Nitrogen Incorporated Photoactive Brownmillerite Ca2Fe2O5 for Energy and Environmental Applications

Ca2Fe2O5 (CFO) is a potentially viable material for alternate energy applications. Incorporation of nitrogen in Ca2Fe2O5 (CFO-N) lattice modifies the optical and electronic properties to its advantage. Here, the electronic band structures of CFO and CFO-N were probed using Ultraviolet photoelectron spectroscopy (UPS) and UV-Visible spectroscopy. The optical bandgap of CFO reduces from 2.21 eV to 2.07 eV on post N incorporation along with a clear shift in the valence band of CFO indicating the occupation of N 2p levels over O 2p in the valence band. Similar effect is also observed in the bandgap of CFO, which is tailored upto 1.43 eV by N+ ion implantation. The theoretical bandgaps of CFO and CFO-N were also determined by using the Density functional theory (DFT) calculations. The photoactivity of these CFO and CFO-N was explored by organic effluent degradation under sunlight. The feasibility of utilizing CFO and CFO-N samples for energy storage applications were also investigated through specific capacitance measurements. The specific capacitance of CFO is found to increase to 224.67 Fg−1 upon N incorporation. CFO-N is thus found to exhibit superior optical, catalytic as well as supercapacitor properties over CFO expanding the scope of brownmillerites in energy and environmental applications.


Results and Discussions
X-ray diffraction patterns of both CFO and CFO-N are shown in figure 1a are in good agreement with the orthorhombic crystal system. Figure 1b shows a shift in peak position for (200) in CFO-N towards higher 2θ values implying the N incorporation 34,35 . Figures 1(c,d) show the crystal structure of CFO and CFO-N. The oxygen sites of FeO 6 octahedra and FeO 4 tetrahedra are partially replaced with N atoms, leading to the formation of FeO 6−x N x and FeO 4−x N x in CFO-N. Figure 2a,b do not reveal any appreciable change in morphology upon N incorporation. EDX mapping of CFO-N ( Fig. 2(c-g)) reveal the presence and uniform distribution of Ca, Fe, O and N elements throughout the compound. The EDX elemental mapping and spectra of CFO and CFO-N are shown in Figs. S1 and S2. EDX elemental mapping of CFO-N revealed that 3.36% atomic percentage of N was incorporated in CFO. Figure 3(a-d) show HRTEM and SAED micrographs of CFO and CFO-N samples having clear resolved crystalline domains corresponding to (200) lattice plane. The lattice spacing corresponding to (200) plane in CFO-N (2.71 Å) is found to be smaller than that of CFO (2.76 Å), which is also consistent with the small shift observed in XRD. The SAED pattern also establishes the crystalline nature of CFO and CFO-N samples.
X-ray photoelectron spectroscopy (XPS) measurements of CFO and CFO-N confirm the valence states of constituent elements, simultaneously pointing towards the presence of nitrogen in CFO-N. XPS spectra corresponding to Ca 2p of CFO and CFO-N ( Fig. 4(a)) reveal a split into doublet due to spin-orbit coupling of Ca 2p 3/2 and Ca 2p 1/2 with binding energy difference ~3.42 eV and ~3.41 eV respectively. The split points toward the existence of Ca in 2+ oxidation state in both these samples. Similarly for Fe, the Fe2p splits into two spin orbits Fe 2p 3/2 and Fe 2p 1/2 peaks ( Fig. 4(b)). The positions of Fe 2p 3/2 and Fe 2p 1/2 spin orbits at 711.62 eV and 724.40 eV respectively for CFO confirms the existence of Fe in Fe 3+ valence state 36 . A small shift in Fe 2p towards lower binding energy in CFO-N arises due to the replacement of O in Fe-O sites by N. Since N has lower electronegativity as compared to O, it leads to lower binding energy of Fe 2p in CFO-N by forming Fe-N x sites 37 . The XPS spectra of O1s for CFO and CFO-N ( Fig. 4(c)) were deconvoluted into two components at 530.72 eV (OI) and 532.75 eV (OII) for CFO and at 530.66 eV (OI) and 532.88 eV (OII) for CFO-N respectively. The OI component arises from the lattice oxygen, whereas OII is typical of hydroxyl and carbonate groups indicating possible chemisorption of oxygen  The electronic structures of CFO and CFO-N were probed by ultraviolet photoemission spectroscopy (UPS) measurements. It is a well-established fact that the valence band maximum (VBM) calculation can be performed by considering the cutoff for the lowest binding energy. A clear decrease in VBM is observed in CFO-N (2.01 eV) compared with CFO (2.27 eV) shown in Fig. 5b. This points towards occupation of N 2p levels over O 2p levels in valence band leading to an effective reduction in bandgap due to the lower electronegativity of N and creation of Fe-N x active sites in CFO-N. Figure 5(a) illustrates the UV-visible absorption spectra corresponding to CFO and CFO-N. The absorption edge of CFO-N shifts towards higher wavelength indicating a significant reduction in bandgap for CFO-N (2.07 eV) over CFO (2.21 eV) (inset Fig. 5a). The corresponding energy band diagrams of CFO and CFO-N are shown in Fig. 5(c).
The bandgap of CFO could also be tailored by N + ion implantation unlike conventional doping. In order to implant the N + ions, Ca 2 Fe 2 O 5 bulk films were fabricated on glass substrates using Doctor blade method. These CFO films were then implanted with 1 MeV of N + ion beam with a beam current of 1 µA using low energy ion beam facility (LEIBF). The implantation was performed at three ion fluencies 10 14 , 10 15 and 10 16 ions per cm 2 , which were named as CFO-N +--E14, CFO-N + -E15 and CFO-N + -E 16 respectively. X-ray diffraction (XRD) measurements were performed to investigate the effect of N ion implantation on the crystal structure and the nature of crystallinity of CFO films. Figure 6(a) shows the XRD patterns of pristine CFO and N implanted CFO films. No additional diffraction peaks corresponding to other phases are detected.
Nitrogen (N) incorporation in CFO by implantation is also expected to change the optical properties by narrowing its bandgap due to reasons mentioned above. In the present case too, N + ion implanted films showed modified optical properties as compared to pristine CFO (see Fig. 6(b)). Figure 6(c) shows the corresponding Tauc plot. The bandgap values were found to be 1.85, 1.67, 1.56 and 1.43 eV for pristine CFO, CFO-N + -E14, CFO-N + -E15 and CFO-N + -E16 respectively. These studies corroborate the effective reduction in bandgap of CFO due to N incorporation.
The electronic structures of CFO and CFO-N were also investigated theoretically by density functional theory (DFT) calculations, as shown in Fig. 7(a,b). These calculations were performed by sampling the Brillouin zone with a set of high symmetry k-points. The effect of Nitrogen incorporation in CFO was analyzed theoretically CFO and CFO-N have excellent optical and catalytic properties. To explore the use of CFO and CFO-N in environmental remediation applications, we have investigated sunlight driven photocatalytic degradation of organic effluent Mythylene blue (MB). The photodegradation experiments were carried out under natural sunlight at a latitude of 13.0087°N and longitude of 80.2371°E (Chennai, India). The initial dye (MB) concentration was kept at C o : 1 × 10 −5 mol/L. Then 50 mg/100 ml catalyst loaded dye solution was ultrasonicated vigorously under the dark condition and allowed to rest in dark for 10 min to ensure good adsorption. The catalyst loaded dye solution was then kept in sunlight and the sample was collected at regular time intervals. Simultaneous light intensity measurements were also carried out using lux meter. For the whole duration of experiment (50 min) the incident intensity of sunlight was almost constant with an average intensity of about 105.40 klux. The degradation efficiency and first-order reaction kinetics are calculated using the following equations.
where C o and C are the concentrations of MB at 0 min and at time interval t respectively. k is the degradation rate constant.
The schematic representation of photodegradation process, along with degradation profile, efficiency and first-order reaction kinetics of CFO andCFO-N are shown in Fig. 8(a-e). The degradation efficiencies of bare MB, CFO and CFO-N are found to be 3.5%, 83% and 99.7% respectively and the first order rate constants of CFO and CFO-N are found to be 0.0375 min −1 and 0.096 min −1 respectively. CFO-N shows superior photocatalytic performance compared to CFO due to a relatively smaller bandgap as well as the presence of Fe-N x active sites. Adsorption curves of MB and photodegradation performance CFO and CFO-N over different photocatalysts are discussed in supporting information. www.nature.com/scientificreports www.nature.com/scientificreports/ The major active species involved in the photodegradation mechanism of MB was determined by the active species trapping experiments. The trapping experiment results are shown in Fig. 9(a,b). We have used various scavengers such as AgNO 3 (1 mmol) as electron (e − ) scavenger, Ethylenediaminetetraacetic acid (EDTA, 1 mmol) as a hole scavenger (h + ) and Isopropyl alcohol (IPA, 1 mmol) as the OH − radical trapping agent. The trapping agents were added to dye-catalyst solution and degradation experiments were performed under sunlight. In absence of trapping agents, the MB could be degraded by 83% and 99.7% of its initial concentration with the presence of CFO and CFO-N respectively. Upon addition of AgNO 3 as an e − scavenger to CFO and CFO-N dye solution the photodegradation efficiencies were found to be 81.4% and 96.4% respectively. No significant change in photodegradation performance leads us to infer that electrons (e − ) are not involved in photodegradation mechanism of MB by both CFO and CFO-N. With the presence of EDTA as h + scavenger photodegradation efficiencies of CFO and CFO-N were reduced to 31.5% and 31.7% respectively, confirming that photodegradation of MB using CFO and CFO-N are governed by holes. Reactive oxygen species (ROS) such as . O 2 , · OH, singlet oxygen ( 1 O 2 ), peroxyl (RO 2 . ), and alkoxyl (RO − ) are highly active in photochemical reactions such as photodegradation of organic pollutants 40 . Among the all ROS, superoxide (O 2 . ) and hydroxyl( · OH) radicals are the most possible generated ROS in photodegradation of organic pollutants. In the present case, it is evident from e-and h + trapping experiments, that photodegrading phenomena are governed by holes (h + ) and there is no participation of e − in photodegradation process. Hence, the generation of -O 2 . species is limited. This is, because -O 2 . can be generated upon reduction of O 2 by e − 41 . In the present photochemical reaction, · OH is one of the major ROS, which is generated by oxidation of H 2 O/OH − by holes (h+). Since · OH has high tendency to degrade organic pollutants 42 , ·OH species trapping experiment was conducted by addition of IPA; as a result of which the photodegradation efficiencies of CFO and CFO-N were reduced to 47% and 51.1% respectively. Hence, the active species trapping experiments suggest that, photodegradations of MB by CFO and CFO-N are attributed to the holes(h + ) and · OH active species. Figure 10 shows a possible MB photodegradation pathway in the presence of catalyst. During the photocatalytic degradation process, active species such as holes(h + ) and hydroxyl radicals ( · OH) are generated which are responsible for mineralization of MB into unstable organic byproducts Cl − , NO 3 − , SO 4 2− , CO 2 , and H 2 O 36,43 . MB degradation process is initiated by active species breaking the S-Cl bonds and then Cl − ions being separated from MB core structure. Several intermediate molecules such as C 8 H 9 NO 2 , C 9 H 9 NOS and C 7 H 6 N 2 S are generated by breaking N-CH 3 bonds, (which are connected to terminal of MB core structure). The active species further oxidizes the -CH 3 groups. It also gives rise to single ring structures (C 7 H 7 N 3 , C 7 H 6 O 2 and C 7 H 7 Cl). These reactions continue until the MB mineralizes into Cl − , NO 3 − , SO4 2− , CO 2 , and H 2 O.  www.nature.com/scientificreports www.nature.com/scientificreports/ Ca 2 Fe 2 O 5 , by virtue of its unique properties as mentioned above, could also be used for energy harvesting applications such a supercapacitors, batteries and fuel cells etc. In this regard,we have explored the feasibility of CFO and CFO-N for supercapacitor applications. Electrochemical studies have been performed on CFO and CFO-N using three electrode system with 0.5 M Na 2 SO 4 aqueous solution as the electrolyte. These techniques are employed for investigating the redox behavior, stability and supercapacitor properties. The working electrodes are prepared on FTO plates with ethanol as solvent using a doctor blade method. Figure 11(a,b) shows the cyclic voltammetry (CV) curves of CFO and CFO-N at different scan rates (10 mV/s and 100 mV/s). CV curves reveal the presence of redox peaks for both the samples, demonstrating the pseudocapacitive nature of both CFO and CFO-N. A notable difference in the peak current was observed for both samples at low and high scan rates of 10 mV/s and 100 mV/s respectively, which can be attributed to diffusion-controlled reaction kinetics as well as faster electronic and ionic transport in the electrode/electrolyte interface. CFO-N sample shows well defined anodic and cathodic peaks even after 1000 cycles (Fig. 11c,d), which is attributed to a higher stability than CFO. Hence, it delivers a higher specific capacitance with good cycling stability due to the  Fig. 12. GCD studies are carried out using chrono-potentiometric analysis at a current density of 1 Ag −1 , which revealed the capacitive behavior of CFO and CFO-N. The GCD studies are consistent with the CV curves. A non-linear discharging behavior is observed with various current densities. The curvature in charge and discharge profile indicates the capacitive behavior of material, which is influenced by both redox reactions (pseudocapacitor) and electrical double layer (EDLC) response, is similar to reported trends [44][45][46][47] . The specific capacitance values were calculated using the expression below (4)    www.nature.com/scientificreports www.nature.com/scientificreports/ curves point towards improved capacitive properties of CFO-N over CFO. The stability of both materials at various current densities (using GCD plots) are discussed in detail in the supporting information.
Electrochemical impedance spectroscopy measurements were carried out for both samples (Fig. 13). From the impedance plot, it is observed that CFO and CFO-N offer negligible series resistance (R s ) at high frequencies.
Both the samples exhibit the same trend exhibiting high-frequency semicircular arc followed by a straight line at lower frequencies. These studies reveal that both are promising candidates for capacitive applications. However, CFO-N is found to be a better candidate than CFO as it offers lower charge transfer resistance. Electrochemical studies revealed better capacitive behavior for CFO-N as compared to CFO. Hence nitrogen doped brownmillerites can be promising materials for energy and environmental applications. conclusion CFO and CFO-N nanoparticles were synthesized using a chemical route followed by thermoammonolysis. The effect of N incorporation on the bandgap and photocatalytic properties of brownmillerite CFO has been examined. The incorporation of N into CFO significantly decreases the band gap by occupying N 2p energy levels over the O 2p level in the valence band. The electronic band structures of CFO and CFO-N were demonstrated using UPS and UV-vis spectroscopy and the trend was supported by DFT studies. Photodegradation of methylene blue under sunlight revealed superior photocatalytic performance of CFO-N over undoped CFO. Electrochemical studies revealed the feasibility for supercapacitor applications, where CFO-N showed better specific capacitance over CFO. Hence, sunlight-driven photocatalytic degradation of MB and supercapacitor properties expand the scope of utilization of CFO and CFO-N for energy and environmental applications.