Enhancement of photocatalytic degradation of MB by recyclable Li/Mn3O4 thin films

Thin films of physical mixture of Hausmannite Mn3O4 and lithium (Li) are synthesized by spray pyrolysis technique. Structural, morphological, optical, electrical, wettability, and photocatalytic properties have been investigated. X-ray diffraction (XRD) and Raman spectra, Scanning electron microscope (SEM) images, and electrical measurements show that Li nanoparticles are formed both on top surface of the film and inside grain boundaries. Bandgap and Urbach energies and optical relaxation time have been determined from transmittance T and reflectance R spectra. Impedance spectroscopy shows that charge separation increases with Li content, which improves photocatalytic efficiency of the film. The best photocatalytic efficiency is obtained for Li/Mn ratio of 15%. Indeed, the degradation of methylene blue (MB) under ultraviolet (UV) and visible light exposure, is improved by a factor of 5.7 and 2.4, respectively, when compared to undoped Mn3O4. In addition, this film exhibits a high photostability (10 cycles consecutively) under solar light. On the other hand, hydrophobicity reveals the hydrophilic character of the films.


Introduction
During the past decade, manganese oxide Mn 3 O 4 has attracted researchers' curiosity due to its magnetic behavior, which is sensitive to crystal size at low temperature [1]. It is also used as anode materials for lithium-ion batteries [2][3][4] and supercapacitor applications [5,6]. Recently, Mn 3 O 4 nanoparticles show moderate antibacterial activity against bacteria V [7]. On the other hand, Mn 3 O 4 is a stable to corrosion, non-toxic, and has good environmental compatibility, so it is a good candidate for sustainable energy devices.
Many techniques have been used for the synthesis of Mn 3 O 4 thin films, including chemical bath deposition technique, pulsed laser ablation, and SILAR (Successive Ionic Layer Adsorption and Reaction) method [8][9][10]. In this work, we have prepared Mn 3 O 4 thin films, by the spray pyrolysis technique, which is a simple, effective, and cheap method with controllable reaction and large-scale production of samples. Hausmannite Mn 3 O 4 is a multivalent oxide that crystallizes in a spinel tetragonal structure with manganese ions (Mn 3? ) in the octahedral sites and (Mn 2? ) in the tetrahedral ones, respectively. This duality among sites is the origin of its important characteristics, which can be improved through metal ion doping. For instance, copper-doped Mn 3 O 4 microcrystal is an excellent reagent for the decomposition of organic contaminants in water [11]. Furthermore, Zhao et al. synthesized Sn-doped Mn 3 O 4 / C nanocomposites for supercapacitor electrode with higher capacitance [12]. Moreover, Li et al. [13] showed that Cr doping could enhance ferromagnetic properties of Mn 3 O 4 nanowires. This material is a good candidate for photodegradation of organic dyes, since ferromagnetic materials such as Fe 3 O 4 are confirmed as good photocatalysts. [14] Likewise, previous work have demonstrated that the effect of Cr doping on physical properties of Mn 3 O 4 thin films is very promising in the field of optoelectronics, photothermal conversion, and chemical sensing [15]. In this context, Li is an efficient electron donor owing to its electronic states. It is also an appropriate doping element to improve the physical properties of many transition metal oxides. For example, Li et al. [16] demonstrated that electrochemical performance and discharge capacity of NiO films are boosted by Li doping. Moreover, numerical simulation and experiments show that stochastic nonlinear dynamic characteristics and safe basin of Li-doped graphene are helpful for hydrogen storage field [17]. Moreover, previous works have demonstrated that addition of Li and Zr to Mn 3 O 4 thin films leads to an enhanced response toward low ethanol concentration [18,19]. Mn 3 O 4 has a large potential in heterogeneous catalysis, in particular advanced oxidation techniques of H 2 O 2 and degradation of dyes [20][21][22]. This photocatalytic activity can be improved by metal doping which increases free-carrier lifetime. Recently, Kanwal et al. showed that Mn 3 O 4 /Al 2 O 3 heterostructure exhibits a bandgap narrowing and enhancement of charge separation due to a strong interfacial contact [23]. The same behavior was obtained by Jamila et al. for N-and Bidoped carbon quantum dots on Mn3O4. They found that N and Bi doping induces a photoluminescence quenching leading to enhancement of photocatalytic activity [24].
In this paper, we have studied the effect of Li addition on the structural, electrical, and optical properties of the Mn 3 O 4 thin film. These properties are related with catalytic activity of the Mn 3 O 4 thin film to enhance its photocatalytic performance in the visible region. The results of this study may be of interest for applications in photocatalysis and optoelectronic domains. Finally, recovering the photocatalyst after degradation of organic compounds remains a major obstacle to this technique. In order to avoid this problem, photocatalyst films are fixed on glass substrate.

Characterization techniques
Firstly, structure and morphology of all samples are examined using XRD (X-Ray Diffraction) (''XPERT-PRO'' diffractometer with Cu Ka radiation at a wavelength k = 1.542 Å ), scanning electron microscope SEM (JEOL-JSM 5400), and Raman spectroscopy (Jobin Yvon Horibra LABRAM-HR with the excitement source He-Ne 632 nm laser). Then, electrical conductivity was determined by an impedance meter (Hewlett Packard 4192 analyzer) at a temperature of 250°C in the frequency range 1 kHz-12.6 MHz. Thereafter, contact angle is measured by the Drop Shape Analyzer DSA100 system from KRÜ SS. The deposit of the droplet of volume 5 l' and the recording of the contact angle are piloted automatically by a specific software. Appropriate model is used to calculate surface energy from the measurement of contact angle between distilled water and thin film [25]. Finally, optical transmittance T, where C 0 , A 0 and C, A are the concentration and absorbance of MB when the reaction time is 0 and t, respectively.  [28]. This result is in good agreement with previous works [29]. Doping of Mn 3 O 4 at low Li rate (3 and 6%) improves the sample crystallinity leading to an enhancement of XRD intensity. For the sample doped at 6%, the main peak at 2h = 18°corresponding to (101) preferential orientation exhibits the maximum intensity. At 9% of Li rate, XRD peaks are reduced due to partial destruction of crystalline structure and increased surface roughness.
Indeed, this sample presents the maximum values of microstrain and dislocations density. At 15%, peak intensity increases owing to lower dislocations and microstrain as well as improved film crystallinity, since crystallite size has clearly increased ( Table 1).
Positions and profiles of XRD peaks can be used to calculate lattice constants (a and c), grain size (D), microstrain (e), and dislocation density (d) which is related to optical and photocatalytic properties. These parameters are calculated using the following equations [30]: where h (in radian) is the Bragg diffraction angle, b (in radian) is the full width at half maximum (FWHM) of the most intense peak and k is the X-ray wavelength. On the other hand, lattice constants (a and c) for tetragonal structure are determined by where (h,k,l) are Miller indices and d hkl is the interplanar distance. In the same context, absolute variations of cell volume and crystallite size are given by  These results are in good agreement with those obtained by Ben Said et al. [22] which proved that Zr is partially incorporated into the Mn 3 O 4 nuclei, it grows as nanoparticles at the film surface and in the grain boundaries. Similar results have been also reported in the literature, concerning the oxidation of zinc in presence of noble metals (Ag, Pt) which leads to pure metal nanoparticles at the surface of ZnO film [31][32][33]. On the other hand, main XRD peaks are used to determine crystallite size, microstrain (e), and dislocation density (d). These parameters are obviously affected by added Lithium. Raman spectra of Li/Mn 3 O 4 thin films shown in Fig. 2 are composed of two small peaks at 315, 370, and a sharp peak at 655 cm -1 . The main peak at 655 cm -1 is owed to Mn-O vibration mode (A 1g ) [24]. Raman peak positions and bandwidths are similar to those of pure Mn 3 O 4 [34]. We conclude that Mn 3 O 4 crystalline structure is not affected by Lithium.
Raman peaks intensity decreases when Li doping rate increases from 0 to 6%, then it increases from 6 to 15%. The main reason of this behavior is the variation in absorbance of the incident laser beam due to bandgap shifts. These variations will be discussed in the following Sect. 3.2.
The variation of intensity can also be related to the variation of crystallite size and dislocation density. From 0 to 6% of Li, the crystallite size is reduced and dislocations are enhanced which causes a loss of light by diffusion and reduces Raman scattering effect. This behavior is reversed when Li rate increases from 6 to 15%.
SEM images (Fig. 3a-e) confirm the results obtained by XRD and Raman spectra. We observe  Table 2 shows the RMS roughness of Li/Mn 3 O 4 physical mixtures. The surface roughness increases by increasing Li rate up to 9%, then it decreases slightly at 15% which corresponds to largest crystallite size and best crystallinity. These factors are of most importance in photocatalytic activity.

Optical study
Transmittance T(k) and reflection R(k) spectra of Li/ Mn 3 O 4 thin films are shown in Fig. 4a and Absorbance in Fig. 4b. We find that optical transmittance increases with Li content in the visible and near-infrared region. Therefore, penetration depth of light into the films is increased leading to excitation of electrons and holes inside the film volume. Similar result has been reported in the literature for Li-doped ZnO thin films [35]. In addition, Li doping induces absorption edge shift toward low wavelengths. Optical absorption spectra a(E) (Fig. 4c) obtained from absorbance spectra (Fig. 4b) are used to determine the band gap energy E g following the Tauc law [17]: We have also determined Urbach energy E u from the equation [36]: where a is the absorption coefficient and E ¼ hm the photon energy of incident light. The obtained energy values of E g and E u are given in Table 2. We notice that band gap energy increases from 2.23 to 2.82 eV as Li ratio increases from 3 to 6%, then decreases slightly to 2.67 eV for the film at 15%. These bandgap shifts may explain the variations of Raman peaks intensity. Indeed, the blue shift obtained from 3 to 6% of Li lead to a drop of the optical absorbance (Fig. 4b), especially for the wavelelength of the Raman laser (632 nm), which leads to a reduction of the Raman effect. From 6 to 15% of Li, the Raman peaks are enhanced as the optical absorption at the laser wavelength is improved.
The band gap energy depends on several factors such as effective mass of charge carriers and Bohr radius. For a semiconductor crystal, the exciton is usually delocalized over a distance much longer than the lattice constant called Bohr radius. When the dimension of the crystallite becomes near to Bohr radius, the exciton is confined and consequently its electronic properties start to change [37,38].
Same results have been reported in previous works on ZnO and WO 3 doped with Lithium [39,40] where localized states appear in the band gap due to Liinduced defects. The analysis of Urbach energy shows that the width of localized states increases with Li content. Indeed, dislocation density, microstrain (Table 1) and roughness ( Table 2) are markedly enhanced by Li doping from 3 to 9%. At 15% of Li, the film exhibits a high value of Urbach energy despite the decrease of dislocations density. This is probably due to other defects induced by Li cations, such as oxygen vacancies and interstitial Mn. The broadening of band gap tails is of a great benefit for photocatalytic activity since the charge hopping between these localized states is improved [41].
On the other hand, dielectric calculations are used to determine optical relaxation time s op which correspond to the lifetime of photo-excited carriers. In the near-infrared region, optical relaxation time s op is related to real and imaginary parts (e 1 and e 2 ) of the complex dielectric constant e via the relation [42,43].
where e 1 is the dielectric constant at high frequencies, k is the wavelength, and c is the light celerity.
The calculated values of relaxation time are shown in Table 2. We notice that pure Mn 3 O 4 film has the smallest value of relaxation time that increases with Li doping. The maximum value is obtained for 15% of Li rate, since the film exhibits highest crystallite size, best crystallinity, and minimum rate of defects.

Photoluminescence
Photoluminescence spectroscopy (PL) is used to study the radiative recombination mechanisms of Lidoped Mn 3 O 4 films when irradiated by 488 nm laser beam. For all samples, PL spectra (Fig. 5) show a broad and intense peak centered at 694 nm neighbored by a small peak at 640 nm. The broadness of main peak is due to many defects caused by dislocation and fractal morphology of the nanostructured surface resulting from deposition technique and Li films. These peaks are originated from the optical transitions due to structural disorder-induced localized states in the bandgap [44]. We conclude that radiative transitions occur between deep levels induced by defects and the valence (or conduction) band [45].
However, increasing Li rate reduces noticeably the PL main peak. Indeed, the growth of Mn 3 O 4 small grains around Li nanoparticles at the film surface and the raise of Li-induced defects provoke the inhibition of radiative recombinations of electron-hole pairs through the hoping process. For high Li doping rate

Impedance spectroscopy
Impedance spectroscopy is a reliable technique to study the electrical properties of a material and to disclose the efficiency of charges carrier trapping. Thus, this technique is carried out to evaluate the chemical interaction between Mn 3 O 4 thin film and Li nanoparticles by electrical conduction. Figure 6 shows the complex impedance spectra obtained by plotting the imaginary part (-Z 00 ) of the impedance versus corresponding real part (Z 0 ). The intercepts of semi-circular impedance curves with real axis (Z 0 ) correspond to the resistance values. The decrease of the curvature radius while increasing Li content implies a decrease of electrical resistance R. Indeed, adding metallic atoms such as Li leads to great amount of free carriers injected in the semiconductor material which enhances its electrical conductivity. This result is in agreement with previous works [35,46] which proved that Li doping improves electrical conductivity and electron mobility for ZnO thin films. Similar results on Mn 3 O 4 -MnO 2 composite had been reported by Zhao et al. [7] who found that the sample giving smallest radius of impedance curve yields the highest photocurrent. The analysis of experimental data of all the materials shows that the semi-circles are slightly sunken. This effect is well described by the Cole-Cole model in which a polarization phenomenon with a distribution of relaxation time occurs in the material. According to this model, the complex impedance is given by the equation [36]: where s el = RC is the mean value of relaxation time distribution, x is the angular frequency, and a is a parameter which characterizes the relaxation time distribution. In order to study relaxation time, electrical conductivity, and capacity of Li/Mn 3 O 4 thin films, we have plotted (-Z 00 ) versus frequency in Fig. 7. The elements values of the equivalent electric circuit are ported on Table 3. The maximum of imaginary part (-Z 00 ) shifts toward low frequencies as the Li content increases suggesting an increase of relaxation time averagevalue and a longer mean free path of charge carriers. Meanwhile, film capacitance C is found to increase with Li content which reveals an enhancement of charge separation in Li/Mn 3 O 4 films. Indeed, a schottky junction appears between Li grains and the p-type Mn 3 O 4 film [47]. Charge separation provokes the enhancement of carrier lifetime [48] which is the determining factor in photocatalysis process [26,[49][50][51][52][53].    [54][55][56] which showed that high surface roughness can improve the hydrophilicity as well as the surface energy leading to a better catalytic performance. The formation of Li nanoparticles increases the specific surface area, which enhances the adsorption of pollutant at the film surface [57,58]. Surface energy calculated from measured values of contact angle is found to be practically constant for different Li content. However, it is substantially higher for Li-doped films compared to that of pure Mn 3 O 4 thin film.

Photocatalytic activity
3.6.1 UV irradiation effect Figure 8 shows the absorption spectra of MB solution in presence of Li/Mn 3 O 4 thin film for 2 h under UV light, at room temperature. Under the same conditions, a blank test is realized without film, in order to evaluate the self-degradation of MB, which is evaluated to about 9%. However, catalytic decomposition of MB in presence of Li/Mn 3 O 4 thin films increases with Li content. In the case of 15% ratio, we notice a degradation efficiency of about 5.7 times higher than that of pure Mn 3 O 4 . Figure 9 shows a satisfactory correlation between photocatalytic efficiency, electrical resistance, and electric capacitance. It is found that photocatalytic activity is improved while increasing the film capacitance. However, it is reduced by the rise of electrical resistance. These results are in good agreement with previous work [35] on Li-doped ZnO thin films. After 2 h of exposure time, we obtain the elimination of 46% of MB molecules which is of the same order with the results obtained with Ag/ZnO nanoparticles (50%) [31].
On the other hand, catalytic activity is dependent on the size and distribution of Li nanoparticles. It can be suggested that particle with small size allows a better dispersion and, thus, a large number of active sites, thus improving the catalytic activity. This result is consistent with previous works [31,52,57,59]. Figure 10 shows the absorption spectra of MB alone and in presence of Li/Mn 3 O 4 thin films under visible light. At first, we notice a same rate for the degradation of MB in the presence of pure Mn 3 O 4 or without catalyst. Then, degradation activity is improved by Li doping and the efficiency is extended from UV to visible region. Indeed, under visible light, the pure Mn 3 O 4 catalyst exhibits no effect on MB degradation; however, a noticeable effect is seen with  [27,31,32,35] on photocatalytic behavior of metaldoped ZnO films. The variation of photocatalytic efficiency and optical relaxation time with Li content is shown in Fig. 11. The photocatalytic efficiency varies proportionally with the variation of optical relaxation time. This may be explained by the increase of photogenerated charges lifetime, which allows electrons and holes to react with water and oxygen before recombination. This result is confirmed by the degradation of visible photoluminescence seen above, especially for the sample with 15% Li rate, which proves that carrier lifetime is enhanced while radiative recombinations are reduced by hopping effect.

Visible light effect
Moreover, photocatalytic reaction is boosted by the great number of free electrons released by metallic Li atoms adjacent to Mn 3 O 4 nanoparticles.
The degradation results analysis of MB under UV and visible light, and those of impedance spectroscopy experiments and optical absorption, reveals reliable information on the degradation mechanism of MB. The almost zero role of pure Mn 3 O 4 to the degradation of MB allows deducing that the recombination of photogenerated electron-hole pairs in pure Mn 3 O 4 is very fast which inhibits the degradation of MB pollutant. The Li addition to the Mn 3 O 4 helps to reduce the electrical resistance and increases  After light absorption, the photogenerated electrons can transfer from Mn 3 O 4 to Li-induced traps such as dislocations and oxygen vacancies, thus improving charge separation and carrier lifetime, which instigate the formation of hydroxyl radicals essential for MB degradation. Photocatalytic process suggested above is in agreement with previous works [31,32,60] which showed that noble metal particles help the transmission of photoelectrons at the film surface and can be the instigator of photogeneration process. Figure 12 displays the evolution of MB degradation in the presence of Li/Mn 3 O 4 at 15% as a function of exposure time to sunlight irradiation. For these experiments, 20 mL of MB solution is used, which corresponds to 3.33 times as the volume is exposed to UV and visible light source. The drop of absorption peak to baseline level showed that degradation of sub-products and the scission of bonds are perpetuated up to mineralization.

Solar energy recyclability
After 4 h of exposure time, we obtain the elimination of 91% of MB molecules which is of the same order with the results obtained with nitrogen-doped carbon quantum dots (NCQD) on Mn 3 O 4 ( 73-100%) [24].
After irradiation for 6 h, we notice that about 2% of MB remains in aqueous solution, so the photodegradation of MB is practically complete.
Photocatalytic degradation of MB follows the pseudo-first-order kinetics: where C0 and C are the MB concentration at t = 0 and t, respectively. The degradation rate k for each sample is obtained from a linear fit of the curve ln(C0/C) versus time given in Fig. 13. This curve shows that photodegradation of MB molecules follows first-order reaction. The first-order degradation rate constants for Lidoped Mn3O4 samples are found to increase with Li rate and the optimum value k = 7.6 10 -3 min -1 is obtained for 15% Li. This value is far from the rate obtained without catalyst which is below 0.9 10 -3 min -1 .
The stability of Li/Mn 3 O 4 at 15% of Li is investigated by comparing the MB absorbance with the same catalyst for the first degradation cycle and after 10 repetitive cycles. As shown in Fig. 14 [61,62]. On the other hand, photocatalytic efficiency can be attributed to the small size of Li nanoparticles, which improves oxygen adsorption at the catalyst surface [31,63]. Finally, considering its

Conclusion
Li-doped Mn3O4 have been successively deposited onto glass substrate by versatile spray pyrolysis technique. The influence of Li addition on structural, morphological, optical, electrical, and photocatalytic properties is investigated. XRD, Raman, and electrical analyses reveal the existence of Li nanoparticles at the surface and in the grain boundaries of Mn 3 O 4 thin films. Their size and distribution are determined by SEM images. Capacitance, electrical conductivity, and optical lifetime are enhanced by Li doping. The increase of photo-carrier lifetime is confirmed by the drop of radiative recombination as photoluminescence peak is drastically reduced while increasing Li doping rate.
Wettability proved that liquid adsorption at the film surface is improved by Li nanoparticles. Moreover, photocatalytic activity of catalyst Mn 3 O 4 is enhanced by increasing Li content in the film. Photocatalytic efficiency of thin films is attributed to long lifetime of photogenerated electron-hole pairs. For this purpose, the film should have low defect density, high crystallinity, and spatial separation of charge carriers. The physical mixture Li/Mn 3 O 4 at 15% of Li rate exhibited the highest photocatalytic efficiency for MB decomposition, 5.7 and 2.4 times higher than that obtained with pure Mn 3 O 4 catalyst under UV and visible lights, respectively. It has been deduced that Li doping broadens the activity spectrum of Mn 3 O 4 catalyst from UV to visible region, which allows a better exploitation of the free solar energy. Li/Mn 3 O 4 with optimal rate of 15% exhibits the highest photostability. Indeed, this film exhibits the same photocatalytic efficiency after 10 consecutive photocatalytic cycles. We conclude that it can be used as immobilized photocatalyst for water purification from organic pollutants and self-cleaning applications under sunlight.

Funding
None.

Declarations
Conflict of interest All authors declare that they have no conflict of interest.