Conversion of methane to acetonitrile over GaN catalysts derived from gallium nitrate hydrate co-pyrolyzed with melamine, melem, or g-C₃N₄: the influence of nitrogen precursors

Abstract

Co-pyrolyzing gallium nitrate hydrate and melamine, melem, or g-C₃N₄ generates gallium nitride (GaN) for the conversion of methane to acetonitrile (AcCN). The solid-state-pyrolysis-made GaN catalysts exhibited better activity than commercial GaN. Among the as-prepared catalysts, GaN made by using g-C₃N₄ with a N/Ga ratio of 2 (i.e., GaN-(C₃N₄)-(2)) was the most attractive: a high initial methane conversion (28.2%), a high initial AcCN productivity (151 μmol g_cat⁻¹ min⁻¹), and a 6 h accumulated AcCN yield (5816 μmol g_cat⁻¹) were obtained at 700 °C with a space time of 3000 mL_CH4 g_cat⁻¹ h⁻¹. GaN-(C₃N₄)-(2) had finely dispersed GaN crystals and enriched amorphous CN species (e.g., sp² N and C≡N groups), and both are important in promoting the methane conversion rate. GaN agglomeration, coke deposition, and depleted CN species contributed to the deactivation of the catalyst, and a nitridation–activation process could rejuvenate the activity partially. The analysis of the structure–activity correlation revealed that the accumulated AcCN yield had an inverse trend with respect to the crystallite size of GaN and the sp³/sp² ratio of the N environment.

---

1. Introduction

Methane is a greenhouse gas with a global warming potential approximately 30 times higher than that of CO₂. Approximately 60% of methane emissions are related to anthropogenic behavior,¹ and a 5% growth rate of annual emissions is forecasted.² Methane is mainly used as fuel, with about 15% used as a chemical feedstock, mostly in steam reforming to produce syngas.³ Owing to the increasing economically feasible methane reserves after the shale gas revolution⁴ and the growing concern over methane emission/leakage from industry, methane utilization in chemical synthesis has been refocused on.

Direct conversion of methane to chemicals is usually categorized into non-oxidative (e.g., methane dehydroaromatization) and oxidative routes (e.g., methane oxidative coupling), which transform methane into olefins and aromatics.⁵ Other methods such as halogenation of methane (e.g., methane bromination into methyl bromide^6–8) and sulfonation to methanesulfonic acid have been reported.⁹ However, most of these approaches are uneconomical due to low methane conversions, high activation temperatures, and demanding separation steps.¹⁰ This has stimulated recent studies on methane activation through photo-,¹¹ electro-,¹² electro-thermal,¹³ and novel heterogeneous catalysis (e.g., tandem¹⁴ and single-atom catalysts¹⁵).

Acetonitrile (AcCN) is an important chemical mostly used as a solvent in organic synthesis and in separation/purification. The major use of AcCN in the petrochemical industry is to extract butadiene from crude C₄ streams in extractive distillation.¹⁶ Currently, AcCN is a byproduct obtained from the ammoxidation of propylene to acrylonitrile (Sohio process).¹⁷ Transforming light alkanes to AcCN has gained considerable attention because of the low cost and abundance of alkanes. Conversion of ethane and ethylene to AcCN by ammoxidation has been reported.^18–21 Still, the growing demand for AcCN stimulates researchers to develop an alternative route for AcCN production from light alkanes. Our group recently discovered that methane conversion to AcCN can be achieved through gallium nitride (GaN)-based catalysts.²²

GaN has been applied in the catalysis of alkane conversion, such as methane conversion to ethylene²³ and aromatics,²⁴ oxidative dehydrogenation of n-butane,²⁵ and dehydrogenation of propane.²⁶ GaN is more active than its oxide form (β-Ga₂O₃). Moreover, the effects of the particle size and purity of GaN had a strong correlation with the alkane conversion activity. Li et al.²⁴ tested GaN and β-Ga₂O₃ in methane conversion, and found that the former showed an approximately five-fold higher activity and a two-fold higher benzene selectivity than the latter. Xing et al.²⁵ investigated the particle size effect of GaN in the oxidative dehydrogenation of n-butane, and revealed that small GaN particles performed better than their counterparts with large sizes. Dutta et al.²³ used anhydrous ammonia for nitridation of β-Ga₂O₃ to GaN. The higher the degree of nitridation (the conversion of β-Ga₂O₃ to GaN), the better the C₂H₄ yield of methane conversion.

Numerous methods have been reported for the preparation of nanostructured GaN particles, such as solvothermal decomposition of Ga–urea complexes,²⁷ the molecular-beam-epitaxy growth method,²⁸ and ammonia nitridation of Ga precursors.^23,29 However, small-scale production, tedious preparation steps, severe synthesis conditions, and corrosion of nitridation systems caused by anhydrous ammonia are challenges that remain to be overcome. A promising method for synthesizing GaN is solid-state pyrolysis.²⁶ Co-pyrolyzing the precursors of Ga (e.g., gallium nitrate hydrate) and N (e.g., melamine) in a one-pot process can be used to prepare supported and unsupported GaN nanoparticles. However, the influence of Ga and N precursors on the derived GaN nanoparticles and the structure–activity correlation of the solid-state-pyrolysis-made GaN in alkane conversion remain to be clarified.

This study investigated the mixtures of gallium nitrate hydrate and the N precursor (including melamine (C₃H₆N₆), melem (C₆H₆N₁₀), or graphitic carbon nitride (g-C₃N₄)) in solid-state pyrolysis and applied their derived GaN catalysts in the anaerobic conversion of methane to AcCN. The use of g-C₃N₄ with a N/Ga ratio of 2 is preferred because the derived GaN catalyst has a higher methane conversion and AcCN productivity. The better activity was attributed to the coexistence of well dispersed GaN nano-crystals and amorphous CN species containing sp² N and C≡N groups. The advantage of using g-C₃N₄ in GaN synthesis was thereby proposed.

2. Experiments

2.1. Materials

Gallium(III) nitrate hydrate (Ga(NO₃)₃·xH₂O, 99.9998%) was purchased from Acros. Commercially available gallium nitride (denoted as c-GaN, 99.99%), beta-gallium oxide (β-Ga₂O₃, 99.99%), and melamine (99%) were purchased from Alfa Aesar. Methane (99.999%) and nitrogen (99.999%) were obtained from Air Products.

2.2. Preparation of nitrogen precursors (melem and g-C₃N₄)

Melem and g-C₃N₄ were prepared by pyrolyzing melamine in N₂ atmosphere.³⁰ Approximately 20 g of melamine was placed in a crucible and sealed with aluminum foil, and then linearly heated (2.5 °C min⁻¹) in the furnace to 450 °C and the isotherm was maintained for 4 h to synthesize melem; on heating to 550 °C for 4 h, g-C₃N₄ was obtained. The obtained samples were ground into powder form.

2.3. Preparation of GaN

The GaN catalysts were prepared by using a modified solid-state pyrolysis method.²⁶ The mixture of the nitrogen precursor and gallium nitrate hydrate was ultrasonicated in methanol (25 mL) for 15 min, and then dried in an oven at 90 °C overnight. The dried powder was transferred into a horizontal quartz tube, pyrolyzed at 800 °C (ramping rate of 1 °C min⁻¹) and kept isothermal for 1 h in a N₂ stream (20 mL min⁻¹). The as-synthesized GaN catalyst was named as GaN-(N precursor)-(X), where X indicates the weight ratio of the N-source to the gallium precursor, namely N/Ga ratios, including 0.5, 1, and 2. For example, GaN-(C₃N₄)-(2) was prepared by using a mixture of g-C₃N₄ and gallium nitrate hydrate with an N-to-Ga weight ratio of 2. Moreover, to reveal the cavity effect of g-C₃N₄, Ga supported on g-C₃N₄ (Ga/C₃N₄) was prepared by pyrolyzing the mixture of g-C₃N₄ and gallium nitrate hydrate (N/Ga = 2) at 550 °C for 1 h.

2.4. Activity test

A horizontal quartz tube (i.d. = 1.3 cm) was used as a fixed-bed reactor for methane conversion. Approximately 0.2 g of catalyst (40–80 mesh) was placed in the center of the tube and sandwiched with quartz wool. Before each test, the catalyst was treated in a N₂ stream (25 mL min⁻¹) at 700 °C with a ramping rate of 10 °C min⁻¹. After pre-activating the catalyst for 0.5 h, the N₂ stream was switched to a methane flow (10 mL min⁻¹) with a gas hourly space velocity (GHSV) of 3000 mL_CH4 g_cat⁻¹ h⁻¹. All tubing and connections were wrapped with heating tape and heated at 150 °C to avoid condensation of the products. The downstream exhaust was split into two lines for on-stream analysis using a quadrupole mass spectrometer (MS, Thermostar GSD 320, Pfeiffer Vacuum) and a gas chromatograph (GC, SRI 8610C, PoraPLOT Q-HT (25 m × 0.53 mm × 20 μm) GC column) with a flame ionization detector and a thermal conductivity detector. The data was recorded immediately after admitting methane to the system by the MS, and was recorded by the GC at a steady-state operation (40 min after admitting methane). Methane conversion and productivity were calculated using the following equations:where CH_4,inlet is the molar flow rate of methane (μmol min⁻¹) at the inlet and C_xH_y is the molar flow rate of the products (μmol min⁻¹) at the outlet. Other products, including C₂ hydrocarbons (ethylene and ethane), hydrogen cyanide (HCN), and aromatics (mostly benzene with trace toluene) were detected. The productivity of C₂ products was calculated by using the MS fragment of 27 because their GC signals were masked by the large tailing of methane. The initial methane conversion and productivity of products (the first data point of the on-stream test, at approximately 5 min after admitting methane) were obtained by MS using fragments of 26 (HCN), 27 (C₂), 41 (AcCN), and 78 (benzene) for calculation. The fragment of 32 (O₂) was used as the relative standard to quantify the aforementioned products since no CO and CO₂ were detected (the leakage-free system). Besides the initial activity, all the data points were obtained by using an on-line GC with a calibrated response factor of each species besides C₂ products.

The effect of reaction temperature (675, 700, and 725 °C) was studied at a GHSV of 3000 mL_CH4 g_cat⁻¹ h⁻¹ using GaN-(melamine, melem, or C₃N₄)-(1). The effect of GHSV (1500 and 4500 mL_CH4 g_cat⁻¹ h⁻¹) was evaluated at 700 °C. Catalyst regeneration was tested by using a calcination–nitridation or a nitridation–activation process. For the calcination–nitridation process, the temperature was cooled to 600 °C (30 min) after a 6 h test under a N₂ stream (20 mL min⁻¹) and maintained for 1 h under a flow of air (25 mL min⁻¹) to remove deposited coke. After calcination, the temperature was increased to 800 °C (30 min) in a N₂ stream (20 mL min⁻¹). The flow was switched to 99.99% NH₃ (20 mL min⁻¹) and kept isothermal for 9 h for nitridation. For the nitridation–activation process, after a 6 h test, the temperature was directly increased to 800 °C (30 min) in a N₂ stream (20 mL min⁻¹). The flow was then switched to 99.99% NH₃ (20 mL min⁻¹) and kept isothermal for 9 h for nitridation. After nitridation, the temperature was cooled under an NH₃ stream to room temperature. The regenerated catalyst was purged with N₂ at room temperature for 1 h, then was activated and tested using the same conditions as its fresh form.

To examine whether or not N₂ can be used to replenish the consumed N species of the spent catalyst, an alternating feeding test of CH₄–N₂–CH₄ was conducted.

2.5. Characterization

X-ray diffraction (XRD) patterns were recorded on a diffractometer (Rigaku D/Max-IIB) using Cu Kα radiation. Elemental analysis of Ga was quantified using an inductively coupled plasma mass spectrometer (ICP-MS, THERMO-ELEMENT XR), and the concentration of N was measured using an elemental analyzer (EA, German Elementar, UNICUBE). The porosity was obtained from N₂-physisorption at 77 K using a Micromeritics ASAP 2020 instrument. The shape and d-space of the GaN crystals were observed using a transmission electron microscope-energy dispersive spectrometer (TEM-EDS, JEOL 3010).

X-ray absorption spectra (XAS) at the Ga K-edge were recorded in transmission mode at the beamline TLS-17C1 of the National Synchrotron Radiation Research Center (NSRRC), Taiwan. The incident X-ray was monochromatized with a Si(111) double crystal monochromator. The Ga K-edge XAS spectra of β-Ga₂O₃ were collected as a reference. The Athena and Artemis software ver. 0.9.26 included in the Demeter package³¹ was used to analyze the spectra. The X-ray absorption near edge structure (XANES) N K-edge was recorded at the high-energy spherical grating monochromator (HSGM) TLS20A (NSRRC) beamline using the total electron yield (TEY) mode, with a base pressure of 5 × 10⁻⁹ Torr and a resolving power of approximately E/ΔE = 8000. c-GaN was used for energy calibration. X-ray photoelectron spectroscopy (XPS) was carried out using a Kratos Axis Ultra DLD instrument with a focused monochromatic Al Kα X-ray source (1486.7 eV). The C 1s signal at 285.0 eV was used to correct binding energy shifts.

Fourier transform infrared spectroscopy (FTIR) was performed using a Thermo Scientific Nicolet iS50 spectrometer. The spectral resolution was 4 cm⁻¹ and 20 scans. The background spectrum was measured at room temperature and was subtracted from each spectrum. In situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) analysis was conducted by using a DRIFT cell (Praying Mantis, Harrick Scientific Products) and an MCT detector in the range of 650–4000 cm⁻¹ with 32 scans at a resolution of 4 cm⁻¹ using a ZnSe window. An approximately 20 mg of GaN-(C₃N₄)-(2) was placed in a holder and was then purged with N₂ (25 mL min⁻¹) for 30 min for each run. The in situ DRIFT analyses of GaN-(C₃N₄)-(2) included temperature-resolved spectra in a N₂ stream (20 mL min⁻¹), in an air stream (20 mL min⁻¹), and in a sealed chamber filled with 50% CH₄ in N₂.

3. Results

3.1. Activity evaluation

Table 1 presents the activity results of the initial run and after a 6 h operation of each GaN catalyst and Fig. S1–S7† show the time on-stream (TOS) profiles. The effect of reaction temperatures (675, 700, and 725 °C) was studied over GaN-(melamine)-(1), GaN-(melem)-(1), and GaN-(C₃N₄)-(1), as shown in entries 1 to 9 of Table 1. The initial methane conversion increased with increasing temperature for each catalyst. Among the tested catalysts, GaN-(C₃N₄)-(1) had higher initial methane conversions (20.0–29.9%) and AcCN productivities (3631–4100 μmol g_cat⁻¹) at 675–725 °C. The accumulated AcCN in the 6 h operation was higher at 700 °C for each catalyst. Accordingly, 700 °C was selected as the operating temperature for further investigation.

Table 1 Initial and final activity results of GaN catalysts

Entry	Sample	T (°C)	GHSV (mLCH4 gcat^−1 h^−1)	CH4 conversion (%)		AcCN productivity (μmol gcat^−1 min^−1)		6 h accumulated AcCN productivity (μmol gcat^−1)
				Initial^a	Final	Initial^a	Final
a Initial methane conversion and AcCN productivity were estimated by using the MS signals.
1	GaN-(melamine)-(1)	675	3000	13.8	1.0	33	1.8	1795
2		700		14.6	1.0	34	1.6	2031
3		725		16.2	0.9	36	1.0	1995
4	GaN-(melem)-(1)	675	3000	18.9	1.4	45	2.8	2736
5		700		27.3	1.2	68	2.0	2993
6		725		28.9	0.8	71	0.9	2665
7	GaN-(C3N4)-(1)	675	3000	20.0	2.1	50	4.7	3917
8		700		26.6	0.9	123	1.4	4100
9		725		29.9	1.1	74	1.5	3631
10	GaN-(melamine)-(1)	700	1500	28.1	2.6	72	6.4	4890
11			4500	10.0	0.6	24	0.5	1168
12	GaN-(melem)-(1)	700	1500	29.7	2.6	71	6.4	4880
13			4500	21.0	0.8	67	1.1	2387
14	GaN-(C3N4)-(1)	700	1500	37.7	1.9	186	8.1	9239
15			4500	16.8	0.6	83	1.0	3097
16	GaN-(C3N4)-(0.5)	700	3000	21.3	1.1	102	0.4	3486
17	GaN-(C3N4)-(2)			28.2	1.7	151	3.3	5816

---

The effect of GHSV was studied over GaN-(melamine)-(1), GaN-(melem)-(1), and GaN-(C₃N₄)-(1), as shown in entries 2, 5, 8, and 10–15 of Table 1 and Fig. S4–S6.† The onset methane conversion and AcCN productivity increased with respect to the decreased GHSV. GaN-(C₃N₄)-(1) showed a better activity in most of these trials. At 700 °C with a GHSV = 1500 mL_CH4 g_cat⁻¹ h⁻¹, GaN-(C₃N₄)-(1) had its initial conversion at 37.7%; initial AcCN productivity, 186 μmol g_cat⁻¹ min⁻¹; accumulated AcCN, 9239 μmol g_cat⁻¹. All these results indicated the potential of GaN-(C₃N₄)-(1) in the conversion of methane to AcCN.

Since GaN derived from g-C₃N₄ showed the best performance, the effect of the N/Ga ratio was investigated using GaN-(C₃N₄)-(0.5, 1, and 2), as shown in entries 8, 16, and 17 of Table 1 and Fig. S7.† All of them exhibited a gradual decrease of methane conversion and AcCN productivity in the 6 h testing from approximately 21.3% to 1.1% and from 102 to 0.4 μmol g_cat⁻¹ min⁻¹ for GaN-(C₃N₄)-(0.5), from 26.6% to 0.9% and from 123 to 1.4 μmol g_cat⁻¹ min⁻¹ for GaN-(C₃N₄)-(1), and from 28.2% to 1.7% and from 151 to 3.3 μmol g_cat⁻¹ min⁻¹ for GaN-(C₃N₄)-(2). Among them, GaN-(C₃N₄)-(2) displayed the highest activity and accumulated AcCN yield (5816 μmol g_cat⁻¹).

The productivities of minor products, including HCN (<14 μmol g_cat⁻¹ min⁻¹), C₂ hydrocarbons (<12 μmol g_cat⁻¹ min⁻¹), and aromatics (<2 μmol g_cat⁻¹ min⁻¹) are shown in Fig. S1–S7.† Declining trends of these products were observed. For comparison, Fig. S8† shows the activity test of c-GaN, which displayed negligible methane conversion (<0.6%) and AcCN productivity (<0.2 μmol g_cat⁻¹ min⁻¹).

The reproducibility test was done over three different batches of GaN-(C₃N₄)-(2), see Fig. S9.† The 95% confidence interval of each data point was less than 5%, indicating that the reproducibility of our catalytic system is acceptable.

The regeneration tests of GaN-(C₃N₄)-(2) through the calcination–nitridation and nitridation–activation processes are shown in Fig. S10† and Fig. 1, respectively. By using the calcination–nitridation method in catalyst regeneration, the activity of the spent GaN-(C₃N₄)-(2) can hardly be rejuvenated: the methane conversion and productivities of all species were less than their respective values at the end of the 1st on-stream test. In contrast, the activity of the spent GaN-(C₃N₄)-(2) regenerated by using the nitridation–activation process was partially rejuvenated. The initial conversion (5.6%) and productivities of AcCN (9.8%), HCN (1.7%), and aromatics (2.7%) of the regenerated GaN-(C₃N₄)-(2) were all higher than their respective values at the end of the 1st on-stream test. Interestingly, C₂ productivities were nearly identical before and after regeneration in the range of 6.6% to 8.4%.

Fig. 1 (a) Methane conversion and (b) productivities of AcCN, C₂, HCN, and aromatics from the regeneration test of GaN-(C₃N₄)-(2) using the nitridation–activation process.

Fig. S11† shows the alternating feeding test of CH₄–N₂–CH₄. When introducing methane, AcCN is produced concurrently and declined gradually. After the reaction for 40 min, the flow was switched to the N₂ flow. In the meantime, AcCN declined sharply to an undetectable level. After flowing N₂ for 80 min, the stream was switched back to CH₄, and the AcCN signal was regained to a similar level to that of its regular test.

In brief, g-C₃N₄ is deemed to be the most favorable N precursor. Moreover, the high N/Ga ratio (2) used in co-pyrolyzing Ga(NO₃)₃·xH₂O and g-C₃N₄ improves the activity in methane conversion to AcCN. Accordingly, further analyses of GaN-(C₃N₄)-(0.5, 1, and 2) in comparison with GaN-(melamine)-(1) and GaN-(melem)-(1) were carried out. Moreover, explanation of catalyst deactivation was obtained by the characterization of the post-reaction catalysts.

3.2. Bulk structure of GaN catalysts

Table 2 presents the results of elemental analysis of the tested samples. The measured Ga contents, ranging from 76.9 to 86.2 wt%, were close to the theoretical value (83.3 wt%; less than 8% variation). The N concentrations varied from 10.6 to 17.9 wt% while the theoretical value is 16.7 wt%. The values of the N/Ga ratio ranged from 0.62 to 1.14. The compositions of Ga and N of post-reaction catalysts were also analyzed, as presented in Table S1.† Slightly lower Ga and N concentrations, compensating for carbon deposition, of the post-reaction catalysts than their fresh forms were found. Nevertheless, the N/Ga ratios of the catalysts before and after reaction were close.

Table 2 The concentrations of Ga and N, and the porosity of tested samples

Catalyst	Ga^a (wt%)	N^b (wt%)	N/Ga molar ratio	S BET (m^2 g^−1)	V P (cm^3 g^−1)	D P (nm)
a Estimated by ICP-MS. b Estimated by EA. c Surface area estimated by using the BET method. d Pore volume and pore diameter estimated by using the BJH method.
c-GaN	86.2	10.7	0.62	4.4	0.025	N.D.
GaN-(melamine)-(1)	79.2	17.2	1.08	30.9	0.088	11.6
GaN-(melem)-(1)	80.9	10.6	0.65	42.3	0.062	6.2
GaN-(C3N4)-(0.5)	81.9	15.6	0.95	31.4	0.086	11.8
GaN-(C3N4)-(1)	78.2	17.9	1.14	23.2	0.061	12.0
GaN-(C3N4)-(2)	79.5	15.5	0.97	21.1	0.046	8.2

---

Fig. S12† shows the N₂ adsorption–desorption isotherms and the pore size distributions, and Table 2 presents the porosities of the tested catalysts. A type V-like isotherm with a type H3 hysteresis loop was observed for each sample, suggesting that disordered mesopores were dominant. The surface areas and pore volumes of the as-prepared catalysts were close, in the range of 16.2 to 42.3 m² g⁻¹ and 0.042 to 0.097 cm³ g⁻¹, respectively. c-GaN had an order of magnitude lower surface area (4.4 m² g⁻¹) and approximately half of the pore volume (0.025 cm³ g⁻¹) compared to those of as-prepared GaN.

Fig. 2a shows the structural analysis of GaN-based catalysts. The XRD patterns of the tested catalysts exhibited the wurtzite GaN phase, similar to that of c-GaN. The crystallite sizes were obtained from the Scherrer equation using the (101) diffraction plane. All the as-synthesized GaN catalysts showed smaller crystallites, as compared to that of c-GaN. The GaN crystallites of g-C₃N₄ made samples (5.8 nm) were smaller than those of GaN-(melamine)-(1) (8.1 nm) and GaN-(melem)-(1) (7.0 nm). Moreover, the crystallinity of GaN-(C₃N₄)-(X) was changed with respect to the N/Ga ratio: the crystallite size of GaN was reduced with the increased N/Ga ratio.

Fig. 2 (a) XRD patterns of solid-state-pyrolysis-made GaN catalysts by using the mixtures of gallium nitride hydrate and different N sources (i.e., melamine, melem, and g-C₃N₄). The patterns of c-GaN and β-Ga₂O₃ were included for reference. The TEM images of (b) c-GaN and (c) GaN-(C₃N₄)-(2).

The crystal structure of the post-reaction catalyst was also analyzed by using XRD. Fig. S13† shows that all the post-reaction catalysts retained their wurtzite GaN structure. The diffraction intensities of the post-reaction catalyst were slightly stronger than those of its fresh form. Thus, the estimated crystallite sizes of the post-reaction catalysts were a little higher than their fresh forms.

Fig. 2b and c exhibit the TEM images of c-GaN and GaN-(C₃N₄)-(2). Regular lattice fringes, including the m-plane (5.3 Å) and the (002) lattice spacing (2.6 Å) of wurtzite GaN,^23,32 were observed for both samples. A clear edge was observed for c-GaN, whereas dislocations were identified for GaN-(C₃N₄)-(2). Accordingly, the crystallinity of wurtzite GaN of c-GaN was higher than that of GaN-(C₃N₄)-(2).

Fig. 3a shows the XANES spectra of the tested catalysts and standards including β-Ga₂O₃ and c-GaN. Each catalyst had its white-line peak amplitude weaker than that of β-Ga₂O₃. Moreover, a weaker signal in the range of 10 378–10 390 eV, where the characteristic oscillations of GaN located, was discovered for each GaN catalyst. In the range of approximately 300 eV higher than the edge (10 367 eV), eight oscillations belonging to GaN were identified for each as-synthesized catalyst with weaker amplitudes than those of c-GaN (see Fig. 3b).

Fig. 3 (a) Ga K-edge XANES of GaN-based catalysts and (b) the eight oscillation peaks (marked with asterisks) in the first 300 eV above the absorption edge. (c) Ga–N phase corrected RDFs of the Ga K-edge of tested catalysts. The solid line represents the experimental data, and the dashed line represents the computer fit.

Fig. 3c shows the phase corrected radial distribution functions (RDFs) acquired from the Fourier transformation (phase corrected) of the k³-weighted EXAFS data and Table 3 presents the parameters obtained from the curve fitting results. The values of the coordination number (CN) of the 1st Ga–N shell (1.92 Å) and the 2nd Ga–Ga shell (3.19 Å) of c-GaN were 4 and 12, respectively, in line with the hexagonal wurtzite structure of GaN. GaN-(melamine)-(1) and GaN-(melem)-(1) had similar RDFs with slightly lower CN values of the 1st (2.9 and 3.3) and 2nd (8.4 and 8.2) shells than those of c-GaN. In contrast, the RDFs of GaN-(C₃N₄)-(0.5, 1, and 2) showed mainly the 1st shell response with a weak/negligible signal of the 2nd shell. The 1st shell CN values of GaN-(C₃N₄)-(0.5, 1, and 2) were all greater than 4, while the parameter uncertainties were large for the fitted values of the 2nd shell.

Table 3 Parameters obtained from the curve-fitting results of the EXAFS data at the Ga K-edge^a

Sample	Scattering path	CN	R (Å)	σ ^2 (Å^2)	R-Factor
a CN = coordination number; R = interatomic distance; σ^2 = Debye–Waller factor.
c-GaN	Ga–N	4.0 ± 0.0	1.92 ± 0.01	0.0037	0.0098
	Ga–Ga	12.0 ± 0.0	3.19 ± 0.01	0.0081
GaN-(melamine)-(1)	Ga–N	2.9 ± 0.5	1.92 ± 0.01	0.0048	0.0126
	Ga–Ga	8.4 ± 1.9	3.20 ± 0.01	0.0093
GaN-(melem)-(1)	Ga–N	3.3 ± 0.4	1.93 ± 0.01	0.0039	0.0112
	Ga–Ga	8.2 ± 1.7	3.20 ± 0.01	0.0087
GaN-(C3N4)-(1)	Ga–N	4.9 ± 0.6	1.93 ± 0.01	0.0070	0.0097
	Ga–Ga	10.8 ± 9.5	3.19 ± 0.03	0.0336
GaN-(C3N4)-(0.5)	Ga–N	5.4 ± 0.7	1.93 ± 0.01	0.0064	0.0110
	Ga–Ga	7.2 ± 4.0	3.19 ± 0.02	0.0170
GaN-(C3N4)-(2)	Ga–N	7.0 ± 0.7	1.96 ± 0.01	0.0088	0.0086

---

To reveal the unique chemistry of GaN derived from g-C₃N₄, an X-ray absorption intensity analysis of the sp² and sp³ environments of the N K-edge XANES spectrum was performed (see Fig. 4). The four characteristic responses (A B, C, and D) corresponding to the s and p hybridized orbitals³³ were deconvoluted into six Gaussian peaks (G₁ to G₆) to simulate the transitions to the bound state. The peaks of G₁, G₃, and G₆ are attributed to the sp² environment; G₂, G₄, and G₅, sp³ environment.³⁴ Accordingly, the peak area ratio of (G₂ + G₄ + G₅)/(G₁ + G₃ + G₆) can be used to estimate the relative composition of the sp³ (interstitial N in wurtzite GaN)-to-sp² (amorphous N) environment.³⁵

Fig. 4 N K-edge XANES of GaN-based catalysts. The fitted peaks of G₁ (red), G₂ (green), G₃ (blue), G₄ (cyan), G₅ (magenta), and G₆ (yellow) were included.

The N K-edge XANES spectra of c-GaN, GaN-(melamine)-(1), and GaN-(melem)-(1) were similar, showing a strong B response between the A and C responses with similar intensities. In comparison, the spectrum of GaN-(C₃N₄)-(0.5, 1, or 2) shows a relatively weak B response compared to those of the A and C responses. Moreover, the relative composition of sp³/sp² decreased in the following order: c-GaN (1.28) > GaN-(melamine)-(1) (1.06) > GaN-(melem)-(1) (1.00) > GaN-(C₃N₄)-(0.5) (0.99) > GaN-(C₃N₄)-(1) (0.96) > GaN-(C₃N₄)-(2) (0.91). That is, relatively high amounts of amorphous N were formed on g-C₃N₄ derived GaN catalysts, and GaN-(C₃N₄)-(2) had the highest relative composition of amorphous N species.

Fig. S14† shows the Ga XANES spectrum, the RDF profile, and the N XANES spectrum of the post-reaction GaN-(C₃N₄)-(2). The post-reaction GaN-(C₃N₄)-(2) had similar profiles to those of c-GaN. The fitted parameters of the 1st and 2nd shells of spent GaN-(C₃N₄)-(2) were 3.1 and 11.6 and were close to the 1st and 2nd CN values of c-GaN. The ratio of the sp³/sp² N environment of spent GaN-(C₃N₄)-(2) was 1.06, lower than that of its fresh counterpart (0.91).

3.3. Surface analysis of GaN catalysts

Fig. 5 shows the XPS spectra of Ga 3d, N 1s, C 1s, and O 1s core levels of the as-prepared GaN catalysts. The Ga 3d spectrum (Fig. 5a) can be deconvoluted into 3 peaks at 17.9 eV (N 1s), 19.5 eV (Ga–N), and 20.3 eV (Ga–O).³⁶ The Ga 3d spectrum of c-GaN is also presented in Fig. S15a.† Each sample showed a relative composition of Ga–N response of >60%; that of Ga–O was <35%; and that of N 1s was <10%.

Fig. 5 The XPS spectra of (a) Ga 3d, (b) N 1s, (c) C 1s, and (d) O 1s photolines of as-synthesized GaN catalysts.

The N 1s spectrum of c-GaN (Fig. S15b†) was fitted by five responses: Auger Ga (392.9, 394.2, and 395.6 eV), N–Ga (397.4 eV), and N–H (399.5 eV).³⁷ Two additional responses of sp² N (e.g., C–N=C at 398.1 eV) and C≡N (400.6 eV) were resolved for the as-synthesized GaN catalysts (Fig. 5b). The sum of the relative compositions of sp² N and C≡N in GaN-(C₃N₄)-(1) (17.6%) was greater than those of GaN-(melamine)-(1) (14.7%) and GaN-(melem)-(1) (15.9%). In addition, by using more g-C₃N₄ as the precursor, more sp² N and C≡N groups were produced: GaN-(C₃N₄)-(2) (25.5%) > GaN-(C₃N₄)-(1) (17.6%) > GaN-(C₃N₄)-(0.5) (15.8%).

Fig. 5c displays the C 1s spectra of as-prepared GaN. The C 1s photoline of c-GaN (Fig. S15c†) shows a sharp signal at 284.6 eV (FWHM = ca. 1.8 eV) and a shoulder at 286.3 eV, corresponding to the C–C bond of adventitious carbon and the C–NH_x bond.^38,39 A broad C 1s signal was found for the as-synthesized GaN catalysts (FWHM > 2.0 eV), and the responses of sp² C (i.e., C–N=C) and C≡N were used for peak deconvolution. Nevertheless, no clear trend was observed for the relative composition of each C species. The O 1s spectrum (Fig. 5d) of each sample displayed mainly the responses of physisorbed O and hydroxyls⁴⁰ with a trivial portion of the Ga–O signal.

Fig. 6 shows the FTIR spectra of the as-synthesized GaN catalysts and c-GaN. The bands of C–N and C=N stretching modes at 1100–1750 cm⁻¹, C≡N stretching mode at 2000–2350 cm⁻¹, and –NH_x (x = 1 and 2)/–OH vibrations at 2500–3750 cm⁻¹ were identified for all the as-synthesized catalysts. The C–N and C=N stretching modes originated from the sp² N species (e.g., triazine units^41,42); the C≡N stretching, cyanogen-like structure;⁴³ and the –NH_x vibrations, primary and secondary amines.⁴⁴ In contrast, these signals are negligible for c-GaN. The FTIR spectra of post-reaction catalysts were also recorded, as shown in Fig. S16.† The intensities of surface CN species declined significantly for the post-reaction catalysts.

Fig. 6 FTIR spectra of as-synthesized GaN and c-GaN.

3.4. In situ IR analysis

To reveal the role of amorphous C and N species in the production of AcCN, in situ IR analysis was conducted. Fig. 7a shows the IR spectra of GaN-(C₃N₄)-(2) recorded from 30 to 700 °C in a N₂ stream (25 mL min⁻¹). With increasing temperature, the amplitude of the –NH_x/–OH vibrations decreased gradually and diminished at 700 °C. That is, the –NH_x groups were thermally unstable and shall be bystanders in methane conversion. In contrast, the amplitudes of C–N, C=N, and C≡N stretching modes were mainly preserved even at 700 °C.

Fig. 7 (a) The temperature-resolved IR spectra of GaN-(C₃N₄)-(2) from 30 to 700 °C in a N₂ stream (20 mL min⁻¹). (b) The IR spectra of GaN-(C₃N₄)-(2) before and after exposure to 50% CH₄ in N₂ at 700 °C for 6 h.

Fig. 7b shows the in situ IR spectra of GaN-(C₃N₄)-(2) before and after being used in methane conversion. The amplitudes of the C–N, C=N, and C≡N bands of spent GaN-(C₃N₄)-(2) decreased compared to those of its fresh form. Moreover, a sharp peak at ∼2260 cm⁻¹ of the spectrum of spent GaN-(C₃N₄)-(2) was observed.

3.5. Synergy of surface CN species and GaN in methane conversion

3.5.1. Oxidative abatement of as-prepared GaN. Fig. S17a† shows the TGA and DTG profiles of GaN-(C₃N₄)-(2) and Fig. S17b† shows the MS fragments of NO (m/z = 30) and CO₂ (m/z = 44) of exhaust gas in an air stream. The weight of GaN-(C₃N₄)-(2) showed a decreasing trend first and then an increasing trend at approximately 650 °C and the MS peaks of NO and CO₂ were at approximately 620 °C. Moreover, the in situ IR analysis of GaN-(C₃N₄)-(2) in an air stream showed that the amplitudes of the C–N, C=N, and C≡N bands were gradually decreased from 30 to 600 °C (see Fig. S17c†). Accordingly, 600 °C was selected as the temperature for the oxidative abatement of GaN-(C₃N₄)-(2). The XRD pattern of the oxidation-treated GaN-(C₃N₄)-(2) (see Fig. S17d†) also shows the diffractions of GaN. These results revealed that amorphous CN species could be removed without deep oxidation of GaN through the oxidative abatement.

Fig. S17e† shows the activity test of GaN-(C₃N₄)-(2) subjected to the oxidative abatement. Compared to its un-treated counterpart, the conversion and AcCN yield dropped significantly to less than 1%. Fig. S17f† shows the test of Ga₂O₃-(C₃N₄)-(2) (coexisting amorphous CN species and Ga₂O₃). A low methane conversion (<1%) and a small amount of AcCN (1 μmol g_cat⁻¹ min⁻¹) were observed.

3.5.2. Thermodynamic analysis. Fig. S18† shows the equilibrium compositions as a function of temperature calculated by using ThermoSolver. The mixed CH₄ and N₂ and mixed CH₄ and CN (in a 9-to-1 ratio of each) were used as the reactants.

For both cases, CH₄ showed a downward trend, while H₂ showed an upward trend with nearly identical compositions. C₂H₄ and C₂H₆ were negligible, and C₆H₆ was gradually increased with increasing temperature. The major difference between these two cases is the trends of HCN and AcCN. Trivial HCN (selectivity <2%) with AcCN slightly increased from 500 °C (selectivity = 0.1%) to 1000 °C (selectivity = 32.8%) was found by using CH₄ and N₂ as the reactants (Fig. S18a†). In contrast, the selectivity to HCN decreased from 11.1% to 1.2% and considerable AcCN selectivity, particularly at temperatures below 800 °C (selectivity >65%), was found by using CH₄ and CN as the feeds (Fig. S18b†). Moreover, limited N₂ was converted (N₂ conversion was found to be <47%), while CN was fully consumed in the temperature range.

4. Discussion

4.1. Physicochemical properties of GaN-(C₃N₄)-(X) and their activity in methane conversion

The solid-state-pyrolysis-made GaN by using g-C₃N₄ as the nitrogen precursor has shown its potential for methane conversion to AcCN compared with those made by using melamine and melem. Since similar bulk and surface compositions, porosities, and the same wurtzite GaN structure were found for the as-synthesized catalysts, the promising activity of GaN-(C₃N₄)-(2) can be correlated with its unique GaN geometry and enriched amorphous CN species.

Based on the XRD and Ga K-edge XAFS analyses, GaN-(C₃N₄)-(2) had the smallest GaN crystals. Moreover, the high CN value of the 1st Ga–N shell (7.0) with the absence of the 2nd Ga–Ga shell of the RDF profile underlined that highly dispersed GaN crystallites were formed.⁴⁵ The small GaN crystals have been reported to be more active than the large GaN particles in methane conversion to hydrocarbons.²³ Accordingly, a similar hypothesis of the particle size effect of GaN in methane conversion to AcCN was proposed.

The amorphous CN species is also important in enhancing the activity. The results of N XANES and N 1s XPS both revealed that GaN-(C₃N₄)-(2) had the highest relative composition of amorphous CN species. These species were consumed after the reaction together with the formation of AcCN as observed by the in situ IR study. Moreover, the study of oxidative abatement and the thermodynamic analysis underlined that methane cannot be converted to AcCN without the presence of amorphous CN species.

The study of oxidative abatement of GaN-(C₃N₄)-(2) and the test of Ga₂O₃-(C₃N₄)-(2) indicated that neither GaN nor amorphous CN species can promote the formation of AcCN solely. That is, the interplay of coexisting GaN and amorphous CN species in methane conversion to AcCN was suggested.

4.2. Structure–activity correlation

To emphasize the cooperation of highly dispersed GaN phases and amorphous CN species in improving the AcCN generation, Fig. 8 shows the accumulated AcCN yield in a 6 h reaction time of the as-synthesized catalyst with respect to the GaN crystallite size and the sp³/sp² N ratio.

Fig. 8 The accumulated AcCN yield with respect to the crystallite size of GaN and the sp³/sp² N ratio of as-synthesized catalysts.

An opposite trend was found: the accumulated AcCN yield increased, while both the crystallite size and sp³/sp² N ratio of the as-synthesized GaN catalysts decreased. This highlights that AcCN production can be increased by reducing the GaN crystallite size and by increasing the amounts of amorphous CN species. GaN-(C₃N₄)-(2) had the smallest crystallite sizes and the most abundant amorphous CN species among the tested catalysts, and therefore generated the highest amounts of accumulated AcCN (5816 μmol g_cat⁻¹).

4.3. Mechanism of GaN made by using g-C₃N₄

The well dispersed GaN phase and amorphous CN species may be related to the effect of six-fold cavities of g-C₃N₄. The six-fold cavities of g-C₃N₄ are the most energetically favorable sites for anchoring metal mono-atoms (e.g., Mn, Fe, Co, and Pd) according to DFT computations.^46,47 In addition, g-C₃N₄ supported single metal atoms such as Fe,⁴⁸ Pd,⁴⁹ and Pt⁵⁰ have been successfully fabricated. Accordingly, the mechanism of the well dispersed GaN phase derived from g-C₃N₄ was proposed, as shown in Fig. 9. The cavities of g-C₃N₄ were likely to be the docking ports of Ga³⁺ cations⁵¹ before solid-state pyrolysis (step i). The nanohole-coordinated Ga³⁺ (step ii) allows the suppression of thermal agglomeration in the solid-state pyrolysis, resulting in highly dispersed GaN crystals with a high dispersion (step iii).

Fig. 9 Proposed mechanism of the nanoholes of g-C₃N₄ as the docking ports of Ga³⁺ in GaN synthesis through the solid-state pyrolysis. Red circles mean Ga³⁺ in the solution; blue circles, N of g-C₃N₄; orange circles, Ga³⁺ bonded in the nanoholes.

To support the aforementioned claim, Ga/C₃N₄ was prepared and analyzed by XRD, as shown in Fig. S19.† Relative to g-C₃N₄, the (002) peak of Ga/C₃N₄ was shifted from 27.5° to 27.7° while the (100) peak was unchanged. The (002) peak shift suggests that the Ga cations of Ga/C₃N₄ mainly existed in the six-fold cavity of heptazine units (step ii of Fig. 9), which merely affects the diffraction of the (002) plane belonging to the accumulated heptazine units.⁵² Accordingly, we proposed that before GaN was formed, at least below 550 °C, Ga cations were mainly coordinated in the six-fold cavities of g-C₃N₄.

4.4. Catalyst deactivation and regeneration

The characterization results of the post-reaction catalysts indicated that deactivation was caused by GaN agglomeration, coke deposition, and depletion of amorphous CN species. Advanced catalyst design to allow methane activation under milder conditions shall limit the extent of GaN agglomeration. Coke removal and supplement of surface N content was conducted in the calcination–nitridation treatment. Regretfully, the activity could hardly be regained. A possible explanation is that in the calcination step, deposited coke and remaining CN species were burnt out. That is, amorphous CN species could not be replenished by NH₃ nitridation on a carbon-eliminated surface, resulting in an invalid regeneration treatment.

Stöhr et al.⁵³ reported that N-doped graphene, containing nitrile and/or pyridinic-like nitrogen species, could be formed through NH₃ nitridation over activated carbon. This method was used to directly nitridize the spent catalyst without the calcination treatment. The nitridation–activation process by using NH₃ as the N source could partially rejuvenate the activity. As evidenced by the FTIR spectra (Fig. S20†), the vibrational bands of C[quadruple bond, length as m-dash]N and C=N species were intensified after direct NH₃ nitridation. This confirmed that the CN species could be replenished, resulting in recovering a certain extent of the initial activity for the regenerated catalyst. The replenished CN species were further consumed after the 2nd on-stream test, since the IR bands of C[quadruple bond, length as m-dash]N and C=N species nearly disappeared.

The XRD patterns (Fig. S21†) of the regenerated GaN-(C₃N₄)-(2) before and after the 2nd on-stream test showed that GaN diffractions were enhanced. This suggested that GaN was persistent. The persistent GaN was reflected by the consistent productivity of C₂, which was formed through C–C coupling catalyzed by GaN.⁵⁴ The alternating feeding test of CH₄–N₂–CH₄ underlined that N₂ cannot be activated for AcCN synthesis under the operating conditions. This result is also in line with the thermodynamic analysis, showing negligible N₂ conversion (2.2%) at 700 °C in the methane stream. To sum up, the schematic diagram of methane conversion to AcCN over the solid-state-pyrolysis-made GaN catalyst, and the CN replenishment of the spent catalyst through direct NH₃ nitridation is shown in Fig. 10. The regeneration process may be improved by using other nitridation agents, e.g., HCN or cyanogen gas, to replenish amorphous CN species of the post-reaction GaN. However, special reactor design should be implemented to handle highly toxic cyanides.

Fig. 10 Proposed catalytic cycle of methane conversion to AcCN over the solid-state-pyrolysis-made GaN catalyst.

5. Conclusion

Co-pyrolyzing gallium nitrate hydrate and melamine, melem, or g-C₃N₄ produced GaN catalysts that can be used for methane conversion to AcCN. g-C₃N₄ is preferred among the N precursors: well dispersed GaN crystals and enriched amorphous CN species can be generated. These two factors are responsible for enhancing methane conversion and AcCN productivity. The six-fold cavities of g-C₃N₄ were proposed to be the docking ports of Ga³⁺, confining Ga³⁺ cations to circumvent agglomeration during solid-state pyrolysis. The activity of the post-reaction catalyst could be partially regained by using a NH₃ nitridation–activation process. This study aims to reveal a novel GaN catalyst design for methane upgrading.

Author contributions

Korawich Trangwachirachai: investigation, formal analysis, validation, writing – original draft. Chin-Han Chen: investigation, formal analysis, validation. Ai-Lin Huang: investigation, formal analysis. Jyh-Fu Lee: XAS data analysis, resources. Chi-Liang Chen: XAS data analysis, resources. Yu-Chuan Lin: supervision, project administration, resources.

Conflicts of interest

The authors declare that they have no competing financial interests.

Acknowledgements

This study was supported by the Ministry of Science and Technology (Projects 109-2628-E-006-011-MY3, 110-2221-E-006-165-MY3, 110-2923-E-006-005-MY3, and 110-2927-I-006-506) and the Higher Education Sprout Project, Ministry of Education to the Headquarters of University Advancement at National Cheng Kung University (NCKU). The authors also appreciate the use of an elemental analyzer (EA000600, MOST 110-2731-M-006-001) belonging to the Core Facility Center of National Cheng Kung University.

Notes and references

1. E. Dlugokencky and S. Houweling, in Encyclopedia of Atmospheric Sciences, ed. J. R. Holton, Academic Press, Oxford, 2003, pp. 1286–1294,  DOI:10.1016/B0-12-227090-8/00223-2.
2. J. Johnson, C&EN Global Enterprise, 2016, 94, 16.
3. Y. Li and G. Veser, Energy Technol., 2020, 8, 2000460.
4. M. E. Casco, F. Rey, J. L. Jorda, S. Rudic, F. Fauth, M. Martinez-Escandell, F. Rodriguez-Reinoso, E. V. Ramos-Fernandez and J. Silvestre-Albero, Chem. Sci., 2016, 7, 3658–3666.
5. P. Schwach, X. Pan and X. Bao, Chem. Rev., 2017, 117, 8497–8520.
6. R. Lin, Y. Ding, L. Gong, W. Dong, J. Wang and T. Zhang, J. Catal., 2010, 272, 65–73.
7. K. Ding, H. Metiu and G. D. Stucky, ACS Catal., 2013, 3, 474–477.
8. V. Paunović and J. Pérez-Ramírez, Catal. Sci. Technol., 2019, 9, 4515–4530.
9. C. Díaz-Urrutia and T. Ott, Science, 2019, 363, 1326–1329.
10. L. Sun, Y. Wang, N. Guan and L. Li, Energy Technol., 2020, 8, 1900826.
11. S. Huo, R. Wu, M. Li, H. Chen and W. Zuo, Ind. Eng. Chem. Res., 2020, 59, 22690–22695.
12. J. Jang, K. Shen and C. G. Morales-Guio, Joule, 2019, 3, 2589–2593.
13. S. Yuan, Y. Li, J. Peng, Y. M. Questell-Santiago, K. Akkiraju, L. Giordano, D. J. Zheng, S. Bagi, Y. Román-Leshkov and Y. Shao-Horn, Adv. Energy Mater., 2020, 10, 2002154.
14. T. Yu, Z. Li, L. Lin, S. Chu, Y. Su, W. Song, A. Wang, B. M. Weckhuysen and W. Luo, ACS Catal., 2021, 11, 6684–6691.
15. S.-Y. He, T.-H. Li, Z.-Q. Huang, Y. Liu, J. Li and C.-R. Chang, J. Chem. Phys., 2021, 154, 174706.
16. P. Pollak, G. Romeder, F. Hagedorn and H.-P. Gelbke, in Ullmann's Encyclopedia of Industrial Chemistry, 2000,  DOI:10.1002/14356007.a17_363.
17. T. A. Hanna, Coord. Chem. Rev., 2004, 248, 429–440.
18. Y. Li and J. N. Armor, J. Catal., 1997, 173, 511–518.
19. Y. Li and J. N. Armor, Appl. Catal., A, 1999, 188, 211–217.
20. E. Rojas, M. O. Guerrero-Pérez and M. A. Bañares, Catal. Commun., 2009, 10, 1555–1557.
21. B. Rhimi, M. Mhamdi, A. Ghorbel, V. N. Kalevaru, A. Martin, M. Perez-Cadenas and A. Guerrero-Ruiz, J. Mol. Catal. A: Chem., 2016, 416, 127–139.
22. K. Trangwachirachai, C.-H. Chen and Y.-C. Lin, Mol. Catal., 2021, 516, 111961.
23. K. Dutta, V. Chaudhari, C.-J. Li and J. Kopyscinski, Appl. Catal., A, 2020, 595, 117430.
24. L. Li, X. Mu, W. Liu, X. Kong, S. Fan, Z. Mi and C.-J. Li, Angew. Chem., Int. Ed., 2014, 53, 14106–14109.
25. T. Xing, Y. Lang, J.-G. Chen, Q.-X. Luo, C. Liu, J. Song, Z.-W. Liu and Z.-T. Liu, Can. J. Chem. Eng., 2019, 97, 3115–3124.
26. L. Zhang, Z.-Y. Wang, J. Song, Y. Lang, J.-G. Chen, Q.-X. Luo, Z.-H. He, K. Wang, Z.-W. Liu and Z.-T. Liu, J. CO2 Util., 2020, 38, 306–313.
27. K. Sardar, M. Dan, B. Schwenzer and C. N. R. Rao, J. Mater. Chem., 2005, 15, 2175–2177.
28. H. P. T. Nguyen, S. Zhang, K. Cui, X. Han, S. Fathololoumi, M. Couillard, G. A. Botton and Z. Mi, Nano Lett., 2011, 11, 1919–1924.
29. C. Sun, M. Yang, T. Wang, Y. Shao, Y. Wu and X. Hao, ACS Appl. Mater. Interfaces, 2017, 9, 26631–26636.
30. H. Zheng, Z. Zhao, J. B. Phan, H. Ning, Q. Huang, R. Wang, J. Zhang and W. Chen, ACS Appl. Mater. Interfaces, 2020, 12, 2145–2151.
31. B. Ravel and M. Newville, J. Synchrotron Radiat., 2005, 12, 537–541.
32. P. V. Chandrasekar, H. Jung, C. G. Kim and D. Kim, CrystEngComm, 2012, 14, 2166–2171.
33. W. R. L. Lambrecht, S. N. Rashkeev, B. Segall, K. Lawniczak-Jablonska, T. Suski, E. M. Gullikson, J. H. Underwood, R. C. C. Perera, J. C. Rife, I. Grzegory, S. Porowski and D. K. Wickenden, Phys. Rev. B: Condens. Matter Mater. Phys., 1997, 55, 2612–2622.
34. K. Lawniczak-Jablonska, T. Suski, I. Gorczyca, N. E. Christensen, K. E. Attenkofer, R. C. C. Perera, E. M. Gullikson, J. H. Underwood, D. L. Ederer and Z. Liliental Weber, Phys. Rev. B: Condens. Matter Mater. Phys., 2000, 61, 16623–16632.
35. K.-H. Lee, J.-W. Chiou, J.-M. Chen, J.-F. Lee, A. Braud, K. Lorenz, E. Alves and I.-G. Chen, J. Am. Ceram. Soc., 2010, 93, 3531–3534.
36. L. Ravi, K. Boopathi, P. Panigrahi and B. Krishnan, Appl. Surf. Sci., 2018, 449, 213–220.
37. P. Li, T. Xiong, L. Wang, S. Sun and C. Chen, RSC Adv., 2020, 10, 2096–2103.
38. O. Elbanna, M. Fujitsuka and T. Majima, ACS Appl. Mater. Interfaces, 2017, 9, 34844–34854.
39. D. Zhang, X. Han, T. Dong, X. Guo, C. Song and Z. Zhao, J. Catal., 2018, 366, 237–244.
40. S. Chen, M. E. A. Warwick and R. Binions, Sol. Energy Mater. Sol. Cells, 2015, 137, 202–209.
41. X. Yuan, K. Luo, K. Zhang, J. He, Y. Zhao and D. Yu, J. Phys. Chem. A, 2016, 120, 7427–7433.
42. Z. Huang, Y. Zhang, H. Dai, Y. Wang, C. Qin, W. Chen, Y. Zhou and S. Yuan, J. Catal., 2019, 378, 331–340.
43. J. Hu, P. Yang and C. M. Lieber, Phys. Rev. B: Condens. Matter Mater. Phys., 1998, 57, R3185–R3188.
44. P. Praus, L. Svoboda, M. Ritz, I. Troppová, M. Šihor and K. Kočí, Mater. Chem. Phys., 2017, 193, 438–446.
45. K. Feng, H. Zhang, J. Gao, J. Xu, Y. Dong, Z. Kang and J. Zhong, Appl. Phys. Lett., 2020, 116, 191903.
46. Z. Chen, J. Zhao, C. R. Cabrera and Z. Chen, Small Methods, 2019, 3, 1800368.
47. M. Chen, W. Chang, Y. Shi, W. Liu and C. Li, Phys. Chem. Chem. Phys., 2021, 23, 11472–11478.
48. S. An, G. Zhang, T. Wang, W. Zhang, K. Li, C. Song, J. T. Miller, S. Miao, J. Wang and X. Guo, ACS Nano, 2018, 12, 9441–9450.
49. F. Hu, L. Leng, M. Zhang, W. Chen, Y. Yu, J. Wang, J. H. Horton and Z. Li, ACS Appl. Mater. Interfaces, 2020, 12, 54146–54154.
50. G. Gao, Y. Jiao, E. R. Waclawik and A. Du, J. Am. Chem. Soc., 2016, 138, 6292–6297.
51. W. Yan, L. Yan and C. Jing, Appl. Catal., B, 2019, 244, 475–485.
52. Y. Li, B. Li, D. Zhang, L. Cheng and Q. Xiang, ACS Nano, 2020, 14, 10552–10561.
53. B. Stöhr, H. P. Boehm and R. Schlögl, Carbon, 1991, 29, 707–720.
54. K. Dutta, M. Shahryari and J. Kopyscinski, Ind. Eng. Chem. Res., 2020, 59, 4245–4256.

---

Footnote

† Electronic supplementary information (ESI) available: XRD patterns of N sources, product productivity, N₂ adsorption/desorption isotherm, XANES spectra, XPS photolines, and TPO profiles. See DOI: 10.1039/d1cy01362a

---