but are not enough to reach the threshold required by the water standards [1,2,3,4]. The advanced oxidation process (AOP) is an attractive technique for water depollution, particularly for the effluents of the pharmaceutical industry [5,6,7]. In this respect, the semiconductor–liquid junction, assimilated at microphotoelectrochemical (PEC) cells, can be used as a photocatalytic system able to decontaminate polluted water [8,9,10]. The spinels based on 3d metals are photocatalysts of choice which begin to gain popularity in the solar energy conversion owing to their chemical stability, nontoxicity, and absorption in the solar spectrum [11, 12]. Under irradiation, they can degrade organic pollutants because the potential of their conduction band is above the O2/O2• couple [13, 14], produce hydrogen [15], or reduce heavy metals to element states [16].

With a forbidden band of ~2 eV, the spinels MM′2O4 absorb ~40% of the solar light; M and M′ belong to the first row of transition metals. The optical transition is of d–d characteristic, it involves lower and upper bands of 3d orbitals coming from the crystal field splitting, and consequently, the photocorrosion is inhibited in aqueous electrolytes [17, 18]. In addition, the band position does not change with pH and can be adequately positioned with respect to redox levels in solution by a judicious choice of pH.

This work is a systematic investigation of the photocatalysis for the degradation of organic molecules (drugs, pesticides, and dyes) in our laboratory [19]. The aim of the present contribution is to report the synthesis and characterization of the spinel NiMn2O4 by sol–gel method and the PEC properties. The technique involves the precursor hydrolysis and a polycondensation to glass-like form. The reagents are mixed at an atomic scale and this should accelerate the reaction rate, leading to the nanocrystallite formation. As an application, the photocatalytic performance is tested through the oxidation of benzamide, a recalcitrant molecule on the heterosystem NiMn2O4/TiO2 upon visible light. The activity is dependent on some parameters such as the sensitizer dose, the benzamide concentration, and pH. The benzamide oxidation was monitored by high-performance liquid chromatography (HPLC) and the conversion rate reached 94% under solar light.

2 Experimental

NiMn2O4 was prepared by sol–gel method; the detailed procedure was reported elsewhere [20]. Briefly, stoichiometric amounts of Ni(NO3)3•7H2O (Merck, 99.5%) and Mn(NO3)2•6H2O (Merck, 99.5%) were dissolved in water containing the gelling agent (Agar-Agar Flucka, 1 g L−1). The solution was heated at 70 °C and the gel was dehydrated at 130 °C under magnetic stirring on a hot plate until apparition of a gray color. At this level, the particles are dispersed in the solvent, and a colloidal suspension is formed. Then, the colloids in the solvent are linked by sol condensation, to form a three-dimensional open grid (gel) and this constitutes the gelation process. Finally, the sample was ground in an agate mortar and heated at 850 °C (18 h) in a muffle furnace with intermediate regrinding, the end product exhibits a black color.

The thermogravimetry analysis (TG) was performed in air at a heating rate of 3 °C min−1 using a thermobalance (Setaram, Setsys 16/18). TiO2 was prepared according to our previous work [21, 22]. The formation of the phases was confirmed by X-ray diffraction (XRD) over the 2θ range (15–100°) using a Siemens diffractometer (Model D-5000). The FTIR analysis was conducted on a pressed pellet using 1 mg of the spinel dispersed in 150 mg of KBr of spectroscopic quality. The TEM image was taken with a Hitachi S2500. The diffuse reflectance was determined with a UV-Visible spectrophotometer (Specord 200 Plus). The photoactivity was tested through the degradation of benzamide. The experiments were performed in a Pyrex reactor equipped with a cooling system whose temperature was maintained at 25 °C. The tests were done in batch mode at neutral pH using 100 mL of benzamide solution at different concentrations (5–20 mg L−1) and variable spinel doses {Y% = x/(x + 125) × 100}, x is the mass of NiM2O4, while the mass of TiO2 (Riedel-de-Haën) is maintained constant (125 mg).

Before irradiation, the mixture was sonicated for 2 min in order to disperse the catalysts and maintained in the dark for 1 h to reach the absorption equilibrium. The catalyst powder was dispersed by magnetic agitation (200 rpm) using a double-walled Pyrex reactor with 0.5 cm of water which absorbs the IR radiation. The artificial light (Tungsten lamp: 5 mW cm−2), measured with a calibrated light meter (Testo 545) and solar irradiation, was used as light sources.

Aliquots (0.5 mL) of the solution were withdrawn at regular time intervals and subjected to filtration to separate the solid particles and analyzed. The remaining concentration of benzamide was titrated by HPLC equipped with a C18 column. An aliquot of 10 µL of solution were filtered through a 0.45-mm Millipore filter (Whatmann) and injected in the chromatograph (Jasco PC1201). The wavelength of the maximum absorption (618 nm) was taken from the UV-Vis spectrum. The photocatalytic yield was calculated from the relation

$${{\eta}} = 100\times \left( C_{\mathrm {o}} - C_{\mathrm{t}} \right){\mathrm{/}}C_{\mathrm{o}},$$
(1)

where Co is the initial concentration and Ct the concentration after irradiation for time (t); no benzamide was degraded by photocatalysis. The solutions were made up with CO2-free distilled water (conductivity ~0.7 MΩ cm).

3 Results and discussion

TG analysis was undertaken to delimit the temperature domain of the synthesis and thermal stability of the spinel; Fig. 1 shows the TG plot of the nitrate mixture in air obtained just after water vaporization. The mass decreases slightly up to ~400 °C due to water removal after which it undergoes a drastic weight loss until ~520 °C followed by a second weight loss at ~690 °C attributed to nitrates decomposition. The weight levels of beyond 800 °C indicate the formation of the spinel. The DTG peak curve at ~770 °C confirms the synthesis temperature of NiMn2O4.

Fig. 1
figure 1

TGA plot of the precursors indicating the formation of NiMn2O4 in air

Full size image

The prepared spinel is the single phase and crystallizes in a cubic symmetry (SG: Fd3m). All XRD peaks (Fig. 2) belong to the inverse spinel Mn[NiMn]O4 and agree with the JCPDS cards N° 11-1110. The structure consists of a close-packed arrangement of O2− ions with one-half of octahedral cavities occupied by Ni2+ and one-eighth of tetrahedral sites occupied by Mn3+ (Fig. 3a). The TEM image (Fig. 3b) gives an average grain size of 310 nm, which is greater than that calculated from the full width at half maximum (L ~63 nm, =0.9 λ (β cosθ)−1) and this clearly indicates the presence of agglomerates which form grains. The L-value gives an active surface area of ~18 m2 g−1 {Ssp = 6 (dexp L)−1}, assuming that compact crystallites with a spherical shape, dexp (=5.01 g cm−3) are the experimental density.

Fig. 2
figure 2

The powder X-ray pattern of NiMn2O4

Full size image
Fig. 3
figure 3

a The spinel structure of NiMn2O4, T tetrahedral coordination, O octahedral coordination. b The TEM micrograph of NiMn2O4 prepared by sol–gel

Full size image

The optical properties of NiMn2O4 are typical of semiconductors and are therefore not different from those of other spinels [23, 24]. The absorption coefficients (α) and the incident energy (hν) are given by the Tauc relation

$$\left( {\mathrm{\alpha h}}\mathrm{\nu} \right)^n{\mathrm{ = }}{\mathrm{Cont}} \times \left( {\mathrm{h}}{\mathrm{\nu}}-E_{\mathrm{g}} \right),$$
(2)

where the exponent n depends on the type of transition: n = 2 or 1/2, respectively, for indirect or direct transitions. The extrapolation of the straight line (αhν)2 to the energy axis shows a direct transition with an Eg value of 1.75 eV (Fig. 4).

Fig. 4
figure 4

Variation of (αhν)2 plot vs. the photon energy (hν) for NiMn2O4

Full size image

The key parameter of a photocatalyst is its semiconductivity and the transport properties are undertaken for this purpose. The small electrical conductivity (σ300 K ~10−5 Ω−1 m−1) is characteristic of semiconducting behavior with a low doping density (NA) and an extended depletion width (δ), where the electron/hole (e−/h+) pairs are separated by the junction electric field (see below). The thermal variation σ(T) follows an Arrhenius-type law with activation energy of 0.33 eV. This trend is confirmed by the positive thermopower (+130 µV K−1 at 300 K) which indicates that the majority of carriers are holes with a concentration of ~1018 cm−3.

Benzamide is weakly dissociated and the intensity-potential J(E) characteristic of NiMn2O4 electrode is plotted in neutral solution (pH ∼7) using Na2SO4 (10−2 M) as a supporting electrolyte. The curve exhibits a plateau region with a dark current less than 0.2 mA cm−2 and which becomes cathodically large below −0.2 V, and the absence of a diffusion plateau indicates a water reduction. The spinel is known to produce hydrogen under illumination [25, 26]. The photocurrent (Jph) appears at ∼0.3 V (photocurrent onset potential Von) and increases toward cathodic potentials, confirming the p-type behavior. However, the flat-band potential (Efb) under the operating conditions is accurately obtained from the Mott–Schottky relation

$$\frac{1}{{C^2}} = \left( {\frac{2}{{\varepsilon \varepsilon _0eN_{\mathrm{A}}}}} \right)\left( {E_{\mathrm{{fb}}} - E - kT{\mathrm{/}}e} \right),$$
(3)

where ε (=50) is the dielectric constant of the spinel, εo is the dielectric constant of vacuum (8.85 × 10−12 F m−1), e is the electron charge, NA is the hole density, kT is the thermal energy (~26 meV at 300 K), and E is the applied potential. The intercept of the fitted line at C−2 = 0 (Fig. 5) and the slope gives, respectively, the potential Efb (−0.20 V) and holes density (NA = 1.3 × 1016 cm−3); the NA value produces an extended depletion width (δ ~400 nm) which is greater than the penetration depth (α−1). The plateau region above −0.2 V corresponds to the accumulation zone with an increased recombination rate of (e−/h+) pairs, while the bending below −0.7 V is due to the inversion of electronic bands. The potentials of the valence (VB) and conduction band (CB) are crucial in photocatalysis and are given by

$$E_{\mathrm{CB}} = 4.75 + e\,E_{\mathrm{fb}} + 0.056\left( {\mathrm{pH} - \mathrm{pH}_{\mathrm{pzc}}} \right){\mathrm{ + }}E_{\mathrm{a}}$$
(4)
$$E_{\mathrm{VB}} = E_{\mathrm{CB}}{\mathrm{ + }}E_{\mathrm{g}}.$$
(5)
Fig. 5
figure 5

The Mott–Schottky plot of p-type NiMn2O4

Full size image

The activation Ea (~0.33 V) was determined from the conductivity measurements on sintered pellets. The value of CB (−3.13 eV/−1.62 V) and VB (−4.88 eV/0.13 V) indicates that both bands are made up of 3d orbital (t2g−eg), which takes its origin from the crystal field splitting of 3d metal.

3.1 Photocatalysis

The adsorption is widely used for the removal of organic molecules; however, it remains a displacement of the pollution and the used pollutant must be converted to less-harmful forms (ideally into CO2 and H2O) to regenerate the catalyst powder and this needs a further energy. By contrast, the environmental photoelectrochemistry is an emerging strategy for water decontamination [27, 28] and AOP requires radicals O2• and/or OH• formed in the conduction and valence bands, respectively, which should destroy the organic matter. Accordingly, the presence of dissolved oxygen is necessary for the photocatalytic process. Indeed, it has been reported that the bubbling solution by nitrogen considerably inhibits the photoactivity [29].

Coupling two SCs with different energy levels, for the synergy, has been actively used. We have established the energy diagram of the heterosystem NiMn2O4/TiO2/benzamide solution on the basis of the physicochemical characterizations (Fig. 6). Generally, the radicals are formed on wide-band-gap semiconductors illuminated by UV light [30, 31]. The injection of charge carriers occurs isoenergetically and the large difference between NiMn2O4–CB of O2/O2• level makes the electronic transfer weak, leading to a slow photocatalytic kinetics. So, TiO2 is used as electrons bridge in order to mediate the electron transfer in solution.

Fig. 6
figure 6

The energy-band diagram of the heterosystem NiMn2O4/TiO2

Full size image

As mentioned above, NiMn2O4–CB is made up of eg orbital and is pH insensitive, whereas the electronic bands of TiO2 change by −0.06 V pH−1, this property has been exploited to have an optimal band bending at the solid interface NiMn2O4/TiO2 and this occurs around pH ~7. On the other hand, the photoactivity is dependent on the morphology of the catalyst and the effect of decreasing the crystallite size on benzamide oxidation is studied. The lifetime of the charge carriers must be long enough to reach the interface; nanosized dimension is desirable in such a case and the sol–gel is appropriate for preparing powders with increased surface-to-volume ratio. Moreover, the porosity of NiMn2O4 reduces the overpotential of the electrochemical reactions and increases the number of the photocatalytic sites.

The dark adsorption is a preamble for the photocatalysis of both organic and inorganic compounds. The isoelectric point (IP) of TiO2 on which the adsorption occurs is obtained by the simple technique of powder addition [32], IP is found to be ~6.5, and the surface is positively charged at neutral pH. The resonance structure of benzamide occurs between the double bond and the lone pair of nitrogen and is likely responsible for binding which favors its access to catalytic sites of TiO2 by electrostatic attraction.

Benzamide is not converted by photolysis, and ~5% is adsorbed as shown by measurement of the concentration before and after keeping the powder overnight in benzamide solution. Therefore, the decrease of the concentration is mainly attributed to the photocatalytic process; the reaction mechanism currently adopted under irradiation is the following [33]:

$${\mathrm{O}}_{2\,\mathrm{ads}} + {\mathrm{e}}^- \to {\mathrm{O}}_2^\bullet + {\mathrm{H}}^+ \to {\mathrm{HO}}_2^\bullet \to {\mathrm{H}}_2{\mathrm{O}}_2\left(+{\mathrm{e}}^-\right) \to {\mathrm{O}}_2 + {\mathrm{OH}}^\bullet + {\mathrm{OH}}^ -$$
(6)
$${\mathrm{O}}_2^\bullet + {\mathrm{e}}^- \to {\mathrm{O}}_2^{\bullet-}.$$
(7)

The illumination time is fixed at 2 h and the principal parameters influencing the photoactivity are the catalyst dose, pH, and benzamide concentration. As expected, the performance increases with increasing the amount of the sensitizer NiMn2O4 (Fig. 7) due to the large reception surface; this occurs because of the high number of PEC sites for the visible photons and in this way the generated (e−/h+) pairs. The first parameter was the spinel dose Y% (=x/(x + 125) × 100) which varies in the range (0–100%) while maintaining the amount of TiO2 constant (125 mg). The optimal dose (Y%NiMn2O4/TiO2) under artificial light is found to be 60% for benzamide degradation (5 ppm) with a half-life of ~100 min. The regression in the activity above the threshold dose is due to the light obstruction of the catalyst powder and the shadowing effect. The normalized benzamide concentration vs. irradiation time is shown in Fig. 8; the linear dependence of the photocatalytic degradation indicates a first-order kinetics:

$$\mathrm {Ln}\,C_{\mathrm {t}}{\mathrm{ = - }}kt + \mathrm {ln}\,C_{\mathrm {o}}.$$
(8)
Fig. 7
figure 7

The first-order kinetics model of NiMn2O4/TiO2 heterosystem for benzamide degradation at different doses

Full size image
Fig. 8
figure 8

The effect of the initial benzamide concentration toward the photoefficiency of NiMn2O4/TiO2 heterosystem

Full size image

The half-life (t1/2), the time needed to oxidize half of benzamide present initially, is found to be 2 h. In addition, the kinetics shows an initial period of relatively rapid degradation. Over irradiation time, the slope decreases progressively, followed by gradual cessation. This tendency to saturation indicates that the layers already adsorbed are first oxidized after which the kinetics becomes governed by the diffusion of benzamide toward the active sites at the interface in which the radicals O2• are generated for further adsorption/photodegradation. This process is self-limited due to the adsorbed layer and the availability of photocatalytic sites. Such result implies an efficient contact by collision which facilitates the electron transfer between NiMn2O4 and TiO2.

The photocatalytic performance under solar irradiation is strongly enhanced compared to artificial light (Fig. 9). Indeed, in addition to NiFe2O4, TiO2 is activated under solar light which accounts for ~5% of UV light and the electron concentration is increased, leading to improved photoactivity up to 94%.

Fig. 9
figure 9

Effect of solar light toward the efficiency of NiMn2O4/TiO2 heterosystem. Experimental conditions: catalyst, NiMn2O4 (60%)/TiO2; [benzamide]O = 15 mg L−1; Vsolution = 50 mL; pH ∼7

Full size image

4 Conclusion

Benzamide is oxidized on the NiMn2O4/TiO2 heterosystem under visible illumination. The nanosized catalyst is elaborated by sol–gel. The XRD exhibits narrow peaks and the oxide is well crystallized. The capacitance measurement (C−2−E) of NiMn2O4 indicates p-type conductivity. The band gap and the flat-band potential permit to build the energy-band diagram which shows the electron transfer from the conduction to dissolved oxygen via TiO2 and the degradation is considerably increased. The rate of benzamide degradation is monitored by HPLC and the spinel dose and benzamide concentration are optimized. The photoactivity is enhanced under solar light because of the activation of TiO2. A conversion of 94% is reported in aerated benzamide solution (15 ppm) in ~1 h under sunlight and the oxidation follows a first-order kinetics with a half-life of 53 min.