3.1. X-ray Diffraction Study
From the XRD patterns of all the synthesized catalysts shown in Fig. 2a & b, all samples including undoped TiO2 calcined at 450 ºC for 5 h showed the formation of anatase phase having characteristic high-intensity diffraction peak at 2θ = 25.3º along with other corresponding small diffraction peaks at 2θ values 37.9º, 48.05º and 54.1º that can be indexed to planes of (101), (004), (200) and (211) of anatase phase respectively (JCPDS No. 21-1272). No extra peak found at 2θ=27.8º, indicated that there is no formation of the rutile phase. As the ionic radii of Mn2+ (0.78Å) and Mg2+ (0.72Å) are closer to the ionic radii of Ti4+ (0.68 Å), Mn2+ and Mg2+ dopant metal ions were expected to substitute Ti4+ ions in TiO2 matrix which is confirmed by the absence of any diffraction peaks related to Mn and Mg oxides or other compounds [17,18]. And it is also known that as the Mn2+ and Mg2+ are more electropositive, the electronic cloud in each TiO2 might be loosely held, favoring the formation of less dense anatase phase . The average crystallite sizes of the undoped, bimetal-doped (MMT) and surfactant-assisted co-doped (MMT-GS) catalysts were calculated based on the FWHM of the characteristic high intensified peak using Scherrer equation  (d = kλ / β Cos θ), where d is the average crystallite size, k is 0.9 (Scherrer constant), λ is 1.5406 Ǻ (X-ray wavelength), β is the full width at half maximum (FWHM) and θ is the diffraction angle and are tabulated in Table 2. From the table, the average crystallite size of the catalysts found to be ranging from 7.21-10.22 nm, 10.86-12.87 nm, and 18.30 nm for surfactant-assisted co-doped (MMT-GS), bimetal doped (MMT), and undoped TiO2 nanocatalysts, respectively. The substitutional doping of metal ions into the TiO2 lattice inhibited the grain growth by formation of Ti–O–Mn and Ti-O-Mg due to which the crystallite size was decreased in MMT catalysts. Further decrease in crystallite size was observed for the catalyst prepared in presence of surfactant, MMT-GS, which can be attributed to the effective capping nature of the Gemini surfactant which controls the nucleation and minimizes the agglomeration of TiO2 nanoparticles during the synthesis process .
3.2. UV-Visible Diffuse Reflectance Spectra Study
From the UV-visible diffuse reflectance spectra of the undoped TiO2, MMT, and MMT-GS nanomaterials shown in Fig. 3a, it is observed that absorbance bands are shifted more towards the higher wavelengths i.e., red-shifted in MMT and MMT-GS catalysts compared to undoped TiO2, which is possible due to the decreased bandgap by co-doping of Mn2+ and Mg2+ into the TiO2 matrix. The decrease in bandgap can be ascribed to the joint effect of these two metal dopant ions (Mn2+ & Mg2+) in MMT and MMT-GS system, where these formed the extra energy levels below the conduction band of TiO2 and thus reduces the electron-hole recombination by trapping electrons and enhances the visible light absorbance. From Fig. 3b it is further confirmed by bandgap energies obtained for all the synthesized catalysts using Kubelka-Monk formalism and Tauc plot method . The corresponding band gap energy values are presented in Table 2. From the table, it is found that the average band gap values of the photocatalysts are in the range of 2.66-2.83, 2.68-2.97, and 3.20 eV for MMT-GS, MMT, and undoped TiO2, respectively. These outcomes revealed that all MMT and MMT-GS catalysts are visible light active and can be used as better photocatalysts. The catalysts MMT5, MMT5-GS1, MMT5-GS2, and MMT5-GS3 have the same weight percentage of Mn and Mg; consequently, they have the absorption peaks nearly at the same wavelength. But in particular, among all the MMT’s and GS assisted bimetal doped catalysts, MMT5-GS2 showed the least bandgap energy i.e., 2.66 eV.
Before proceeds to further characterization, we have conducted trial photocatalytic degradation experiments for all the synthesized catalysts using MR dye. From the results it was noticed that among all the bimetal doped TiO2 catalysts, 0.25 wt% Mn2+ & 1.00 wt% Mg2+ bimetal doped TiO2 (MMT5) exhibited better photocatalytic activity and moreover MMT5 assisted with 10 wt% of GS, MMT5-GS2 has shown some more enhanced photocatalytic activity. Hence, we selected these two particular catalysts for further characterization with SEM-EDX, TEM, BET, FT-IR, and photoluminescence study.
3.3. Scanning Electron Microscopy and Energy Dispersive X-ray Spectroscopy
The FESEM images shown in Fig. 4a, b and c illustrate the surface morphology of undoped TiO2, MMT5, and MMT5-GS2, respectively. On comparison, it can be clearly seen that the morphology of the TiO2 has been changed from large aggregates with a scratchy surface in undoped TiO2 (Fig. 4a) to multiparticle agglomerated irregular shaped particles with rough surface in MMT5 and pseudospherical less agglomerated nanoparticles with smooth surface area and small particle size in MMT-GS2. This is clearly indicating that Mg2+ and Mn2+ bimetal ions doping and capping action with the Gemini surfactant have a significant impact on the morphology of TiO2 nanoparticles which controlled the grain growth and particle nucleation.
Also, the chemical composition of MMT5-GS2 sample was investigated by EDX and the results are depicted in Fig. 5. Along with the Ti and O elements of TiO2, the dopant elements Mn and Mg are found in the spectra supports the presence of dopants in the TiO2 matrix and there are no GS associated peaks are found in the spectra indicated complete elimination of GS after the calcination process. The quantitative analysis results are listed in a table inserted in Fig. 5 describes weight percentage as well as the atomic percentage of the compositional elements of the sample which indicates good compatibility with the dopant concentration used for the synthesis of TiO2 nanoparticles.
3.4. Transmission Electron Microscope
The TEM images of MMT5 and MMT5-GS2 photocatalysts are shown in Fig. 6a and b respectively. The TEM image (Fig. 6a) of MMT-5 shows pseudospherical shaped with multiparticle agglomeration and an average particle size of 6.6 nm. Whereas in MMT5-GS2 (Fig. 6b), most of the TiO2 particles are well dispersed with very low agglomeration compare to Fig. 6a and the particle size is in the range of 2.3-5.4 nm, with an average particle size of 3.8 nm. Thus, it clearly indicates that the GS effectively inhibited the particle overgrowth and aggregation and resulted in small particle size with increased surface area, which is well correlated with BET results given in Table 2. The selected area electron diﬀraction (SAED) patterns shown in Fig. 6c conﬁrmed the anatase phase with good crystallinity indexed by the concentric rings, which is in good agreement with XRD diﬀraction patterns. In addition to SAED, from the HR-TEM image of MMT5-GS2 (Fig. 6d) the observed lattice fringes with d-spacing of 0.352 nm correspond to (101) plane of anatase and further confirms the single-crystal nature and high crystallinity of anatase TiO2 .
3.5. Fourier Transform Infrared Spectroscopy
The incorporation of metal ion dopants, Mn2+ and Mg2+ into TiO2 lattice was further confirmed by FT-IR results. The FT-IR spectra of the undoped TiO2, MMT5, and MMT5-GS2 (before and after calcination) were shown in Fig. 7a-d. The peaks at 3403.01, 2926.10, 1616.36, 1383.78, 3378.14, 2918.69, 1626.67, and 1372.07 cm-1 corresponds to stretching vibrations of surface O-H and 3351.79, 2918.72, 1619.40 and 1383.80 cm-1 corresponds to the and bending vibrations of adsorbed H2O molecules . The stretching vibrations of Ti-O and bending vibrations of Ti-O-Ti observed at 575.41 and 1375.78 cm-1 in undoped TiO2 (Fig. 7a) were deformed/shifted to 620.32 and 1300.12 cm-1 in MMT5 (Fig. 7b) and 540.01 and 1182.01 cm-1 in MMT5-GS2 (Fig. 7c & d) respectively, which can be attributed to the presence of dopants in TiO2 lattice. Hence, the FT-IR study confirms that Mn2+ and Mg2+ are substitutionally doped into TiO2 lattice by replacing Ti4+ and formed a new network i.e., Ti-O-Mn and Ti-O-Mg which are in good agreement with the previous reports [24, 25]. The FT-IR spectra of GS are shown in Fig. S1 as a supplementary file. The bands situated at about 2945.62, 1447.80, 1345.72, 1246.03, 1149.79, 1059.62, and 996.59 cm-1 in GS were shifted to 2855.77, 1734.95, 1459.87, 1206.73, 1136.50, 1094.06, and 982.86 cm-1 in MMT5-GS2 before calcination (BC) shown in Fig. 7c which confirms the existence of strong electrostatic interaction between GS and surface of catalyst in MMT5-GS2 before calcination .
From Fig. 7d, the absence of these peaks confirms that there was no surfactant remained in the synthesized catalyst, MMT5-GS2 after calcination (AC). This indicated that due to calcination at 450 °C, the surfactant is completely eliminated from nanocatalyst.
3.6. Brunauer-Emmett-Teller surface area analysis
To study the effect of bimetal doping and GS on the surface area and porosity nature of the as-synthesized TiO2 nanoparticles of undoped TiO2, MMT5 and MMT5-GS2, N2 adsorption-desorption isotherms and their corresponding Berret- Johner- Halenda (BJH) pore size distribution plots were recorded and presented in Fig. 8a-b. From Fig. 8a, it results in a type -IV isotherm with H2 hysteresis loop, characteristic of the ordered mesoporous structure of the catalyst . The average surface area, ABET (m2g-1) of all prepared catalysts was determined and tabulated in Table 2.
From Table 2 the MMT has shown increased surface area (112.03 m2g-1) compared to undoped TiO2 (64.09 m2g-1) due to crystal growth suppression by dopants. Whereas catalyst prepared in the presence of GS surfactant, MMT5-GS2 showed a higher surface area (230.20 m2g-1) compared to both MMT5 and undoped TiO2. It could be strong evidence for the decreased particle size of the TiO2 resulted from the effective capping ability of the surfactant, which restricts the particle growth and nucleation during the synthesis process. To better compare the effect of bimetal doping and capping of GS on as-synthesized TiO2 nanoparticles over single Mn2+ and Mg2+ doped TiO2, comparative results are tabulated by referencing Mn  and Mg  single doped literature reports in Table 3.
3.7. Photoluminescence Spectra
One of the essential reactive species in the process of photocatalysis is hydroxyl radical (•OH) and is responsible for oxidation reactions. Because of the high reactivity and short life of hydroxyl radical, it is impossible for direct detection. To inspect the production of hydroxyl radicals from the catalyst during the photocatalysis reaction, a photoluminescence technique has been adopted using coumarin as a fluorescent probe molecule, which on reaction with hydroxyl radicals yields the 7-hydroxy coumarin . In this technique, 0.10 g of catalyst is dispersed in 100 mL of 10 ppm coumarin solution in acidic conditions and illuminated to visible light irradiation. Small aliquots of reaction solution samples were withdrawn for every 30 min, filtered and photoluminescence intensity was measured in the range of 350-600 nm with excitation fixed at a wavelength of 435 nm. It has been observed from Fig. 9 that in-between between 440-460 nm there exist photoluminescence spectra of the generated 7-hydroxy coumarin with maximum emission at 450 nm. A linear increase in photoluminescence intensity was observed with increasing irradiation time. However, in the absence of irradiation, no excitation was observed for the sample (0 min in Fig. 9).
From the above results, it could be understood that the number of hydroxyl radicals produced on the catalyst surface was directly proportional to the irradiation time. The results further assure that the synthesized nanomaterial showed the enhanced rate of formation of hydroxyl radicals under irradiation to visible light. This is due to the fact that the already formed photo holes in the valence band of bimetal doped TiO2 could directly react with H2O/OH- to produce hydroxyl radical .