Ternary metal oxide nanocomposite for room temperature H2S and SO2 gas removal in wet conditions

A ternary Mn–Zn–Fe oxide nanocomposite was fabricated by a one-step coprecipitation method for the remotion of H2S and SO2 gases at room temperature. The nanocomposite has ZnO, MnO2, and ferrites with a surface area of 21.03 m2 g−1. The adsorbent was effective in mineralizing acidic sulfurous gases better in wet conditions. The material exhibited a maximum H2S and SO2 removal capacity of 1.31 and 0.49 mmol g−1, respectively, in the optimized experimental conditions. The spectroscopic analyses confirmed the formation of sulfide, sulfur, and sulfite as the mineralized products of H2S. Additionally, the nanocomposite could convert SO2 to sulfate as the sole oxidation by-product. The oxidation of these toxic gases was driven by the dissolution and dissociation of gas molecules in surface adsorbed water, followed by the redox behaviour of transition metal ions in the presence of molecular oxygen and water. Thus, the study presented a potential nanocomposite adsorbent for deep desulfurization applications.

www.nature.com/scientificreports/ concentration of 500 and 100 ppm for H 2 S and SO 2 was adopted for their industrial application and suitability in capturing these pollutants in the toxicity range for humans. The oxide showed better adsorption performance in wet conditions with complete mineralization to non-toxic by-products. Besides studying the factors affecting the adsorption process, the adsorption mechanism was studied in detail using various microscopic and spectroscopic techniques. The study confirmed that the oxide nanocomposite has the potential to eliminate and mineralize low concentrations of gaseous H 2 S and SO 2 in dry-wet conditions.

Methods
Chemicals. Manganese A higher ratio of divalent to trivalent cations was adopted for the formation of MnO 2 and ZnO in the nanocomposite (for a higher acidic gas adsorption capacity). Under vigorous stirring, 2.0 mol L −1 NaOH solution was added dropwise until the solution pH reached 12.5. This pH was sufficient for the formation of ternary oxide nanocomposites as reported earlier 18 . After stirring for 2 h, the precipitate was phase separated and dried at 393 K overnight in a hot air oven. The use of an excess of water for washing was avoided to reduce the overall impact on the environment during the material fabrication process.
Analytical instruments. The morphology in the surface and transmission mode was probed over field emission scanning electron microscopy (FE-SEM, Hitachi S-4300, Hitachi, Japan) and field emission TEM (FE-TEM, JEM-2010F, JEOL Ltd., Japan), respectively. SEM analysis was done on finely grounded dried samples after coating them with a gold-platinum alloy by ion-sputtering (E-1048 Hitachi ion sputter). The elemental analysis was done using energy-dispersive X-ray spectroscopy (EDAX, X-Maxn 80 T, Oxford Instruments, United Kingdom). The specific surface area and porosity were determined by analysing the standard N 2 adsorption-desorption isotherm at 77 K using a Gemini 2360 series (Micromeritics, Norcross, United States) instrument after degassing at 423 K for 6 h with a mass of 0.324 g. The powder X-ray diffraction (PXRD) patterns were obtained at room temperature (2θ = 5-50°) on an Ultima IV (Rigaku, Japan) X-ray diffractometer with Cu Kα radiation (λ = 1.5406 Å) and a Ni filter. Fourier-transform infrared (FTIR) spectra of samples were recorded using KBr pellets over a Cary670 FTIR spectrometer (Agilent Technologies, United States). For X-ray photoelectron spectroscopy (XPS) analysis, a K-alpha XPS instrument (Thermo Fisher Scientific, United Kingdom) was used with a monochromatic Al K α X-ray source. The pressure was fixed to 4.8 × 10 −9 mbar. Spectra were charge corrected to the main line of the C 1 s (aromatic carbon) set to 284.7 eV. Spectra were analysed using CasaXPS software (version 2.3.14).
Breakthrough protocol. The gas adsorption experiments were performed by taking 0.5 g of the adsorbent in a Pyrex tube (height: 50 cm and diameter: 1 cm) at 298 K. The sample was fixed between the glass wool and supported on silica beads 19 . The H 2 S gas (0.05 Vol.%) or SO 2 gas (0.01 Vol.%) was passed through it at a fixed flow rate. The outgoing gas was analyzed using an H 2 S gas analyzer (GSR-310, Sensoronic, Korea) or an SO 2 gas analyzer (GASTIGER 6000, Wandi, Korea). The analyzer recorded the effluent gas concentration every minute in real-time until the breakthrough points of 20% (100 ppm for H 2 S and 20 ppm for SO 2 ) were reached and after that the experiment was completed. The wet samples were prepared by passing water vapours (80% relative humidity) directly through the adsorbent bed before passing the gas through it. The gas adsorption capacity was measured using the following equation: where C 0 -initial concentration (mg L −1 ), C-concentration at time 't' (mg L −1 ), Q-flowrate (L min −1 ), m-the mass of adsorbent (g), and t b -breakthrough time (s).

Results and discussion
The SEM micrograph of oxide nanocomposite showed irregularly shaped nano-globules, which were uniformly distributed in the entire region (Fig. 1a). The controlled release of base (precipitation agent) assisted in regulating the nucleation and particle growth kinetics, which prevented the aggregation of metal oxides in the ternary nanocomposite. A more detailed investigation of morphology was conducted over high-resolution TEM, which confirmed that the nano-globules were constructed of polyhedral nanoparticles (Fig. 1b). The crystallite planes of nanoparticles were assigned by measuring the fringe width and correlating with the interplanar spacing (d) values from the XRD pattern. The fringe width of 0.308, 0.530, and 0.261 nm were assigned to the MnO 2 (110) 20 , ZnO (0002) 21 , and MFe 2 O 4 (311) 22 , respectively. The EDAX elemental analysis confirmed peaks for Mn, Zn, Fe, and O at respective energies having the atomic contribution of 13.80, 7.14, 3.57, and 75.48%, respectively (Fig. 1c). The 2D elemental mapping showed an abundant density of Mn and Zn with low-density regions in the 'Fe' map (respective high-density regions marked in 'Mn' and 'Zn' maps). The possible reason for such a distribution could be the formation of pure, binary, and ternary metal oxides in the nanocomposite (Fig. 1d). www.nature.com/scientificreports/ The PXRD pattern of MFZO nanocomposite has diffraction peaks for β-MnO 2 (purple circle) 23 , ZnO (pink circle) 24 , ferrites (green square) 25 , and NaNO 3 (blue square) 26 , which confirmed a poly-oxide nature of the composite (Fig. 2a). A similar report is available for the fabrication of Cu-Zn-Mn ternary oxide nanocomposite, where CuO, ZnO, and MnO 2 nanoparticles were confirmed 18 . The absence of a washing step during the MFZO fabrication was responsible for the presence of NaNO 3 in the sample. The N 2 adsorption-desorption isotherm of MFZO nanocomposite exhibited a Type III behaviour, generally expected in macro-porous materials (Fig. 2b) 23 . The nanocomposite possessed a BET surface area of 21.03 m 2 g −1 and a pore volume of 0.07 cm 3 g −1 . These values are higher than other nanocomposites like Mn 2 O 3 /Fe 2 O 3 (6.18 m 2 g −1 , 0.12 cm 3 g −1 ) 27 , CeO 2 /Mn 2 O 3 /Fe 2 O 3 (15.64 m 2 g −1 , 0.09 cm 3 g −1 ) 28 , and Fe 2 O 3 /Na 2 SO 4 (2.89 m 2 g −1 , 0.01 cm 3 g −1 ) 29 used for the same applications. The spectrum has a broad band centred at 621 cm −1 for the metal-oxygen stretching vibrations 30,31 . The bands at 835 and 1385 cm −1 were attributed to the asymmetric stretching vibration (ν 3 -NO 3 − ) and out-of-plane bending vibration (ν 2 -NO 3 − ), respectively, of nitrate 32 . The bands at 3433 and 1635 cm −1 were assigned to the stretching and bending vibration modes of adsorbed water molecules, respectively ( Fig. 2c) 33 . The XPS survey of MFZO confirmed peaks for Na 1 s, Zn 2p, Fe 2p, Mn 2p, O 1 s, and N 1 s at their respective binding energy. The Na 1 s and N 1 s peaks were associated with the presence of NaNO 3 (Fig. 2d).
The HRXPS Mn 2p 3/2 signal of MFZO deconvoluted into three contributions at 640.7, 641.6, and 642.8 eV for Mn 2+ (23.2%), Mn 3+ (43.9%), and Mn 4+ (32.9%) oxidation states of Mn ions, respectively 34 . The analyses showed that multivalent Mn ions were related to the formation of MnO 2 and Mn-based ferrites (Fig. 3a, Table S1). The HRXPS Zn 2p spectrum has two peaks at 1021.4 and 1044.3 eV for 2p 3/2 and 2p 1/2 signals of Zn 2+ ions, www.nature.com/scientificreports/ respectively (Fig. 3b, Table S2) 35 . In the HRXPS Fe 2p spectrum, the 2p 3/2 signal was deconvoluted into two contributions at 710.8 and 712.9 eV for Fe 2+ (70.4%) and Fe 3+ ions (29.6%), respectively (Fig. 3c, Table S3) 36 . These two contributions were due to the formation of ferrites. The HRXPS O 1 s spectrum has three deconvoluted peaks at 530.0, 531.4, and 532.9 eV for lattice oxygen (56.9%), surface hydroxyl groups/nitrate ions (24.6%), and water molecules (18.4%), respectively (Fig. 3d, Table S4) 37 . The synthesized nanocomposite was tested for H 2 S removal in breakthrough columns both in dry and wet conditions (Fig. 4a). The adsorption capacity of 0.73 mmol g −1 was achieved in the dry condition. However, in the wet condition, the capacity increased to 1.03 mmol g −1 , showing the positive role of water in the adsorption process mediated by the dissolution and dissociation of H 2 S molecules in the water film over the oxide surface. The effect of parameters, i.e., gas flow rate (Fig. 4b) and adsorbent loading (Fig. 4c) on the adsorption capacity was studied in the wet conditions. The adsorption capacity decreased with the increasing flow rate, where the highest capacity of 1.21 mmol g −1 was achieved with a flow rate of 0.1 L min −1 . Increasing flow rate disfavoured the adsorbate-adsorbent interaction due to a lowering in the gas retention time, which negatively impacted the capacity 38 . The adsorption capacity decreased with the increasing adsorbent loading and the maximum capacity of 1.31 mmol g −1 was achieved for 0.2 g of adsorbent and 0.2 L min −1 of flowrate. This behaviour could be associated with the presence of an unutilized mass of the oxide likely due to the cluttering of wet adsorbent particles in the adsorbent bed, which reduces the effective surface area for the reaction to occur 31 . However, no breakthrough experiment was conducted below 0.2 g. Since the material has a high density, loading adsorbent below 0.2 g led to a narrow bed length (since the tube diameter was 6 mm), which had a poor adsorbate-adsorbent interaction. The maximum adsorption capacity of 1.31 mmol g −1 achieved for the synthesized nanocomposite is similar to or higher than those reported for commercial ZnO (1.16 mmol g −1 ) 39  The nanocomposite was also studied for SO 2 adsorption in dry and wet conditions (Fig. 5a). The adsorption capacity of 0.22 mmol g −1 in the dry condition nearly doubled to 0.41 mmol g −1 in the wet condition. Such www.nature.com/scientificreports/ behaviour has been reported for oxidative SO 2 adsorption over MnO 2 16 . SO 2 adsorption in the presence of water molecules significantly accelerated the sulfate formation reaction, which favoured the overall adsorption process. The increasing gas flow rate negatively affected the adsorption process due to poor adsorbate-adsorbent interaction (Fig. 5b). The SO 2 adsorption capacity of the composite significantly improved with the decreasing adsorbent loading, where the maximum adsorption capacity of 0.49 mmol g −1 was confirmed with 0.2 g of adsorbent and 0.2 L min −1 of flowrate. Here also, the negative role of increasing bed loading was according to the effect witnessed for H 2 S gas adsorption (Fig. 5c). Thus, in the optimized experimental conditions, the adsorbent could remove 0.49 mmol g −1 of SO 2 gas. This value is highly significant and comparable to the reported values for ZnO (0.28 mmol g −1 ) 41 , MnO 2 (0.48-1.23 mmol g −1 ) 42 , and NaM x O y (0.73 mmol g −1 ) 7 in similar experimental conditions. The nanocomposite possessed a higher H 2 S adsorption capacity compared to the SO 2 gas. The superior H 2 S adsorption was related to an easy dissociation of H 2 S molecules due to the much lower energy barrier and higher adsorption energy compared to SO 2 , which has been previously demonstrated for Zn-MoSe 2 structure through computational calculations 43 .
The SEM and TEM micrographs post-H 2 S and SO 2 adsorption showed no significant variation in the surface morphology (Figs. S1, S2) except for the transformation of a chunk of nanoparticles to cluttered nanorods. It could be due to the combined effect of moisture and gas acidity as the SEM micrographs of dry samples showed no such change in the surface morphology. The EDAX analysis of gas-exposed samples confirmed a new peak at ~ 2.3 keV for sulfur. The intensity of the S peak in the H 2 S-adsorbed sample is much higher than that of the SO 2 , which agreed with the experimental results (Fig. S3). The 2D elemental mapping of the H 2 S and SO 2 -adsorbed samples confirmed a high density of sulfur atoms over the oxide surface, which was uniformly distributed over the nanocomposite (Fig. 6).
The PXRD patterns of gas-exposed samples showed insignificant changes in the diffraction peaks, except for new peaks in the H 2 S-adsorbed sample (marked as purple stars). These two peaks were assigned to the presence of ZnSO 3 44 . The absence of additional new peaks in these samples could be related to the formation of oxidized sulfur species on the surface (Fig. 7a). The N 2 adsorption-desorption isotherms are shown in Fig. 7b. For the H 2 S-adsorbed sample, the surface area and pore volume decreased by 26 and 17%, respectively (Table 1). However, for the SO 2 -exposed sample, a minimal drop in these values was observed. The drop in the surface area and porosity was linked to the deposition of oxidized sulfur species, which may have clogged the pores 29 . This www.nature.com/scientificreports/ clogging was expected more in the H 2 S-adsorbed sample due to a higher gas volume adsorption and subsequent mineralization onto the surface. More detailed information on the adsorption mechanism was deduced for the XPS analysis of MFZO nanocomposite after the gas adsorption process. In the HRXPS Mn 2p spectrum of the H 2 S-adsorbed sample, all three contributions for Mn 2+ , Mn 3+ , and Mn 4+ are present at a slightly lower binding energy with variation in the proportion of these oxidation species. The redshift in the binding energy could be associated with the partial sulfidation of the Mn oxides 31 . Moreover, the variation in the oxidation state proportions could be linked to the involvement of Mn 2+ /Mn 3+ /Mn 4+ redox cycles during the chemisorption process 17 . However, for SO 2 -adsorbed samples, the only proportion of oxidation states varied with an insignificant shift in the position, which was related to the Mn redox behaviour responsible for the oxidation of SO 2 (Fig. 7c) 17,45 . The HRXPS Zn 2p spectrum of H 2 S-adsorbed showed a minimal redshift in the peak position probably due to the formation of ZnSO 3 species. However, no such shift was witnessed for the SO 2 -adsorbed sample, which further suggested the delocalized nature of the chemisorption process (Fig. 7d). In the HRXPS Fe 2p spectrum of H 2 S-adsorbed MFZO, the peak position shifted slightly but with a minimal change in the proportion of Fe 2+ and Fe 3+ species. DFT calculations have predicted that H 2 S dissociatively reacts better on the FeO (Fe 2+ sites) than Fe 2 O 3 (Fe 3+ sites) 46 . Even in our previously reported work on the adsorption of H 2 S over Mn 2 O 3 /Fe 2 O 3 , bulk Fe 2 O 3 phase did not take part in the oxidation process 27 . The slight variation in the peak position could be linked to the involvement of Fe 2+ sites in the H 2 S adsorption process. For the SO 2 -adsorbed sample, the Fe 2+ and Fe 3+ peak positions red-shifted by 0.1 and 0.4 eV, respectively, with a significant drop in the Fe 2+ proportion (70.4-62.3%). It has been proven that the SO 2 molecules are much more reactive to the Fe 2+ sites than the Fe 3+ sites. Thus, a drop in the Fe 2+ contribution suggested that the divalent Fe sites catered the oxidation of SO 2 molecules (Fig. 7e) 47 . In the HRXPS O 1 s spectrum of H 2 S-adsorbed sample, the metal-oxygen bond contribution decreased, whereas the contribution at 531.6 eV for −OH/O-N increased due to the formation of metal-sulfide and sulfite (SO 3 − ) species, respectively. For the SO 2 -adsorbed sample, the 531.5 eV peak improved even further due to the consumption of lattice oxygen and the formation of sulfate species (Fig. 7f).
The HRXPS S 2p spectrum of H 2 S-adsorbed MFZO was deconvoluted into three sets of doublets with their 2p 3/2 peaks observed at 161.3, 163.6, and 167.9 eV for sulfide (36.1%), elemental sulfur (25.1%), and sulfite (38.8%) 48 . While the formation of metal-bound sulfide is initiated by the dissociated adsorption of H 2 S (into H + and HS − ) in the presence of water molecules. The formation of elemental sulfur and sulfide is mediated by    49 . The HRXPS S 2p spectrum of SO 2 -adsorbed MFZO has a set of doublets with a 2p 3/2 peak at 168.4 eV, which was assigned to the sulfate species (Fig. 8, Table S5) 48 . The adsorption of SO 2 over the oxide surface is generally driven by the reactive interaction of SO 2 molecules with the lattice oxygen or surface hydroxyl groups to form sulfite/bisulfite, which further oxidized to sulfate via redox behaviour of transition metal oxide and gaseous oxygen molecules 7,17 . Moreover, just like H 2 S dissolution in the surface water, SO 2 could be readily adsorbed and hydrolysed by surface water molecules, which makes the oxidation of SO 2 molecules, energetically favourable 17 .

Conclusion
In conclusion, we have fabricated an Mn-Zn-Fe metal oxide nanocomposite via a one-step coprecipitation reaction. The fabricated nanocomposite has MnO 2 , ZnO, and ferrites with a surface area and pore volume of 21.03 m 2 g −1 and 0.07 cm 3 g −1 , respectively. The nanocomposite was tested for room-temperature adsorptive removal of H 2 S and SO 2 in dry and wet conditions. The oxide exhibited better gas adsorption capacity in wet conditions owing to the dissolution and dissociation of gaseous molecules in the surface water film. The adsorbent showed a better adsorption capacity at a lower flow rate and adsorbent loading. In the optimized conditions, a maximum of 1.31 and 0.49 mmol g −1 of H 2 S and SO 2 was removed by the nanocomposite, respectively. The indepth spectroscopic analysis confirmed the mineralization of H 2 S gas into sulfide, sulfur, and sulfite, which was mediated by the Fe and Mn redox cycles in the presence of adsorbed water and molecular oxygen. Though Zn ions did not participate in the oxidation process, Zn 2+ probably interacted with the sulfides and sulfites. The SO 2 mineralization was associated with the formation of sulfates, driven by the redox behaviour of Fe and Mn in an oxidative environment. Thus, we have presented a novel adsorbent material for the successful mineralization of toxic sulfurous gases, which could be suitable for deep desulfurization applications.

Data availability
Data is available from the corresponding author after reasonable request.