Effect of Mg Doping on the Physical Properties of Fe2O3 Thin Films for Photocatalytic Devices

Undoped and Mg-doped (y = [Mg2+]/[Fe3+] = 1, 2, 3, and 4 at.%) Fe2O3 thin films were synthesized by a simple spray pyrolysis technique. The thin films were extensively characterized. X-ray diffraction (XRD) and energy-dispersive spectroscopy (EDS) analysis confirmed the successful insertion of Mg in the rhombohedral structure of Fe2O3. In addition, scanning electronic microscope (SEM) and confocal microscope (CM) images showed a homogenous texture of the film, which was free of defects. The rough surface of the film obtained by spray pyrolysis is an important feature for photocatalysis and gas sensor applications. The direct band gap of the doped Fe2O3 films obtained for [Mg2+]/[Fe3+] = 3 at.% was Edir = 2.20 eV, which recommends the Mg-doped iron oxide as an optical window or buffer layer in solar cell devices. The photodegradation performance of Mg-doped Fe2O3 was assessed by studying the removal of methylene blue (MB) under sunlight irradiation, with an effective removal efficiency of 90% within 180 min. The excellent photodegradation activity was attributed to the strong absorption of Mg-doped Fe2O3 in the UV and most of the visible light, and to the effective separation of photogenerated charge carriers.


Introduction
Maintaining a clean environment has become an inspiring research ambition in the environmental-science-related communities. As a global environmental issue, water pollution threats our life. For example, eco-clean technology is urgently needed to deal with dyes that are manufactured, used, and thrown into the water. Photocatalysis draws attention to the need for extensive research and investigations into water purification and generation of renewable energies [1][2][3][4]. The oxide semiconductor photocatalyst has received special attention as a beneficial material with easy-to-control properties. The basic idea of photocatalysis is the activation of the semiconductor through an artificial or a natural source of light, which breaks down the organic compounds and dyes to purify water [3]. Ferric oxide (Fe 2 O 3 ), as an environmental green oxide with outstanding physical and chemical properties, has been the subject of a great deal of research [5,6]. Fe 2 O 3 , as a mid-band gap semiconductor, utilizes sunlight effectively in the photocatalysis process as compared to TiO 2 , which is the most researched photocatalyst [7][8][9][10]. Numerous methods were adopted to enhance the photocatalytic activity in metal oxide semiconductors, such as the formation of junctions and doping to prevent the electron-hole (e − /h + ) recombination and to enhance the photocatalysis process, where dopant ions act as charge traps, reducing the recombination of e − /h + [11][12][13]. The photocatalytic degradation of rhodamine B (RhB) with sprayed Fe 2 O 3 films under sunlight illumination was investigated. The efficiency of RhB decomposition was observed to be 73% during 400 min [14]. Jiamprasertboon et al. prepared an α-/Fe 2 O 3 /ZnO and ZnO/α-Fe 2 O 3 heterojunction using aerosol-assisted chemical vapor deposition (AACVD). The α-Fe2O3/ZnO exhibited the highest photocatalytic activity under UVA light, which was approximately 16 and 2.5 times higher than that of the Fe2O3 and ZnO layers. In contrast, the reverse heterojunction architecture was less active [15]. Furthermore, Fe 2 O 3 and Zn:Fe 2 O 3 nanoparticles were prepared via sol-gel with different Zn ratios. The photodegradation analysis of 4 at.% Zn:Fe 2 O 3 showed 87% of RB dye degraded in 90 min in the UV light compared to 63% with pure Fe 2 O 3 , and a further increase in Zn content decreased the degradation efficiency [16]. Gajendra K. Pradhan [19].
Undoped and doped Fe 2 O 3 thin films are commonly fabricated by various chemical and physical techniques, such as sol-gel [20], chemical bath deposition [21], chemical spray pyrolysis [22], electrodeposition [23], and SILAR [24]. In addition, low-cost thin films can be deposited on different substrates by chemical spray pyrolysis for various industrial applications [25,26]. In this paper, the structural, morphological, elemental, and optical properties of Mg-doped Fe 2 O 3 obtained by chemical spray pyrolysis technique are presented. The photocatalytic activity of Mg-doped Fe 2 O 3 on the photodegradation of methylene blue under sunlight irradiation was investigated, which, to our knowledge, has not yet been studied in the literature.

Preparation of Undoped and Mg-Doped Fe 2 O 3 Thin Films
Undoped and Mg-doped Fe 2 O 3 thin films were grown on ordinary glass substrates through chemical spray pyrolysis technique (CSP). Before the deposition process, all the glass substrates were carefully cleaned via immersion in an ultrasonic bath. Iron III chloride (FeCl 3 and 6H 2 O) and (MgCl 2 and 6H 2 O) precursors were acquired from AppliChem (Council Bluffs, IA, USA). Both precursors was dissolved in 100 mL of bi-distilled water, where MgCl 2 was added as a dopant. The Mg ratio was adjusted at y = [Mg 2+ ]/[Fe 3+ ] = 1, 2, 3, and 4 at.%. On the other hand, the iron chloride concentration was kept at 0.14 mol·L −1 . The as-prepared solutions were mixed until obtaining homogenous mixtures. The aqueous solution was sprayed with a flow rate equal to 5 mL·min −1 by means of compressed air on preheated substrates located 25 cm from the nozzle at 400 • C for 20 min.

Photocatalytic Activity
The photocatalytic activity was assessed by evaluating the degradation of methylene blue (MB) solution under sunlight illumination. The Mg-doped Fe 2 O 3 film was immersed in 50 mL aqueous MB solution with a concentration of 5 mg/mL and kept under sunlight irradiation for various times (0-180 min). The photodegradation of MB was then estimated by the maximum absorbance at a wave length of 664 nm using Vis-spectrophotometer (Perkin Elmer Lambda 950). According to the Beer-Lambert law, the degradation efficiency of methyl blue was computed starting from the absorbance spectra using the following equation [27]: where A 0 and A are the values of MB solution absorbance at reaction times of 0 and t, respectively. The reaction kinetic was estimated using the following formula [27]:

Characterization
All the samples were characterized by several techniques as discussed below. The crystalline quality of Fe 2 O 3 thin layers was characterized by X-ray diffraction (XRD) using X-ray "XPERT-PRO" diffractometer (Malvern Panalytical Ltd., Malvern, UK) with CuKα (λ = 1.54 Å) radiation over a scanning angle (2θ) ranging continuously from 25 • to 60 • . The experimental XRD spectra were refined with the MAUD software. The surface morphologies and cross-section of the thin films were observed using scanning electronic microscope (SEM) "ZEISS" (Carl Zeiss Microscopy, New York, USA) in surface (1 µm (EHT = 3.00 kV; Mag = 20.00 K) and 200 nm (EHT = 3.00 kV; Mag = 100.00 K)) and cross-section. In addition, 3D characterization was performed using a confocal microscope (CM) (SENSOFAR, Barcelona, Spain). The elemental composition was analyzed using energy-dispersive X-ray spectroscopy (EDS) "ZEISS". The optical measurements were carried out using a spectrophotometer, Perkin Elmer Lambda 950 (Bridgeport, CT, USA), over the wavelength range of 250-2000 nm.

Structural Properties
The influence of Mg-doping on the crystalline structure of the as-synthesized iron oxide composites was examined by X-ray diffraction. As shown in Figure 1 [28,29]. No new peak formation was observed, which confirms that magnesium has been successfully substituted into the Fe 2 O 3 matrix. The X-ray analysis shows that all films are polycrystalline and have a preferential orientation along the (104) plane, regardless of the Mg doping level. A slight shift to higher diffraction angles was detected in the XRD scans of the doped samples, especially for 1 at.% and 2 at.%. The shift was attributed to the seamless incorporation of magnesium ions into the Fe 2 O 3 structure, due to the smaller ionic radius of Mg 2+ compared to Fe 3+ [30,31]. We note that the highest intensity of the principal orientation was obtained at 3 at.% Mg doping concentration. The preferred orientation degree (Tc) in the (hkl) orientation was determined using the following empirical relation [27]: where (hkl) are the Miller indices, I0 is the standard intensity, I is the measured intensity, and N is the number of reflection peaks. The variation of the Tc (104) and Tc (110) values with the doping concentration of Mg is represented in Figure 2, which shows that the tendency of crystallites to develop along the (110) plane increases with the increase of the Mg doping concentration up to 3 at.%. Although (104) is the major orientation, the variation of the intensity of the (110) plane, which presents a random orientation of crystallite, confirms the polycrystalline character of the Mg-doped thin films. The preferred orientation degree (Tc) in the (hkl) orientation was determined using the following empirical relation [27]: where (hkl) are the Miller indices, I 0 is the standard intensity, I is the measured intensity, and N is the number of reflection peaks. The variation of the Tc (104) and Tc (110) values with the doping concentration of Mg is represented in Figure 2, which shows that the tendency of crystallites to develop along the (110) plane increases with the increase of the Mg doping concentration up to 3 at.%. Although (104) is the major orientation, the variation of the intensity of the (110) plane, which presents a random orientation of crystallite, confirms the polycrystalline character of the Mg-doped thin films.
The lattice constants a and c of the nanocrystals and the cell volume of all the samples were estimated according to the following equations [7]: where d hkl is the interplanar distance. The estimated values of lattice parameter are presented in Table 1. The calculated values of the unit cell parameters of Mg-doped Fe 2 O 3 were found to be lower than those of the undoped films, which result in a decrease in the cell volume. Magnesium incorporation caused local changes in the Fe 2 O 3 matrix, which confirms the successful fabrication of Mg-doped Fe 2 O 3 thin film [32][33][34]. For further investigation of the impact of Mg doping on the film microstructure, the average crystallite size and macrostrain of the films were estimated using Williamson-Hall formula [22]: where β is FWHM in radian, D is the grain size in nanometers, ε is the microstrain, and λ is X-ray wavelength in nanometers. The values estimated from Figure 3 are presented in Table 1, which shows that D decreased from 60.8 nm to 55.3 nm regardless of the doping concentration. Consequently, the microstrain increased from 0.31 × 10 −3 to 0.64 × 10 −3 . The decrease in the lattice parameters (a,c) and cell volume (V) presented in Table 1 is expected due to the stoechiometric replacement of Fe ions with Mg ions with smaller ionic radii. The smallest volume, V, was obtained for y = 3 at.%, corresponding to the best crystalline quality (Figures 1 and 2), which is also in agreement with the substitution of Fe by Mg in the lattice [33]. The lattice constants a and c of the nanocrystals and the cell volume of all the samples were estimated according to the following equations [7]: where dhkl is the interplanar distance. The estimated values of lattice parameter are presented in Table 1. The calculated values of the unit cell parameters of Mg-doped Fe2O3 were found to be lower than those of the undoped films, which result in a decrease in the cell volume. Magnesium incorporation caused local changes in the Fe2O3 matrix, which confirms the successful fabrication of Mg-doped Fe2O3 thin film [32][33][34]. For further investigation of the impact of Mg doping on the film microstructure, the average crystallite size and macrostrain of the films were estimated using Williamson-Hall formula [22]: βcosθ 1 sinθ

Rietveld Analysis
Rietveld analysis (using MAUD software) was used for the 3 at.% Mg-doped Fe2O3 thin films to check the α-Fe2O3 phase purity. The refinement plot is shown in Figure 4, which represents the experimental pattern as black dots and the pattern calculated with

Rietveld Analysis
Rietveld analysis (using MAUD software) was used for the 3 at.% Mg-doped Fe 2 O 3 thin films to check the α-Fe 2 O 3 phase purity. The refinement plot is shown in Figure 4, which represents the experimental pattern as black dots and the pattern calculated with Rietveld refinement as red solid lines. In addition, the lower part of the graph indicates the difference between the values of experimental and calculated intensities. The structural Nanomaterials 2022, 12, 1179 7 of 14 fitting quality was checked by the goodness-of-fit factor (GoF = Rwp/Rexp), where Rwp is the weighted residual error and Rexp is the expected error. The GoF was found to be 1.26, which describes a well-fitting model with low discrepancies between the experimental and calculated XRD patterns [7,35]. The results obtained confirm that the Fe ion was substituted by the Mg ion in the Fe 2 O 3 phase.

Rietveld Analysis
Rietveld analysis (using MAUD software) was used for the 3 at.% Mg-doped Fe2O3 thin films to check the α-Fe2O3 phase purity. The refinement plot is shown in Figure 4, which represents the experimental pattern as black dots and the pattern calculated with Rietveld refinement as red solid lines. In addition, the lower part of the graph indicates the difference between the values of experimental and calculated intensities. The structural fitting quality was checked by the goodness-of-fit factor (GoF = Rwp/Rexp), where Rwp is the weighted residual error and Rexp is the expected error. The GoF was found to be 1.26, which describes a well-fitting model with low discrepancies between the experimental and calculated XRD patterns [7,35]. The results obtained confirm that the Fe ion was substituted by the Mg ion in the Fe2O3 phase.

Morphological Properties
The effect of magnesium doping on the morphology of Fe2O3 films is shown in Figure 5a-e. The SEM micrographs show a granular morphology with a uniform distribution through the substrate surface for the undoped and doped samples. A similar morphology was mentioned in literature [5,22,28]. The cross-section images show a

Morphological Properties
The effect of magnesium doping on the morphology of Fe 2 O 3 films is shown in Figure 5a-e. The SEM micrographs show a granular morphology with a uniform distribution through the substrate surface for the undoped and doped samples. A similar morphology was mentioned in literature [5,22,28]. The cross-section images show a compact and homogenous textured film with a thickness of about 384 and 340 nm for undoped Fe 2 O 3 and 3 at.% Mg-doped Fe 2 O 3 , respectively (Figure 5a,d). No defects are observed on the samples surface. The undoped films appear to have a randomly distributed grain agglomeration on the surface of the film. The Mg-doped films show a more uniform distribution with smaller grain size compared to the undoped films, which recommends them for photocatalysis applications [36]. At a higher Mg-doping concentration, the rhombohedral shape of the grains appears more clearly defined (inset Figure 5d,e).
In order to better understand the effect of Mg doping on the morphological properties of the Fe 2 O 3 thin films, the topography evolution of the deposited films was investigated using confocal microscope (CM). Figure 6 presents the 3D CM images of the thin films. Overall, these images are in agreement with those observed by SEM ( Figure 5). The 3D micrographs show a homogenous layer free from voids and cracks. The surface roughness parameters, including the root mean square (Sq) and arithmetic average of absolute values (Sa), were extracted from CM data. The values of surface roughness of undoped and Mg-doped Fe 2 O 3 thin films as a function of Mg concentrations are given at layers have a relatively high surface roughness, which can offer more available active sites and, consequently, improve the adsorption of pollutants during the photocatalysis process [18]. The highest values of Sa and Sq were obtained for a Mg content of 3 at.% ( Table 2). undoped Fe2O3 and 3 at.% Mg-doped Fe2O3, respectively (Figure 5a,d). No defe observed on the samples surface. The undoped films appear to have a ran distributed grain agglomeration on the surface of the film. The Mg-doped films s more uniform distribution with smaller grain size compared to the undoped films, recommends them for photocatalysis applications [36]. At a higher Mg-d concentration, the rhombohedral shape of the grains appears more clearly defined Figure 5d,e).  absolute values (Sa), were extracted from CM data. The values of surface roughness of undoped and Mg-doped Fe2O3 thin films as a function of Mg concentrations are given at Table 2. The Sa and Sq values vary in the range [60.6-76.6] nm and [80.6-104.1] nm, respectively. Mg-doped Fe2O3 layers have a relatively high surface roughness, which can offer more available active sites and, consequently, improve the adsorption of pollutants during the photocatalysis process [18]. The highest values of Sa and Sq were obtained for a Mg content of 3 at.% (Table 2).   The EDS spectra of the glass substrate and 3 at.% Mg-doped Fe 2 O 3 film shown in Figure 7 reveal many peaks. The EDS analysis confirmed the presence of the expected elements, iron (Fe), magnesium (Mg), and oxygen (O), all in addition to those attributed to glass substrate, confirming the grown layer free from impurities. The EDS spectra of the glass substrate and 3 at.% Mg-doped Fe2O3 film shown in Figure 7 reveal many peaks. The EDS analysis confirmed the presence of the expected elements, iron (Fe), magnesium (Mg), and oxygen (O), all in addition to those attributed to glass substrate, confirming the grown layer free from impurities.

Optical Properties
To investigate the doping effect of Mg (y = 1, 2, 3, and 4 at%) on the optical properties of Fe2O3 thin films, transmission (T(λ)) and reflection (R(λ)) measurements were carried out. The T (λ) and R(λ) spectra are presented in Figure 8, which shows that the as-deposited films have a high transmission coefficient (≥60%) within the interval of 1000-2500 nm. Below 650 nm, a sharp fall in all the R-T spectra is observed, which is due to the very strong absorption of these films in the UV and most of the visible light region. Remarkably, in the SEM cross-section images ( Figure 5), interference fringes are

Optical Properties
To investigate the doping effect of Mg (y = 1, 2, 3, and 4 at.%) on the optical properties of Fe 2 O 3 thin films, transmission (T(λ)) and reflection (R(λ)) measurements were carried out. The T (λ) and R(λ) spectra are presented in Figure 8, which shows that the as-deposited films have a high transmission coefficient (≥60%) within the interval of 1000-2500 nm. Below 650 nm, a sharp fall in all the R-T spectra is observed, which is due to the very strong absorption of these films in the UV and most of the visible light region. Remarkably, in the SEM cross-section images ( Figure 5), interference fringes are observed in the transmission spectra, referring to the films excellent thickness homogeneity [22,28]. The optical reflection data allow the estimation of the band gap values through the differential reflectance spectra (dR/dλ) as a function of wavelength. The optical direct (Edir) and indirect (Eind) band gaps values obtained for Fe 2 O 3 and Mg-doped iron oxide semiconductor are typical of those for Fe 2 O 3 (Table 3) [7]. No significant variation was observed for indirect transition (E ind ), but a slight increase of the direct transition value (E dir ) with Mg doping content was observed from 2.15 eV for Fe 2 O 3 to 2.20 eV for 3 at.% Mg-doped Fe 2 O 3 (Table 3). We observed that the films crystallinity improved with the Mg insertion ( Figure 1). In addition, a slight increase in the direct band gap was noticed, to reach a maximum value equal to 0.05 eV with a magnesium doping ratio equal to 3 at.%. This finding is in good correlation with the crystallinity observed at the same doping ratio, which reduced the structural defects.

Photocatalytic Activity
It is well recognized that the photocatalysis process occurs through the photogeneration of electron-hole pairs under light irradiation [2,4]. When photons reach a semi-conductor catalyst, an electron absorbs the photon energy to move from valence band to occupy conduction band levels, leaving behind an electron vacancy (hole); thus, creating the electron-hole pairs. On the other hand, the lifetime of electron-hole pairs is very short [13,15,17]. To ensure the conservation of the charges, that is when the electron and/or hole is filled by an arbitrary charge, and the photocatalytic reaction may take place. In order to study the photocatalytic activity of Mg-doped Fe2O3 thin films, many parameters such as adsorption process, particle size, morphology, and crystallinity catalyst performance were considered. Mg-doped Fe2O3 with y = 3 at.% provides a good platform for the photodegradation of organic dyes. Figure 9a shows the absorbance spectra recorded in the wavelength range from 400 to 800 nm. As shown in the absorption spectra of MB, the main peak intensity occurred at 664 nm over the reaction

Photocatalytic Activity
It is well recognized that the photocatalysis process occurs through the photogeneration of electron-hole pairs under light irradiation [2,4]. When photons reach a semiconductor catalyst, an electron absorbs the photon energy to move from valence band to occupy conduction band levels, leaving behind an electron vacancy (hole); thus, creating the electron-hole pairs. On the other hand, the lifetime of electron-hole pairs is very short [13,15,17]. To ensure the conservation of the charges, that is when the electron and/or hole is filled by an arbitrary charge, and the photocatalytic reaction may take place. In order to study the photocatalytic activity of Mg-doped Fe 2 O 3 thin films, many parameters such as adsorption process, particle size, morphology, and crystallinity catalyst performance were considered. Mg-doped Fe 2 O 3 with y = 3 at.% provides a good platform for the photodegradation of organic dyes. Figure 9a shows the absorbance spectra recorded in the wavelength range from 400 to 800 nm. As shown in the absorption spectra of MB, the main peak intensity occurred at 664 nm over the reaction time using Mg-doped Fe 2 O 3 as catalyst. As irradiation was carried out, the absorption intensity decreased, with about 90% of degradation achieved within 180 min (Figure 9b) This film shows higher degradation efficiency compared with the undoped Fe 2 O 3 thin films, which degraded almost 50% of MB under (Figure 9a) sunlight irradiation. The substitution of Fe 3+ by Mg 2+ sited onto the Fe 2 O 3 lattice led to the enhancement in the photocatalytic activity. The possible reaction mechanism for the photocatalytic process of the mentioned sample is shown below [12]: MB + (OH • , O 2 •− ) → Intermediates → Decomposition product (10) leading to an enhancement in the photocatalytic degradation efficiency. Many researchers have reported that the activity of semiconductor materials in dye degradation processes is affected by metal transition dopants [13,17,18]. In order to investigate the kinetic of MB photodecomposition on the catalyst surface, Figure 9b shows the curve fitting of the kinetic equation. A straight line was obtained, indicating that the reaction is of pseudo-first order. Thus, 3 at.% Mg-doped Fe2O3 thin layer presents a performant photocatalyst.

Conclusions
Fe2O3 and Mg-doped Fe2O3 films were grown using chemical spray pyrolysis technique and then characterized using several analytical techniques. A structural characterization study revealed that all the films are highly crystallized in the rhombohedral structure with (104) as principal orientation, which confirms that Mg was well-incorporation onto the Fe2O3 lattice, especially for 3 at.% of Mg-doped Fe2O3. The morphological properties of the films showed a rough surface with granular texture, with the highest values of Sa and Sq for 3 at.% Mg-doped Fe2O3. Moreover, the roughness values obtained for 3 at.% Mg-doped Fe2O3 may improve the sensitivity of iron oxide to As a result, the linked works of the photogenerated electrons and holes cause the MB degradation, where the e-reacts with oxygen molecules to produce superoxide anion radicals (O 2 •− ) and the h + reacts with water to form hydroxyl radicals. The inserted magnesium ions act like a charge sink, which consequently can enhance the separation of the photogenerated charge pairs, giving rise to further superoxide and hydroxyl radicals, leading to an enhancement in the photocatalytic degradation efficiency. Many researchers have reported that the activity of semiconductor materials in dye degradation processes is affected by metal transition dopants [13,17,18]. In order to investigate the kinetic of MB photodecomposition on the catalyst surface, Figure 9b shows the curve fitting of the kinetic equation. A straight line was obtained, indicating that the reaction is of pseudo-first order. Thus, 3 at.% Mg-doped Fe 2 O 3 thin layer presents a performant photocatalyst.

Conclusions
Fe 2 O 3 and Mg-doped Fe 2 O 3 films were grown using chemical spray pyrolysis technique and then characterized using several analytical techniques. A structural characterization study revealed that all the films are highly crystallized in the rhombohedral structure with (104) as principal orientation, which confirms that Mg was well-incorporation onto the Fe 2 O 3 lattice, especially for 3 at.% of Mg-doped Fe 2 O 3 . The morphological properties of the films showed a rough surface with granular texture, with the highest values of Sa and Sq for 3 at.% Mg-doped Fe 2 O 3 . Moreover, the roughness values obtained for 3 at.% Mg-doped Fe 2 O 3 may improve the sensitivity of iron oxide to detect toxic gases, which recommends the 3 at.% Mg-doped Fe2O3 thin film for gas sensor devices. The Mg doping did not significantly affect the indirect optical band gap (Eind), while the direct band gap increased from 2.15 eV for undoped films to 2.20 eV for 3 at.% Mg-doped Fe 2 O 3 . These band gaps values of Mg-doped Fe 2 O 3 recommend the use of these films as an optical window or buffer layer in photovoltaic devices. Otherwise, magnesium incorporation enhanced the photogenerated charge separation, greatly improving the photocatalytic performance of iron oxide material. The dye degradation activity of the 3 at.% Mg-doped Fe 2 O 3 catalyst reached 90% after 180 min of sunlight irradiation.

Data Availability Statement:
The data is available on reasonable request from the corresponding author.