Synergistic effect of N-decorated and Mn2+ doped ZnO nanofibers with enhanced photocatalytic activity

Here we report a high efficiency photocatalyst, i.e., Mn2+-doped and N-decorated ZnO nanofibers (NFs) enriched with vacancy defects, fabricated via electrospinning and a subsequent controlled annealing process. This nanocatalyst exhibits excellent visible-light photocatalytic activity and an apparent quantum efficiency up to 12.77%, which is 50 times higher than that of pure ZnO. It also demonstrates good stability and durability in repeated photocatalytic degradation experiments. A comprehensive structural analysis shows that high density of oxygen vacancies and nitrogen are introduced into the nanofibers surface. Hence, the significant enhanced visible photocatalytic properties for Mn-ZnO NFs are due to the synergetic effects of both Mn2+ doping and N decorated. Further investigations exhibit that the Mn2+-doping facilitates the formation of N-decorated and surface defects when annealing in N2 atmosphere. N doping induce the huge band gap decrease and thus significantly enhance the absorption of ZnO nanofibers in the range of visible-light. Overall, this paper provides a new approach to fabricate visible-light nanocatalysts using both doping and annealing under anoxic ambient.

Another important feature in catalytic performance is that there is no driving force to trap electrons on doping centers and hence they will have better photoconductivity. A high spin d 4 ion like Mn 3+ , the tremendous intra-atomic exchange energy gained upon converting to a d 5 configuration provides an enormous driving force to trap electrons that are diffusing nearby. By contrast, d 0 , d 10 , and high spin d 5 ions gain no exchange energy by trapping an electron (e.g., Fe(III), Mn(II), and Cu(I)), which makes them the ideal candidates for active dopants. Among the metal ions, Mn 2+ ions with half-filled electronic configuration show excellent catalytic activity. They could trap the charge carriers shallowly and efficiently to facilitate the detrapping process of charge carriers to the surface of the catalyst which accelerate the interfacial charge transfer processes 17 . Moreover, previous studies of N-doped TiO 2 or ZnO have shown a significant red shift in light absorption wavelength and notable improvement in photocatalytic activity 14,18,19 . Here we report a significant enhancement in visible-light photocatalytic activity of ZnO by creating intentional defects in its crystal lattice via Mn 2+ -doping and N 2 atmosphere annealing. Mn 2+ was chosen because it will not act as a recombination center but will create more intermediate states.
To date, various methods have been employed to obtain high-performance photocatalysts, such as hydrothermal 20 , sol-gel 21 , microwave irradiation 16 , chemical precipitation 22 , electrospinning 23 , etc. Among these approaches, electrospinning appears to be the most facile and practical technique for manufacturing nanofibers with small crystal size 23,24 . This technique performance is not only feasible and simple but also it is an ideal way to prepare nano-materials with specific composition, large specific surface area and recoverable character. It is also reported that an increasing number of nanocatalysts are fabricated by this approach [25][26][27] . Some notable examples include ZnO, TiO 2 , SnO 2 , ZnS and BiVO 4 1, [28][29][30][31] , etc. Hence, in this research, a simple and efficient electrospinning method combined with the controlled heat treatment were applied for the synthesis of Mn 2+ -doped and N-decorated ZnO nanofibers. The morphologies, crystalline structures, optical properties, and photocatalytic performances for all samples were explored to clarify the correlation between nanostructures and photocatalytic properties. Furthermore, underlying mechanisms responsible for this novel photocatalytic material have been proposed. The ensemble of results showed that enhanced photocatalytic performance and excellent efficiency are owing to both the Mn 2+ and N doping, which induce large amounts of surface defects. It is envisioned that our work may provide a new approach for preparing nano-sized ion-and N-doped ZnO photocatalysts with high density of oxygen vacancies.

Results
Characterization of Mn-ZnO nanofibers. Figure 1a shows the SEM images of the as-spun 0.15 Mn-ZnO nanofibers with smooth surface and a range of diameters about 300 nm. By using electrospinning, all the fibers collected on the aluminum substrate show uniform and high density. After the as-spun fibers calcined for 1 h under inert atmosphere, the continuous structure of the fibers were maintained, and the surface became rough. The diameter after annealing reduces to 100 ± 20 nm (see Fig. 1b,c).
To further investigate the internal structure, the nanofibers were dispersed in ethanol evenly and viewed under TEM. It is obvious that the nanofiber shows a fibrillar structure after annealing. A selected-area electron diffraction (SAED) pattern reveals that the nanofibers are comprised of polycrystalline and fine grains of ZnO with no preferential orientation. The intensity of the wurtzite rings is weak, however, the presence of (002) (101) (100) and (103) lattice spacings can be resolved. As shown in Fig. 2e, recorded from the area marked by the red circle in Fig. 2d, the high resolution TEM (HRTEM) image indicates that the lattice constants between two adjacent parallel atomic lattice-fringes are 0.19 nm, which can be attributed to the ZnO (102) plane. Both the SAED and HRTEM results demonstrate the annealed 0.15 Mn-ZnO nanofibers are composed of only hexagonal ZnO and no other phases, such as MnO or MnO 2 . Furthermore, Energy-dispersive X-ray spectroscopy (EDS) analysis confirms the concentration of Mn is found to be 14.85 at.% (Fig. 2f).
The phase composition and structure of the annealed samples were examined by X-ray diffraction (XRD). Figure 2 shows XRD patterns of a series of Mn-ZnO nanofibers with different concentrations of Mn 2+ ions annealed at 550 °C. The main diffraction peaks are indexed to hexagonal wurtzite phases, and there is no other manganese oxide phase, which is in accord with the TEM result. All the diffraction peaks fit well with the hexagonal phases, as labeled in Fig. 2a, manifesting Mn atoms were successfully doped into the lattice. Since Mn easily donates its electron to O, the preferable site of substitution of Mn in ZnO lattice is Zn site. However, XRD spectra is unable to detect a crystal structure if the content of it is below 3%. Thus, an excess Mn 2+ doping have done to identify the saturated concentration of Mn 2+ , as shown in Figure S1. The crystalline structures corresponding to MnO and ZnO phases are confirmed, and no other peak relates to manganese compound with high valence is detected. Moreover, the concentration of these two phases were measured by the quantitative analysis using jade 8. The concentration of MnO phases in 0.5 Mn-ZnO sample is about 30 at.%, which indicates nearly 30 at.% Mn 2+ could be loaded in ZnO lattice. Thus, all Mn 2+ atoms were successfully doped in the lattice even after 15 at.% loading in this study. Moreover, it is noted that the diffraction peaks for Mn-ZnO samples gradually shift to lower angles with the increase of Mn concentration. Clearly, it could also be seen from the step scanning XRD patterns that the single diffraction peak in the range of 55.2-58° shifts to lower angles (Fig. 2b). With increasing Mn content, a clear increase of the lattice constants could be observed from the results of lattice calculation (Supplementary Table S1). Since the Mn 2+ ions occupy the positions of the Zn 2+ ions, the lattice parameters and cell volume accompany with an increase, leading to a shift to lower diffraction angles of XRD peaks. As the ionic radius of Mn 2+ (80 pm) is larger than that of lattice Zn 2+ (74 pm), the hypothetic Mn 4+ (67 pm) doping could be excluded 32,33 . Photocatalytic Activity and Stability. The photocatalytic activities of the nanofibers with different Mn concentration were evaluated by the degradation of RhB dyes under visible light irradiation (400 < λ < 750). Temporal changes in the concentrations of RhB were monitored by examining the variations in the maximal absorption peak at 554 nm in the UV− vis spectra. The total optical power impinging on the solution is 100 mW mL −1 , and the effective surface area of the nanofibers measured by BET test is 18.86 ± 2.23 m 2 /g. Figure 3b shows the photocatalytic removal of the RhB using 0.15 Mn-ZnO nanofibers under visible light for 100 min. The inset of Fig. 3b exhibits SEM images of the nanofibers after photocatalytic measurement, which displays the catalysts still keep their three-dimensional structure but some are broken up into pieces during ultrasonic suspension. Comparative experiments were also carried out to investigate the photocatalytic activity of the nanofibers with different Mn composites under identical conditions. In Fig. 3c, when Mn 2+ was doped in ZnO nanofibers, a remarkable enhancement in photocatalytic activity is observed clearly. The optimal loading is around 15 at.%, since RhB is completely degraded after irradiating 100 min. The RhB adsorbed on the sample surface has also been proved to be removed totally (see FTIR results, Supporting Information Figure S2). There are no observable changes of the peaks in the 0.15 Mn-ZnO nanofibers before and after photocatalytic cycles. Hence, not only the photocatalytic activity of ZnO nanofibers is highly enhanced by Mn 2+ doping, they also show a stable photocatalytic activity under visible light. Figure 3d displays no decay of this sample for photocatalytic activity within several cycles. As known, the photodegradation of RhB can be considered as a pseudo-first-order process 34,35 , and its kinetics can be described by the formula as follows: where C is the concentration of RhB that at the reaction time t, C 0 is the initial concentrations of RhB, and k is the degradation rate constant. The 0.15 Mn-ZnO nanofibers exhibit the highest catalytic activity with constant rate (k) of 2.5 × 10 −2 min −1 , which is more than 50 times than that of ZnO nanofibers (4.7 × 10 −4 min −1 ), shown in Fig. 3e. Obviously, RhB is started to degraded under visible light with the help of 5 at.% Mn 2+ -doped catalyst (see Fig. 3c), and the rate constant is 4.8 × 10 −3 min −1 , which is 10 times than pure ZnO. Furthermore, the rate constants increase to 1.1 × 10 −2 and 2.5 × 10 −2 min −1 after 10 and 15 at.% doping, respectively. To quantize the apparent quantum efficiency (AQE) of all samples, AQE can be defined as follows 31 : /dt is the initial rate for RhB degradation per unit volume, which equals to kC 0 in our case (C 0 = 2.5 × 10 −5 mol/L); d[hν] inc /dt is the total optical power (TOP) radiating on the sample. As the catalyst loading is 1 g/L and the density for ZnO is 5.606 g/cm 3 , the TOP radiating on the nanofibers is 0.01784 mW/mL. Then, the AQE for all samples could be calculated by using formula 2 (see Fig. 3e). In this paper, TOP is defined by the incident photons on the sample surface instead of the photons absorbed by the photocatalyst. Thus, the real quantum efficiency might be even higher than the obtained value. In spite of this there are indeed significant improvement of photocatalytic properties in Mn-ZnO nanofibers. The AQE is 2.47% for the 5 at.% Mn 2+ -doped sample, which improves 10 times compared with pure sample (0.24%). After loading up to 15 at.%, the AQE increases to 12.77%, which is 50 times than that of pure ZnO. This excellent degradation behavior enables ZnO a promising alternative for visible light active photocatalysts.
Photocatalytic Mechanism. From the bandgap energy point of view, we have studied the improved photocatalytic mechanism of Mn 2+ doping in the visible light range. Figure 3a represents UV− vis diffuse reflection spectra of ZnO and Mn-ZnO nanofibers. It is noteworthy that there are varying degrees of absorption for all Mn-ZnO samples in the visible light range, which maybe contribute to doping or surface defects. The bandgaps could be obtained from the diffuse reflection spectra using the Tauc plot (inset of Fig. 3a) 36 . Obviously, the ZnO nanofiber sample (3.25 eV) is much higher than Mn 2+ -doped samples, but a little lower than the theoretical value (3.37 eV), which may be affected by the nanometer size effect 19 . And the bandgaps for Mn-ZnO nanofibers are further narrowed by an increased loading (see Table S2). More intuitively, the middle inset of Fig. 3a indicates the color turned to yellow after loading with 15 at.% Mn 2+ ions. There are probably two reasons for this: One is due to the doping ions in lattice that lower the band gap 17 ; another is related to annealing in N 2 ambient, which introduces N-doping 37 or large amount of surface defects 7 , such as dangling bonds or oxygen vacancies. It is documented that surface defects are well-known to enhance photocatalytic activity evidently despite various other crystal defects 38,39 . Note that surface defects are also associated with a number of states such as zinc interstitials ) and zinc vacancies (V Zn ", V Zn ' and V Zn ) 40 . To elucidate the photocatalytic mechanism that involves Mn 2+ doping or rich vacancy defects on the nanofiber surface, we have studied the surface element compositions and chemical states of all samples using XPS analysis and the band gap of Mn 2+ -doped ZnO crystal by first-principles calculations.
For XPS analysis, Zn and O elements are observed in all samples in Fig. 4a, but Mn and N are only detected in Mn-ZnO samples. There is a distinct peak related to N 1s in the 0.15 Mn-ZnO sample, owing to the nitridation on the nanofiber surface. To study the chemical state of each individual element, high-resolution XPS peaks for Mn 2p, N 1s, Zn 2p, and O 1s states are provided. For Mn-ZnO nanofibers (Fig. 4b), a double peak at 641.2 and 655 eV is corresponded to the Mn 2+ -2p level of p 3/2 and p 1/2 , respectively, indicating that the chemical state of Mn is present as Mn 2+ in all Mn-ZnO samples. It further confirms that Mn 2+ -doped ZnO nanofibers have been successfully prepared. In Fig. 4c, the core level spectrum of the N 1s region shows a symmetric peak centered at 399.3 eV, which suggests that the N chemical state is present. Because this peak lies in between the typical binding energy for zinc nitride (396-397 eV) 41 and NO type species (above 400 eV), it can be attributed to the N 1s of oxynitride (O-Zn-N) 18 . This result indicates N atoms occupy at O sites on the ZnO nanofibers surface during annealing, when N 2 as an N precursor favors the formation of Zn-N bond. Since all samples are calcinated under the same condition, the Mn 2+ doping seems to facilitate the surface nitridation, which could be attributed to that Mn owns the more chemical states than Zn and thus easier to be nitrogenized 42,43 . In addition, doping destabilizes the crystal lattice, leading to an access for N atoms embedded in it. As shown in Fig. 2d, the peaks related to Zn 2p for all samples are clearly observed. There are two symmetric peaks in the Zn 2p region. The peak centered at 1022.2 eV corresponds to Zn 2p 3/2 and another one centered at 1045.4 is assigned to Zn 2p 1/2 , indicating a normal state of Zn 2+ in the Mn-ZnO nanofibers. Note that the diffraction peaks of both Zn 2p 3/2 and 2p 1/2 shift monotonically with the increase of Mn 2+ ions, which can be attributed to replacement of Zn 2+ by Mn 2+ and an added Zn-O-Mn binding energy 44,45 . Figure 4e displays the high-resolution O 1s spectra of all doped and undoped samples. It shows the asymmetric O 1s core spectrum deconvoluted with three peaks. The first peak centered at 529.9 eV is corresponded to the O 2− ions in the wurtzite structure (O-Zn) 46 , while the medium peak at 531.6 eV is assigned to O 2− ions in the oxygen deficient regions 47   O-O, respectively. It is noteworthy that the medium binding energy component, centered at 531.6 eV, is connected to the variations in the concentration of oxygen vacancies, therefore, the intensity changes of this component can be connected in part to the variations in the concentration of surface oxygen vacancies. The intensity of this peak increases with the increasing Mn 2+ and N concentration, as Table 1 shows. For 0.15 Mn-ZnO nanofibers, the intensity of the peak at 531.5 eV is obviously stronger than that of others, while the area ratio of the peak at 531.6 eV to the one at 529.9 eV is about 1.15, and this ratio is only 0.09 in the ZnO nanofibers. It suggests that the oxygen vacancies in fibrous surface are highly increased after doping with Mn 2+ while annealing at N 2 atmosphere; this could partly explains the former bandgap results.
Another possible reason for the gap reduction is the partial replacement of Zn 2+ by Mn 2+ in ZnO lattice. To investigate Mn doping effect on orderly ZnO crystal, the bandgap for Mn 2+ -doped ZnO (Mn-ZnO) crystal has been calculated from first-principles using a 48 atom supercell, and the structure of Mn-ZnO (1 0 1 0) surface is shown for replacing one Zn atom with a Mn atom, which corresponds to 4.2 at.% Mn doping, shown in Fig. 5a. Figure 5b shows the total density of states (DOS) for pure and 4.2% Mn 2+ -doped ZnO (0.042 Mn-ZnO), where the Fermi level is set at zero and the "scissor operator" has been set to 1.26 eV. The calculated band gap for bulk ZnO and 0.042 Mn-ZnO crystals after using scissor operator are 3.37 and 3.23 eV, respectively. Thus, it is evident that 4.2% Mn doping could lower the band gap by 0.14 eV, compared with the undoped counterpart. Figures 5c,d depict the partial DOS of states, which indicate that the valence band is mainly contributed by O 2p and Zn 3d states, while the conduction band comes mainly from Zn 4s and O 2p states. In the inset of Fig. 5d Table S2), which displays pale yellow color (Supplementary Figure S3). One of the reason is that ZnO with nanostructure has a narrower band gap nature induced by trap states on the nanofiber surface 4,49 , comparing with that of bulk structure. Another reason may be due to the limited effect of Mn doping independently. To observe the individual catalytic effect of Mn 2+ , the Mn 2+ -doped ZnO sample was prepared by annealing under vacuum circumstance. As expected, the photocatalytic properties of this sample is much lower than the Mn 2+ -doped and N-decorated ZnO samples, which confirmed the vital N doping effect for improving photocatalytic activity (see Figure S4).

Discussion
Combined with the above conclusions, the predominant effects on such a significant band gap decreasing for ZnO can be contribute to the high density of surface defects induced by N doping. It is documented that the band gap of a material could be substantially narrowed by introducing defects and disorder. For instance, Chen et al. 50 fabricated the black TiO 2 nanoparticles with large amounts of defects and disorder introduced by hydrogenating and resulting in a great enhancement of photocatalytic activity. Compared with TiO 2 , it is hard to nitrogenize ZnO under the same conditions. Yang et al. 18 modified it by annealing under ammonia and fabricated N-doped ZnO nanowires. Although N concentration in this sample is only 4%, it shows a significant enhancement in conversion efficiency in the visible region. Since doping introduces more local lattice structure distortion and surface defects, which favors the diffusion of the N atoms, doping with ions make it more easily for N atoms embraced on nanofiber surface. In this work, we demonstrate a 0.15Mn-ZnO nanofiber sample with 10% N concentration in the surface, depending on XPS analysis (Fig. 4c), and show high quantum efficiency in visible-light range. The significant enhanced photocatalytic activities come from the synergetic effects between Mn 2+ and N, which induce defects states and large amounts of oxygen vacancies. The DFT calculation indicates Mn 4d state that brings in the lower-energy mid-gap states is the effect state for reducing the band gap of ZnO. For N doping effect, it is proved that the oxygen vacancies (V O s) in 0.15Mn-ZnO nanofibers surface were highly increased when annealing at N 2 atmosphere (see the XPS results). Normally, V O exist as V O + state and captures charge carriers, thereby delaying the recombination process 13 . Figure 6 shows a schematic of the gain process in photocatalytic property with the presence of rich V O states at the fibrous surface. Here we propose a mechanism to be effective in Mn-ZnO nanofibers that electrons are trapped at the surface oxygen vacancies and multiple holes separate out, then efficiently oxidize the dyes. Upon illumination, when the photocatalysts are excited by photon energy surpassing its band gap, electron-hole pairs are generated and electrons are readily trapped at the V O states, releasing holes from nanofiber surface, which enhance the oxidation reaction with organic pollutants (Fig. 6a). The top schematics in    51,52 , as described by the equations following: where V O * state is 0.86 eV below the CB and V O ++ state is 1.16 eV above the VB, relatively to a typical band gap of 3.37 eV 53 . Figure 6b shows the normal states of V O s on fibrous surface under dark. Upon visible-light illumination, not only the V O s provide the mid-gap states, which serve as intermediate steps for the photoexcitation process, they also capture electrons from CB, helping the holes release from the catalyst, and exist in the grain boundaries (V O ++ ), trapping holes at V O ++ and electrons at V O * , consequently delaying the recombination process (Fig. 6c). Hence, the high level of V O + , trapping at nanofiber surface, drastically affects the band gap and the quantum efficiency of Mn-ZnO nanofibers. Moreover, Figure S5 illustrates our proposed mechanism of the synergetic effects for Mn 2+ and N doping in ZnO photocatalyst. The electrons, instead of generating from the VB, could be first excited from N2p levels to the defect energy states, then transfer to the Mn4d states, or excited from N2p levels to Mn4d states directly 54 . Thus, the synergetic interactions can impart multiple charge transfer pathways and long mean free path of charge carriers in ZnO nanofibers, which highly improve the photocatalytic efficiency.
In summary, we have successfully prepared a Mn 2+ -doped ZnO photocatalyst with high density of surface oxygen vacancies via a facile electrospinning method combined annealing under N 2 ambient. The remarkable enhancement in photocatalytic activity and stable photocatalytic cycles are clearly observed. For 0.15Mn-ZnO nanofibers, the apparent quantum efficiency could reach up to 12.77%, which is 50 times than that of pure ZnO. The significant enhanced visible photocatalytic properties for Mn-ZnO nanofibers are due to the synergetic effects of both Mn 2+ doping and N decorated. The DFT calculation and XPS spectra indicate that Mn 4d and O-Zn-N bonds induce the huge band gap decrease and thus significantly enhance the absorption of ZnO nanofibers in the range of visible-light. Large density of V O s induced by surface N doping provide the mid-gap states, which serve as intermediate steps for the photoexcitation process. Not only they capture electrons from CB, helping the holes release from the catalyst, but also trap holes at V O ++ and electrons at V O * at the same time, consequently delaying the recombination process. Further investigations exhibit that the Mn 2+ -doping facilitates the formation of N-decorated and surface defects when annealing in N 2 atmosphere. Lastly, the method adopted here is not limited to Mn 2+ -doping and can be applied to the other transition metal doping in wide-bandgap semiconductors for visible-light photocatalysis, photoelectrode and photovoltaic cell. Since doping facilitates the nitridation on the surface, we recommend nitriding while doping efficient elements simultaneously such as Fe 3+ , Cu + , Ti 3+ , etc.

Methods
Fabrication of Mn 2+ -doped ZnO nanofibers. Samples of 0, 5, 10 and 15 at.% Mn 2+ -doped ZnO nanofibers were synthesized using an electrospinning approach (identified as ZnO, 0.05Mn-ZnO, 0.1Mn-ZnO, and 0.15Mn-ZnO). Zinc acetate (0.5 g) and manganese acetate were first dissolved in 2 g of deionized water. For chemical reduction of both Mn 4+ and ZnO, certain amounts of glucose as a carbon source had been introduced to the solution. A PVP (0.4g MW = 1 300 000) solution with a concentration of 8 wt % was then prepared from PVP powders and ethanol with stirring. Afterward, the two prepared solutions were mixed and stirred at room temperature for 2 h. Subsequently, the clear transparent precursor solution was drawn into a hypodermic syringe. The positive voltage of 15 kV and a flow rate of 0.5 mL h −1 was applied to the tip and the distance between the needle tip and the collector was 15 cm. Finally, nanofiber samples were removed from the collector plate and taken for thermal treatment in the middle of a furnace tube (volume, 1 L). With a heating rate of 10 °C min −1 , the prepared samples were annealed under a controlled anoxic environment, which was 50% in air and 50% in nitrogen atmosphere with a flow of nitrogen at 10 ml min −1 , then after heating to the temperature of 550 °C, the sample would be in pure nitrogen atmosphere. A small amount of oxygen was introduced to decompose PVP, and the second thermal process was carried out for chemical reduction of Mn ions and surface O 2+ .
Characterization. The external morphologies of the as-prepared nanofibers were examined using a scanning electron microscope (SEM, JSM-7001F, JEOL, Tokyo, Japan), and a transmission electron microscope (TEM, JEM-2010F, JEOL, Tokyo, Japan). The X-ray diffraction (XRD) patterns of all samples were analyzed using a diffractometer (XRD, D/max-2500, Rigaku, Tokyo, Japan) equipped with Cu Kα radiation. Further surface compositions and oxygen vacancies of the Mn-ZnO nanofibers were determined by X-ray photoelectron spectroscopy (XPS, PHI-5300 ESCA, PerkinElmer, Boston, MA).
The photocatalytic activity was evaluated by the degradation of rhodamine B (RhB) dyes under visible light irradiation (λ > 400 nm) using a 300 W Xe lamp with a 400 nm cutoff filter, and the average visible light intensity was 100 mW mL −1 . The prepared films were immersed in RhB aqueous solution (2.5 × 10− 5 M, 10 mL). Before illumination, the RhB aqueous solution was stirred in the dark for 2 h to ensure an established adsorption/desorption equilibrium. During photodegradation testing, the RhB solution with the photocatalyst film was continuously stirred using a dynamoelectric stirrer, and the absorbance of the resulting solution was monitored by colorimetry with a UV− visible (UV− vis) spectrophotometer (Shimadzu, UV3600).
First-principles calculation. Our calculations based on the first-principles density functional theory (DFT) were carried out by the Vienna Ab initio Simulation Package (VASP). The projector augmented wave (PAW) method was used in the description of exchange-ion interactions. The exchange and correlation functional were Scientific RepoRts | 6:32711 | DOI: 10.1038/srep32711 described within the generalized gradient approximation (GGA) with the Perdew-Burke-Ernzerhof (PBE). A Monkhorst-Pack k-point mesh of 2 × 2 × 1 was used for irreducible Brillouin zone sampling, while the plane-wave cutoff energy was set to 400 eV. In all calculations, self-consistency was achieved when the total energy converged to less than 1.0 × 10 −4 eV/atom and the geometry relaxation tolerance was below 0.01 eV/Å in all forces. On the basis of the wurtzite structure with space group symmetry of P63mc, the 2 × 2 × 3 supercell containing 48 atoms was adopted for pure ZnO. The calculated bulk lattice parameters are a = b = 3.249 Å, c = 5.205 Å, α = β = 90°, and γ = 120°, which are obtained from XRD results. For the Mn-ZnO sample, one Zn atom is replaced by a Mn atom giving the doping molar concentrations of 4.16 at.% (see Fig. 6a).
It is well-known that first-principle calculations tend to underestimate the absolute band gap values of semiconductors 55 . Some authors introduced coulomb interactions between the d electrons or calculated both the spin-polarized band structures and density of states of doped configurations to correct the calculated values 56,57 . Since we intended to compare the relative changes in energies induced by Mn 2+ doping, we applied a simple "scissor operator" to correct the band gap to the experimental value 58 . In the scissor operator method, the conduction band was shifted upwards by 1.26 eV.