Near-Infrared-Driven Selective Photocatalytic Removal of Ammonia Based on Valence Band Recognition of an α-MnO2/N-Doped Graphene Hybrid Catalyst.

Near-infrared (NIR)-response photocatalysts are desired to make use of 44% NIR solar irradiation. A flower-like α-MnO2/N-doped graphene (NG) hybrid catalyst was synthesized and characterized by X-ray diffraction spectroscopy, transmission electron microscopy, Raman spectroscopy, UV-vis-NIR diffuse reflectance spectroscopy, and X-ray photoelectron spectroscopy. The flower-like material of α-MnO2/NG was oval-shaped with the semi major axis of 140 nm and semi minor axis of 95 nm and the petal thickness of 3.5-8.0 nm. The indirect band gap was measured to be 1.16 eV, which is very close to 0.909 eV estimated by the first-principles calculation. The band gap can harvest NIR irradiation to 1069 nm. The coupling of α-MnO2 with NG sheets to form α-MnO2/NG can significantly extend the spectrum response up to 1722 nm, improving dramatically the photocatalytic activity. The experimental results displayed that the α-MnO2/NG hybrid catalyst can recognize ammonia in methyl orange (MO)-ammonia, rhodamine B (RHB)-ammonia, and humic acid-ammonia mixed solutions and selectively degrade ammonia. The degradation ratio of ammonia reached over 93.0% upon NIR light irradiation in the mixed solutions, while those of MO, RHB, and humic acid were only 9.7, 9.4, and 15.7%, respectively. The products formed during the photocatalytic process were followed with ion chromatography, gas chromatography, and electrochemistry. The formed nitrogen gas has been identified during the photocatalytic process. A valence band recognition model was suggested based on the selective degradation of ammonia via α-MnO2/NG.

Introduction

Near-infrared (NIR) light (44%) and visible light (48%) account for 92% of solar irradiation spectrum. Therefore, it is urgently desired to develop NIR-response photocatalysts to make full use of solar spectrum. Efforts have been paid to find out NIR-response photocatalysts, in which up-conversion materials composed of rare earths,[1−6] carbon quantum dots,[7−10] and plasmon composites composed of Au and Ag[6,11−13] have been synthesized. The efforts inspired scientists to develop novel NIR-response photocatalysts for thorough usefulness of solar irradiation.

It is known that α-MnO2 is a semiconductor material with various band gaps with the size of particles or sheet thickness. The band gap depends on the size and structure of this material. The bulk MnO2 materials are usually regarded as indirect semiconductors with very small band-gap energies of about 0.20–0.37 eV,[14,15] while an ultralong nanowire α-MnO2 with the length of more than 10 mm and width of 30 nm produced a band gap of 0.98 eV.[16] One of α-MnO2 nanofibers with typical diameters of 20–60 nm and lengths of 1–6 μm formed a band gap of 1.32 eV;[17] another ultrathin MnO2 of nanosheets with the thickness of about 0.5 nm achieved a band gap of about 2.23 eV.[18] The shift of band gap to higher energies is attributed to the carrier confinement in small semiconductor particles.[16] Obviously, the band gap of α-MnO2 can been tailored by the size of the material to fulfill the goal of wide spectrum absorption and full use of solar irradiation.

In addition, the surface modification such as copper phosphide decorated on g-C3N4 nanosheets,[19] Ag on Bi4Ti3O12 and CuBi2O4, improved markedly the photocatalytic activity.[20,21] CdSe decorated N-doped TiO2 to boom the generation of H2.[22] N-doped graphene (NG) can enhance the photocatalytic activity of photocatalysts because of the absorption of incident light, fast electron transportation from semiconductors to NG, and the reduced recombination of the photogenerated electron–hole pairs[23,24] when it was coupled with semiconductors. Therefore, the as-coupled product, α-MnO2/NG, was expected to be more active to NIR irradiation.

Selective photocatalysis is a very interesting research from organic synthesis to environmental applications.[25−30] It can synthesize specific target organic molecules or degrade specific target pollutants. These selective photocatalytic reactions are based on the interaction forces such as van der Waals and coordination between hosts and guests. However, the potential recognition based on the valence band (Ev) level has not been reported. Herein, we reported the NIR-driven selective degradation of ammonia via the α-MnO2/NG nanophotocatalyst, and a valence band recognition model was achieved. It is expected to highlight NIR photocatalysis.

Experimental and Calculations

Chemicals

Potassium permanganate (KMnO4) was purchased from Shanghai General Chemical Reagent Plant, and manganese sulfate monohydrate (MnSO4·H2O) was purchased from Sinopharm Chemical Reagent Co., Ltd. Graphite powder (average particle size = 30 μm) was purchased from Shanghai Colloid Chemical Plant, and sodium hydroxide (NaOH) was purchased from Tianjin Damao Chemical Factory. Ammonium sulfate ((NH4)2SO4), ammonium chloride (NH4Cl), rhodamine B (RHB), and methyl orange (MO) were obtained from Nanjing Chemical Reagent Co., Ltd. All reagents were of analytical grade and used without further purification. All solutions were prepared with 18.2 MΩ·cm deionized Milli-Q water.

Synthesis of NG

NG sheets were prepared via a one-pot hydrothermal reaction of graphene oxide (GO) with urea at 170 °C for 12 h.[31] GO was obtained by a modified Hummers method as we reported previously.[32,33] The general procedures are as follows. First, natural graphite powders were oxidized with concentrated KMnO4 in the presence of H2SO4; then, the excess of KMnO4 was reduced with H2O2 and exfoliated by sonication. After GO was obtained, nitrogen insertion into graphene frameworks using urea was achieved by a facile hydrothermal reaction at 170 °C for 12 h. The nitrogen content could be controlled by the mass ratio of GO to urea. A typical mass ratio (GO/urea) used in our case was 1:300 to obtain an optimal nitrogen content. Eighty milligrams (80 mg) of GO was dispersed into 80 mL of water with sonication, and then 24.0 g of urea was introduced in the 80 mL suspension solution for continuing sonication for 2 h. The mixture was transferred to a Teflon-lined stainless steel autoclave with a volume of 100 mL, sealed, and heated to 170 °C for 12 h to form NG (nitrogen content: ca. 7%).[31] Finally, the resulting product was filtered, washed, and dried at 60 °C in a vacuum chamber.

Synthesis of α-MnO2/NG

The α-MnO2/NG catalyst with a loading of 1.0 wt % NG was synthesized as follows. First, 6 mg of NG was dispersed into 20 mL of water with sonication, and 0.2672 g (1.35 mmol) of MnCl2·4H2O was added into the suspension solution under sonication for 6 h. Then, the suspension solution was transferred into a 50 mL three-necked flask, and then it was heated to 85 °C under magnetic stirring. Finally, 0.1422 g (0.9 mmol) of KMnO4 was dissolved in 5 mL of deionized water, and the KMnO4 solution was titrated into a three-necked flask at 85 °C under continuous stirring. This temperature was maintained for 12 h to form α-MnO2/NG. After that, it was cooled to room temperature, and the precipitates were collected, washed at least for three times, and dried at 60 °C in a vacuum chamber for 8 h. The resulting product was denoted as α-MnO2/NG. Similarly, α-MnO2/reduced graphene oxide (rGO) and pure α-MnO2 samples were synthesized for comparison. The reaction formula was denoted as follows:

Structure Characterization and Mechanism Investigation

X-ray diffraction (XRD) measurements were performed using an X-ray diffractometer (D/Max 2500 PC). The X-ray source was Cu Kα radiation with a wavelength of 0.154 nm, a tube voltage of 40 kV, and a tube current of 40 mA. α-MnO2/NG and α-MnO2/rGO powders were dispersed in water using an ultrasonic device, placed on carbon-coated copper grids, and dried under ambient conditions for morphological observations using a transmission electron microscopy (TEM) system (Tecnai G220, FEI, USA). UV–vis–NIR diffuse reflectance spectra (UV–vis–NIR DRS) were recorded on a Shimadzu UV–vis–NIR spectrometer (UV-3600 plus, Japan), and the wavelength scope is 190–3400 nm. An X-ray photoelectron spectrometer (ESCALAB 50 XI) was utilized to characterize ammonia adsorbed on α-MnO2 samples and chemical surrounding of Mn in α-MnO2 samples before and after ammonia was adsorbed to elucidate the molecular recognition mechanism. The binding energy (Eb) scale was calibrated with respect to the carbon 1s peak at 284.60 eV. The aluminum Kα 1.2 line (hν 1486.60 eV) was used as the X-ray excitation source, and each powder sample was dispersed on a gold-plated copper surface. The spectra were recorded in a fixed analyzer transmission mode to achieve maximum instrumental resolution. The instrument was operated under a vacuum of 1 × 10–9 Torr in the analysis chamber. Wide and high-resolution spectra were recorded at a constant pass energy of 50 eV and channel widths of 1.0 and 0.1 eV. The α-MnO2 samples were collected before and after adsorbing ammonia to detect the Eb of N and Mn atoms on the catalyst.

Cyclic voltammetry was performed using a CHI660C electrochemical workstation (CH Instrument Company, TX, USA). The electrochemical experiments were conducted in a conventional three-electrode system. A ring-disk platinum electrode was used as the working electrode, a platinum foil as the counter electrode, and a saturated calomel electrode as the reference electrode. In order to obtain the stable cyclic voltammograms of oxygen dissolved, the working electrode was scanned between 0.0 and −1.4 V for 10 cycles before the measurement. The voltammetric measurements were carried out in a sealed container to prevent O2 from air.

A Dionex (Dionex, Sunnyvale, CA, USA) model ICS-900 ion chromatograph was used for separation and detection of nitrite and nitrate. The analytical column was Dionex IonPac AS14A (250 mm × 4 mm, i.d.), and the guard column was IonPac AG14A (50 mm × 4 mm, i.d.). The eluent was a mixture of 4.5 mmol/L sodium carbonate and 0.8 mmol/L sodium bicarbonate, the eluent flow rate was set at 1.2 mL/min, the column temperature was 25 °C, and the injection volume was 2 mL. A conductivity detector CD produced by the company Dionex (USA) and conductivity suppressor ASRS-300 4 mm (Dionex, USA) were used for the detection of anions.

In order to investigate the mechanism of photocatalysis, a sealed photocatalytic reaction system (Labsolar 6A photocatalytic system, Perfect light Co. Ltd., Beijing, China) was used, in which a 100 mL aqueous solution containing 100.0 mg/L NH3–N was irradiated under NIR light and a 465 mL gas with an initial ratio of oxygen to argon at 1:3 was cycled and the initial internal pressure was maintained at 26 kPa. The sealed system was vacuumed before the mixed gas of oxygen and argon was injected. The reaction system was matched with a GC-7806 gas chromatograph (Shiwei Puxin Instruments Co. Ltd., Beijing, China). The chromatographic column of 5 m length × 2 mm i.d., filled with 5 Å molecular sieve, was utilized as a separating column. High-purity argon was utilized as the carrier gas. The flow rate of the carrier gas was set at 23 mL/min, and the thermal conductivity tank was used to detect N2 and O2. The column temperature was set at 80 °C, the inlet temperature at 120 °C, and the detector temperature at 150 °C.

Selective Degradation of Ammonia

Photocatalytic experiments for degradation of ammonia were conducted under NIR light irradiation (λ > 780 nm). A 300 W UV–visible lamp (OSRAM, Germany) was used as the light source. Briefly, 50 mL of ammonia solution was subjected to photocatalytic degradation in a 100 mL beaker at room temperature (25 ± 2 °C). The distance between the lamp and the test solution was approximately 10 cm. The wall of the beaker was protected from the surrounding light with an aluminum foil. NIR light was obtained with a cutoff filter of λ > 780 nm (see the picture of the cutoff filter in Figure S1 in the Supporting Information). The filter covered the window of the beaker to absorb UV–vis light and allow NIR light of λ > 780 nm to pass through. Typically, 50 mL of test solutions was used in the degradation experiments. A specific concentration of ammonia–organic compound solution was prepared according to the desired concentration. Briefly, 0.10 g of the catalysts was used for photocatalysis. NaHCO3–Na2CO3 buffer was used to control the pH of the solutions tested.

A double-beam TU-1901 spectrophotometer was used to determine the ammonia concentration by using the Nessler reagent during photocatalysis. The Nessler reagent [alkaline solution of dipotassium tetraiodomercurate(II)] was prepared according to the following procedures. Briefly, 10 g of HgI2 and 7 g of KI were dissolved in water, NaOH solution (16 g of NaOH in 50 mL of water) was added to the mixture, and the mixture was diluted to 100 mL by adding deionized water. The Nessler reagent was stored in a dark bottle and diluted properly before analysis. Ammonia reacts with the Nessler reagent to obtain colored solutions by the reaction given in the following. The absorbance at 392 nm approached the peak and was recorded at the wavelength of 392 nm. The decline of the peak indicated the degradation of ammonia during the photocatalytic process.

The products degraded were identified by ion chromatography, gas chromatography, electrochemistry, and UV–vis spectroscopy during the photocatalytic process (see Figures S5–S7 in the Supporting Information).

First-Principles Calculation

The calculations of band structure and density of states (DOS) were performed by using the CASTEP package of Material Studio 7.0. In the plane-wave calculations, a cutoff energy of 340 eV was applied. The generalized gradient approximation was adopted with Perdew–Burke–Ernzerhof functional. The calculations were performed in a ferromagnetic spin-polarized configuration. The self-consistent field convergence criterion and energy tolerance were set as fine levels. A Monkhorst–Pack scheme with a 1 × 1 × 5 k-point grid was employed. The Taylor series method was used for dispersion-corrected density functional theory correction. The α-MnO2 phase is of a tetragonal structure with the space group 14/m (no. 87) and the lattice parameters a = 9.7980, c = 2.8513 Å, and α = β = γ = 90°, in which the Mn and O atoms occupy the fractional coordinates of Mn (0.35, 0.17, 0.00), O1 (0.16, 0.205, 0.00), and O2 (0.16, 0.458, 0.00). These accurate data validated the applicability of the CASTEP package for calculating the α-MnO2 structures.

Results and Discussion

XRD Characterization

The as-synthesized samples were determined by XRD spectroscopy, and the resulting patterns are shown in Figure A. It can be seen from Figure that the diffraction peaks appeared at 2θ = 12.8°, 18.0°, 28.8°, 37.4°, 50.0°, 56.4°, 60.2°, 65.0°, and 69.7°, which are indexed to a pure tetragonal phase of α-MnO2 (JCPDS, no. 41-0141) with I4/m space group and lattice constants a = 9.7980 and c = 2.8513 Å, corresponding to the Bragg reflections of (110), (220), (310), (211), (301), (411), (600), (521), (002), and (541) planes, respectively.[34−37] There are one-dimensional 2 × 2 (4.6 Å × 4.6 Å) and 1 × 1 (1.9 Å × 1.9 Å) tunnels intergrowth along the c axis in the lattice shown in Figure B.[37−40] No characteristic peaks are observed for the impurities, indicating high purity and crystallinity of the as-synthesized samples. In addition, the diffraction peak at 26.31° was indexed to the plane (102) of rGO and NG, indicating that rGO and NG were formed during the hydrothermal process. As aforementioned, the 2 × 2 tunnels constructed of octahedral [MnO6] have a 4.6 Å × 4.6 Å space, which enables to accommodate cations. The tunnel structure makes it attractive to be used as photocatalysts because the photocatalytic process is a redox one involving electron transfer and associated cation shift to maintain the electrical neutrality equilibrium.

Figure 1

(A) XRD patterns of the samples: (a) α-MnO2/NG; (b) α-MnO2/rGO; (c) NG; (d) rGO; and (e) α-MnO2. (B) Crystal structure of α-MnO2 using the [MnO6] model demonstrating [1 × 1] and [2 × 2] tunnels. The purple balls denote Mn atoms and red balls O atoms.

Raman Spectrum Characterization

To investigate the vibrational properties of the as-synthesized samples, the characterization of Raman spectroscopy has been conducted. Figure presents the Raman spectra of α-MnO2/NG (a), α-MnO2/rGO (b) composite, and α-MnO2 (c), rGO (d), and NG (e) components. The vibration peaks at 646 cm–1 were assigned to the Ag mode originating from breathing vibrations of the MnO6 octahedra, indicating the α-MnO2 crystallite.[41] This observation is in well-agreement with the XRD results. The G band at 1580 cm–1 in the rGO sample is characteristic of sp2-hybridized C–C bonds in a two-dimensional hexagonal lattice, and the band shifted to 1600 cm–1 after nitrogen atoms were doped in the lattice as shown in the α-MnO2/NG (a) and NG (e) samples. The D band located at 1335 cm–1 corresponds to the defects and disordered carbon in the graphitic layers.[42] In addition, the broader 2D peaks appeared at around 2645 and 2905 cm–1, which is consistent with that of the few-layer NG reported by Yu.[43]

Figure 2

Raman spectra of the as-synthesized samples: (a) α-MnO2/NG; (b) α-MnO2/rGO; (c) α-MnO2; (d) rGO; and (e) NG.

TEM Observation

Figure A,B shows the representative TEM images of the obtained α-MnO2/NG samples. Dispersive nanoflowers can be observed throughout the samples. Most of the flowers are oval-shaped with the semi major axis of 140 nm and semi minor axis of 95 nm, as shown in Figure B. The thickness of prevailing petals is among 3.5–8.0 nm, and this nanostructure favors to adsorb the randomly approaching molecules to decline the surface tension. High-resolution TEM image (Figure C) indicates the polycrystalline feature of α-MnO2 in the as-synthesized α-MnO2/NG sample. A finger distance of 0.50 nm corresponds to the d-spacing value of the plane (200) in α-MnO2,[44] which is in agreement with the above XRD analysis. The other finger distance adjacent to α-MnO2 was found to be 0.34 nm in Figure C, which is the same as the d-spacing value of the plane (002) in pure NG sheets in Figure D. It also shows the compact alignment between α-MnO2 and NG.

Figure 3

TEM images of the as-synthesized samples: (A,B) α-MnO2/NG; (C) high-resolution TEM image; and (D) pure NG.

XPS) Characterization

XPS measurements of the as-synthesized α-MnO2/NG samples have been carried out to further confirm the structure and composition. The as-measured results are presented in Figure . The total survey revealed the presence of Mn, O, and N atoms in the as-synthesized α-MnO2/NG sample, as shown in Figure A. High-resolution XPS pattern indicated two peaks of 642.27 and 654.07 eV (Figure B), which are attributed to the binding energy of Mn 2p3/2 and Mn 2p1/2, respectively. The spin energy separation of 11.80 eV is in very good accordance with the reported datum on MnO2, confirming the presence of α-MnO2,[40,45] which is also consistent with the XRD result as reported previously. The XPS spectrum of O 1s binding energy is shown in Figure C, and the two peaks centered at 530.1 and 531.29 eV corresponded to Mn–O–Mn and Mn–O–H, respectively, in the α-MnO2/NG composite. The presence of Mn–O–H indicates that oxygen atoms linked to some of Mn atoms in the MnO6 octahedra may interact with hydrogen atoms via a hydrogen bond or a covalent coordination bond because water and related hydrides (H3O+) can also be accommodated in the tunnels.[40,46] Three different types of nitrogen configurations (Figure D), including pyridinic N (397.13 eV), pyrrolic N (399.75 eV), and quaternary N (graphitic N, 401.68 eV) revealed that nitrogen atoms were bonded in the graphene framework.[47]

Figure 4

XPS spectra of the as-synthesized samples: (A) wide-scan XPS spectra of α-MnO2/NG (a) and α-MnO2 (b); (B) Mn 2p XPS spectra in α-MnO2/NG (a) and α-MnO2 (b); (C) O 1s XPS spectra in α-MnO2/NG (a) and α-MnO2 (b); and (D) N 1s XPS spectra in α-MnO2/NG.

UV–Visible–NIR Reflection Spectra

Figure shows the UV–visible–NIR reflection spectra of the as-synthesized α-MnO2 (a), α-MnO2/rGO (b), and α-MnO2/NG (c) species. From the spectra, the absorption peaks were observed at 380 nm in the UV region, and an absorption band for the pure α-MnO2 species was observed in the visible and NIR region as shown in curve (a). The absorption peak in the NIR region is assigned to the electron transition derived from the mixing of metal d–d and oxo-to-metal charge transfer,[48−52] which is also confirmed by our calculation using the CASTEP method below. On the basis of the UV–visible–NIR reflection spectra, the direct and indirect optical band gaps of the as-synthesized α-MnO2 (a), α-MnO2/rGO (b), and α-MnO2/NG (c) species can be determined by Tauc’s relationwhere A is the constant, hν is the photon energy, and α is the absorption coefficient, while n = 2 for direct and n = 1/2 for indirect transition. It has been known that α-MnO2 is an indirect semiconductor.[46] Thereby, the indirect band-gap values are given, as shown in Figure , which are equal to 1.16, 1.00, and 0.72 eV for α-MnO2, α-MnO2/rGO, and α-MnO2/NG, respectively. The indirect band gap for pure α-MnO2 is 1.16 eV, which is very close to 0.98 eV reported in the literature.[16] The transition corresponds to the NIR wavelength of 1069 nm, which provides a profound basis for using NIR irradiation as power for photocatalytic applications. The band gap is further reduced to 0.72 eV after it was combined with NG as shown in curve (c) in Figure , implying that the α-MnO2/NG species can absorb more NIR irradiation as a photocatalyst for a more effective photocatalysis. As predicted, the combination of rGO with α-MnO2 led to a narrow indirect band gap, which is equal to 1.00 eV. These facts prove the strong interaction between α-MnO2 and NG or rGO.

Figure 5

UV–vis–NIR diffuse reflection spectra of the as-synthesized samples: (a) α-MnO2; (b) α-MnO2/rGO; and (c) α-MnO2/NG.

Figure 6

Tauc plot of (αhν)1/2 vs photo energy for the indirect band gaps of the samples: (a) α-MnO2; (b) α-MnO2/rGO; and (c) α-MnO2/NG.

Band Structure and DOS

The band structure (Figure A) and total DOS and partial DOS (PDOS) of α-MnO2 (Figure B) were calculated by the first-principles method, resulting in an indirect band gap of 0.909 eV, which is close to 1.160 eV determined by our experimental measurements, which is also in very good agreement with that calculated previously.[16,46,53] As seen from Figure B, the total DOS are provided by s, p, and d orbitals. Obviously, 3d orbitals in Mn atoms (Figure C) were split into the low energetic level of t2g and high energetic level of eg in the MnO6 octahedral field[46,48−52] and contributed to the valence band and conduction band (Ec), together with p orbitals provided by O atoms (Figure D). This band structure of α-MnO2 favors the absorption of incident NIR irradiation.

Figure 7

Band structures of α-MnO2 along the Z−Γ–X–P–N−Γ lines (A). Orange red and blue curves denote up- and down-spin energy levels at each k-point, respectively. Total DOS and PDOS of α-MnO2 (B). PDOS of Mn atoms (C). PDOS of O atoms (D). The energies are aligned to the Fermi level marked by a vertical dashed line.

Selective Degradation

As seen in Figure S2 in the Supporting Information, the α-MnO2/NG catalyst can degrade ammonia under NIR light irradiation, and Figure S3 shows 3.0% of the NG content reaching the maximum degradation ratio (Figure S3). Given the coexistence of ammonia together with organic pollutant in practical waters, RHB and MO were chosen as organic model pollutants to examine the selectivity of the α-MnO2/NG catalyst. The results indicated that the degradation ratio of ammonia reached over 93.0% under NIR light irradiation in NH3–MO mixed solution or NH3–RHB mixed one as shown in Figure , while the ratios were only 9.7 and 9.4% for MO and RHB, respectively. The degradation ratio of ammonia is much higher than that of MO or RHB under the same conditions, indicating that the α-MnO2/NG photocatalyst can recognize and degrade ammonia selectively from MO or RHB in their coexistence. Under the similar conditions, humic acid was used to examine the selectivity of the α-MnO2/NG photocatalyst, and the results indicated that 15.7% of humic acid was degraded for 7.0 h (Figure S4), while over 93.0% of ammonia was eliminated in 50.0 mg/L humic acid and 50.0 mg/L NH3–N mixed solution. It confirmed further the selectivity of the photocatalyst for ammonia.

Figure 8

Selective degradation of ammonia in the presence of 50 mg/L NH3–N (a) plus 50 mg/L MO (b) in 50 mL solution under NIR light irradiation and in the presence of 50 mg/L NH3–N (c) plus 50 mg/L RHB (d) in 50 mL solution under NIR light irradiation.

The subsequent investigations indicated the adsorption of ammonia, MO, and RHB on the catalyst; the adsorption ratios were 17.6 and 2.1% for ammonia and MO, respectively, in NH3–MO mixed solution in dark for 8 h and 17.1 and 3.1% for ammonia and RHB, respectively, in NH3–RHB mixed solution (Figure ).

Figure 9

Adsorption of NH3, MO, and RHB on the α-MnO2/NG catalyst in 50 mL solutions in the coexistence of 50 mg/L NH3–N (a) plus 50 mg/L MO (b) and in the coexistence of 50 mg/L NH3–N (c) plus 50 mg/L RHB (d) for 8 h in dark. The mass of the α-MnO2/NG catalyst was 0.10 g, and pHs were adjusted to 10.50 by adding 0.1 mol/L NaHCO3–Na2CO3 buffer.

XPS measurements showed the adsorption of NH3, MO, and RHB molecules on the catalyst with the same shift of 0.2 eV at Mn 2p3/2 (Figure ), implying a nonspecific adsorption. That is, the adsorption is a physical one but not a chemical one. In other words, the selective degradation of ammonia using α-MnO2/NG as the photocatalyst in NH3–MO or NH3–RHB mixed solution under NIR irradiation could be attributed to another reason.

Figure 10

Binding energies of Mn 2p in α-MnO2 samples: (a) pure α-MnO2; (b) after immersing in 50 mg/L NH3–N solution for 8 h; (c) after immersing in 50 mg/L MO solution for 8 h; and (d) after immersing in 50 mg/L RHB solution for 8 h.

Insights of Selectivity and Reaction Mechanism

As estimated by experiments, the indirect band gap of α-MnO2/NG is 0.72 eV, which corresponds to the energetic difference between valance band and conduction band. Figure presents the valence band XPS spectrum of the α-MnO2 species. The top of valance band is the highest binding energy of the filled electrons and is defined as the intercept at 0.00 eV on the background. Therefore, the valence band of α-MnO2/NG can be obtained as 1.44 eV based on the XPS measurement because the valence band is mainly composed of O 2p and Mn 3d orbitals as seen in Figure B. The conduction band can be estimated as 0.72 eV because the band gap is 0.72 eV.

Figure 11

Valence band measurement of α-MnO2 by XPS spectrum.

As far as photocatalytic reactions are concerned, they involve the oxidization process driven by photogenerated holes and the reduction process driven by photogenerated electrons. In a viewpoint of electrochemistry, the photogenerated holes can be regarded as oxidizing agents for the oxidization process, while the photogenerated electrons can be regarded as reducing agents for the reduction process. The photogenerated holes could oxidize the species collided only when the holes on top of the valence band are more positive than the redox potential of the species collided (reducing agents). It has been known that the electrode potential for E0(N2/NH3) is equal to 0.057 V versus normal hydrogen electrode (NHE), whereas the anodic peak potential is equal to about 1.18 V versus NHE for the oxidization of RHB[54,55] and 1.497 V versus NHE for the oxidization of MO.[56] Under consideration of over-potentials, the oxidization of ammonia to dinitrogen is much easier than that of MO or RHB to their oxidized states. Therefore, the preference of the valence band of α-MnO2 resulted in the selective oxidization of ammonia from MO or RHB in solutions.

For the reduction process, the photogenerated electrons could reduce the species collided (oxidizing agents) only when the electrons excited on the bottom of the conduction band are more negative than the reduced potential of the species collided. The Fermi level for the perfect graphene is just located at the so-called Dirac point. For the NG, the Fermi level is shifted up into the conduction band, which is due to the electron donation from the doped-nitrogen atom to the antibonding π* band.[57,58] Graphitic nitrogen atoms in NG resulted in shifting up with the Fermi energy level of 0.66 V versus NHE,[59] that is, the conduction band level of NG is 0.66 V versus NHE. The level (0.66 V) of conduction band is much negative than that of 1.23 V versus NHE, which shows the potential of oxygen being reduced to water. Therefore, oxygen can accept photoelectrons from the conduction band of NG to form water. Furthermore, the band gap of NG is in the range of 0.045–0.550 V, depending on the nitrogen-doped content from 0.5 to 8.0%.[60] In the case here (ca. 7.0%), the band gap is about 0.48 eV, so the valence band level of NG is 1.14 V. The valence band level of NG could accept the photogenerated electrons transferred from the conduction band of α-MnO2. The band structure of NG was combined with that of α-MnO2, establishing the Z-scheme photocatalytic system for the removal of ammonia.[61,62] Therefore, the overall mechanism for selective oxidization of ammonia can be presented in Figure . The product has been identified as N2 gas by means of UV–vis spectroscopy, electrochemistry, ion chromatography, and gas chromatography (see Figures S5–S7 in the Supporting Information). It is concluded that the matched band structure of α-MnO2/NG oxidized ammonia selectively.

Figure 12

NIR-driven selective photocatalytic degradation of ammonia via the α-MnO2/NG catalyst.

Conclusions

The α-MnO2/NG hybrid catalyst is synthesized via a one-pot method. The theoretic band gap (0.909 eV) based on the first-principles calculation is very close to that (1.16 eV) measured by UV–vis–NIR diffuse reflectance spectroscopy, indicating that the theoretic simulation approaches to the real band structure of α-MnO2. The analysis based on first-principles calculation reveals that the valence band is composed of hybrid O 2p and Mn 3d orbitals and that the conduction band is composed of Mn 3d and O 2p orbitals. The band structure of α-MnO2 results in the response to NIR irradiation, which makes it as a NIR photocatalyst. The combination of α-MnO2 with NG to form α-MnO2/NG significantly extends the spectrum response to 1722 nm. The results show that α-MnO2/NG as a photocatalyst can make full use of NIR irradiation. α-MnO2/NG can degrade ammonia selectively in a mixed solution of ammonia and MO or RHB or humic acid under NIR light irradiation to release nitrogen gas, which is attributed to the matched band structure of α-MnO2/NG and the valence band level of α-MnO2.