Sodium hydroxide

Sodium doped flaky carbon nitride with nitrogen defects for enhanced photoreduction carbon dioxide activity

Jianbo Wang, Yanan Wang, Guojun Li, Yongheng Xiong, Mingjia Zhang, Shule Zhang ⇑, Qin Zhong ⇑

h i g h l i g h t s

● Sodium doped flaky g-C3N4 with nitrogen defects was rationally fabricated.
● The photocatalyst realizes the efficient reduction performance of CO2.
● The introduction of nitrogen defects promotes the increase of active sites.
● Sodium doping and flaky structure are conducive to electron transfer and accelerate the reaction.
● The photocatalyst shows good stability and reusability in CO2 reduction.

a b s t r a C t

Sodium doped flaky carbon nitride (g-C3N4) with nitrogen defects (bmw-DCN-x) were synthesized via two steps method to enhance photocatalytic reduction of carbon dioxide (CO2). After ball milling and cal- cination, dicyandiamide was evenly dispersed on the sodium chloride (NaCl) template to form a flaky structure. The NaCl not only provided part of sodium (Na) source for Na doped g-C3N4, but also intro- duced a large number of nitrogen (N) defects. Meanwhile, sodium hydroxide (NaOH) significantly enhanced Na doping. The bmw-DCN-30, a proportion of modified g-C3N4, showed heightened photo- reduction CO2 performance, with satisfactory carbon monoxide (CO) and methane (CH4) productivity at a rate of 30.6 lmol·g—1·h—1 and 5.4 lmol·g—1·h—1 respectively. This productivity was 15 and 11 times as much as that of bulky g-C3N4 (BCN). The related characterizations confirmed that N defects produced more reactive sites and enhanced the adsorption capacity of carbon nitride to CO2. The accompanying Na doping and flaky structure characteristics improved the optical absorption ability and the effective sep- aration of photogenerated carriers. Accordingly, this work provides further insights into constructing modified materials based on carbon nitride for CO2 reduction.

Keywords:
g-C3N4
N defects Na doping
Flaky structure
CO2 photoreduction

1. Introduction

The rapid development in industrialization over the past few decades has exerted, enormous pressure on several aspects of sustainable development, especially in the areas of environment and energy [1]. Most importantly, following rapid growth of global human population and the high industrial demand for fossil fuels, some energy resources are being exhausted [2]. Moreover, fossil fuels release abundant pollutants in the process of industrial uti- lization, which causes a series of environmental problems like car- bon emissions that may lead to climate change. With the increasing amounts of CO2 emissions, effects of global warming caused by the greenhouse effect are becoming more conspicuous which seriously threatens safety of human life [3]. To curb on the effects of fossil fuels, several remediation techniques have been proposed and tested. They include, but are not limited to, solar energy, ion exchange, membrane filtration, precipitation and pho- tocatalysis. Solar energy has been revealed to correspond appropri- ately with the trend of energy development in the future because of its rich, pollution-free, and endless characteristics. Elsewhere, photocatalytic technology has been indecisively studied in produc- tion of hydrogen, synthesizing ammonia [4,5] and degrading pollu- tants [6]. Therefore, from the perspective of sustainable development, it is of great significance to carry out photoreduction of CO2 for environment and energy.
Interestingly, the photocatalytic reduction of carbon dioxide produces C1 compounds such as methane, carbon monoxide, methanol and formic acid [7]. These chemical products have high value-added characteristics since they can be utilized as the initial raw materials of other products [8]. Several researches have reported about photocatalytic reduction of CO2. Nonetheless, according to the studies, the utilization rate of CO2 is relatively lower, and the yield is far less than that of photodecomposition of water. The low utilization and less yield have been attributed to high bond energy of C@O in CO2 molecule, multiple electron transfer processes and complexity of mechanism path [9]. There- fore, there is a need to come up with novel approaches that may be incorporated to improve on photocatalysis technology.
In recent years, two-dimensional (2D) materials have been widely exploited for their outstanding photoelectric and excep- tional physicochemical properties [10]. Currently, more 2D materi- als are conspicuously being used in photocatalysis research. Such material include, layered double hydroxide (LDH) [11], MXene [12], graphene derivatives [13] and graphitic carbon nitride (g- C3N4). Following its suitable energy band, comprehensive source of raw materials, easy synthesis, chemical stability and insoluble in different media, g-C3N4 has attracted a lot of interest in the field of photocatalysis globally [14]. However, the photocatalytic activ- ity of pristine g-C3N4 is unsatisfactory because of the poor effi- ciency of photogenerated carrier separation, high photogenerated charge recombination rate and narrow light absorption range [15]. To further improve separation efficiency of photogenerated carriers and accelerate electron transmission, studies focusing on metal or nonmetal doping [16], molecular copolymerization [17], nanoscale generation [18] and construction of heterojunction com- posite structure [19–21] have been emphasized. In particular, the heterojunction formed by the composite of g-C3N4 and other mate- rials is being widely adopted [22,23].
Meanwhile, constructing lattice defects is also a feasible method that may improve photocatalytic activity. The introduction of defects has been suggested to act as electron traps which pro- mote separation of electrons and holes. This is in addition to for- mation of defect energy levels and change in the band structure of g-C3N4. Abundant N defects, as primary adsorption sites, tend to activate reaction, especially for acidic oxides such as CO2. Niu et al. prepared carbon nitride with N defects by heat-treating bulky g-C3N4 in H2 atmosphere. Elsewhere, Lv et al. obtained the same results by argon resintering [24,25]. Chen et al. adopted magne- sium thermal denitrification technology to reduce the nitrogen content of g-C3N4 and then introduce N defects [26]. Liu et al. used NaBH4 high-temperature reduction and Zhang et al. utilized KOH base-assisted thermal polymerization to obtain N defects [27,28]. Luo et al. prepared Co — C3N4 nanosheets by Co-doped g C3N4 [29]. Jiang et al. prepared g-C3N4 doped with alkali metal ions (Na+) through high-temperature treatment [30].
In this paper, we prepared sodium assisted doping of layered g- C3N4 with N defects using a simple two-step process (NaCl grind- ing and NaOH drops). Since NaCl is easily washed away with water, the method achieves green utilization. Addition of NaCl intensely affects thermal polymerization of dicyandiamide. This possibly leads to the introduction of N defects on the edge of the interlayer plane of g-C3N4 and the formation of flaky structure of g-C3N4. On the other hand, moderate NaOH can promote more Na doping. It has been revealed that separation efficiency of electron-hole pair is improved, and the electron transfer rate is accelerated under vis- ible light irradiation. As compared with the pristine g-C3N4, the introduction of N defects created more adsorption sites and enhanced the photocatalytic reduction ability of CO2 for bmw- DCN-30. This work provided a simple and more effective way to improve the photocatalytic activity of g-C3N4 modified using the synergistic reaction of NaCl and NaOH.

2. Materials and methods

2.1. Materials

The chemical reagents were purchased from accredited suppli- ers. They included Dicyandiamide (99%) from Aladdin, Sodium hydroxide (NaOH, AR) from Shanghai HUSHI, Sodium hydroxide (NaCl, AR) from Shanghai Macklin Biochemical Co., Ltd and Tri- ethanolamine from Guangdong Guanghua Technology Co., Ltd. All the chemical reagents were used directly without further purification. The water used in all experiments was purified through a Millipore system.

2.2. Preparation of bulky g-C3N4 (BCN)

Precisely, 10.0 g of dicyandiamide were placed in a temperature-programmed tubular furnace for 4 h at 550℃ to achieve calcinations. A constant heating rate was maintained at 2.0 ℃·min—1 under N2 atmosphere and then bulky g-C3N4 (BCN) was obtained.

2.3. Preparation of g-C3N4 with nitrogen defects (bmw-DCN-X)

Four portions of 20.0 g sodium chloride were independently mixed with 2.0 g, 1.0 g, 0.667 g and 0.5 g of dicyandiamide by ball-milling at 300 r/ min for 2 h. Then 2 mL, 1 mL, 0.667 mL, and 0.5 mL of sodium hydroxide solution (20 g/L) were separately added into the above ball-milled mixed powder samples dropwise. The samples contents were stirred evenly and dried in a vacuum drying oven at 60℃ for 16 h. The dried sample was further ground and then calcined under the same conditions as aforementioned under BCN (See 2.2 above). After cooling to room temperature, the obtained product was washed with a large amount of deion- ized water, six times to remove superficial NaCl and thereafter it was collected using vacuum drying. The collected portions were marked as bmw-DCN-10, bmw-DCN-20, bmw-DCN-30, and bmw-DCN-40. (X: mass ratio of sodium chloride to dicyandiamide). A 1.0 g dicyandiamide was mixed with 1 mL of sodium hydrox- ide solution (20 g/L). The mixture content was stirred, dried and calcined (named wCN). A 0.667 g dicyandiamide with 20.0 g sodium chloride by ball-milling and the mixture content calcined. Then the product was washed and dried as previously described (named bmCN-30).

2.4. Characterizations

Powder X-ray diffraction (XRD) patterns were carried out on Bruker D2 PHASER. Field emission scanning electron microscopy (JSM-IT500HR) and high-resolution transmission electron micro- scopy (JEOLJEM 2100), with an energy dispersive X-ray spec- troscopy (EDS), were employed to estimate the composition and morphologies of synthesized materials. The Transmission electron microscope (TEM) was used to observe the fine structure of the samples. Brunner Emmet Teller (BET) measurements were per- formed on Micro ASAP 2460 with nitrogen adsorption at 77 K, and a pretreatment degassing temperature of 250℃. Fourier trans- form infrared (FTIR) spectroscopy was recorded on a Nicolet 6700 spectrometer. The UV–vis diffuse reflectance spectra (DRS) were conducted on a Shimadzu UV-2600 UV–vis spectrophotometer with a wavelength range of 220–800 nm. To approximate the assessment of sample element ratio, X-ray photoelectron spec- troscopy (XPS) measurements were carried out with a Thermo- VG Scientific Escalab 250 spectrometer. The CO2 temperature pro- grammed desorption (CO2-TPD) measurement with 0.1 g catalyst was performed on micromeritics AutoChem II 2920. Specific con- tent of elements was analyzed using a Perkin Elmer inductive cou- pled plasma emission spectrometer (ICP-OES). The electron paramagnetic resonance (EPR, Bruker EMX-10/12-type spectrome- ter) was used to characterize the electronic property of photocatalyst.

2.5. Photo-electrochemical performances

The photo-electrochemical measurements of required samples were carried out on a CHI 760D electrochemical workstation (Shanghai Chenhua, China). A standard three-electrode pool includes a working electrode, a Pt wire as counter electrode and a standard Ag/AgCl in saturated KCl as reference electrode. Herein, a 0.5 M Na2SO4 solution was used as the electrolyte. The working electrode was prepared by adding 5 mg of the catalyst into a mixture of 30 lL Nafion solution and 1 mL ethyl alcohol for 2 h under ultrasonically scattering. Then, the obtained slurry was dropped onto a 1 cm × 1 cm FTO glass electrode. This was realized with the help of scotch tape at a rate of 50 µl for six times. Finally, the obtained film was dried in a vacuum at 80℃ for 12 h. The Mott-Schottky (MS) was recorded at a frequency of 1000 Hz under dark conditions. The light on–off cycle per 20 s was employed in photocurrent (IT-Curve) under light irradiation (>420 nm). As well, the electrochemical impedance spectroscopies (EIS) were tested under an AC perturbation signal of 10 mV.

2.6. Photocatalytic performance

The photocatalytic activity was presented through photocat- alytic reduction of CO2 at room temperature (26℃) and negative atmospheric pressure (80 Kpa). Photocatalyst (10 mg), 2 mL of tri- ethanolamine (TEOA) and 20 mL H2O were added into a glass reac- tor (80 mL incapacity) in sequence. Then, the high-purity CO2 was vacuumed three times alternately to remove the residual air in the device. Next, CO2 was fully absorbed by the solution for 1 h before the lamp turns on. Finally, the reactor was filled with CO2 at a pres- sure of 80 Kpa, and the circulating water temperature was set at 10℃ in the glass reactor. A 300 W Xe arc lamp (Aulight CEL- HXF300, Beijing) equipped with a 400 nm cutoff filter was used as the light source. In the process of photocatalysis, the reaction system was stirred appropriately with a magnetic agitator. The

3. Results and discussion

The synthetic schematic process of sodium doped flaky g-C3N4 with nitrogen defects is illustrated in Scheme 1. In this system, dicyandiamide was first evenly dispersed in NaCl crystals after ball-milling. Subsequently, moderate sodium hydroxide was mixed and then calcined in nitrogen atmosphere. Owing to the higher melting point and sufficient content of NaCl, the g-C3N4 was extruded into flakes and produced a large number of nitrogen defects during the annealing process. The melting of NaOH during heating process may have further promoted the doping of sodium. Finally, the sodium doped flaky g-C3N4 with nitrogen defects was obtained after rinsing with water to dislodge the redundant NaCl crystals.
DCN-X (10, 20, 30 and 40) samples. No evident peak value of NaCl was observed in a series of samples, indicating that all of NaCl was removed after repeated washing. Two prominent diffraction peaks of BCN located at 13.1°and 27.4° were confirmed to be (1 0 0) and (0 0 2) planes, respectively. They corresponded to the in-plane repeated units and the interlayer stacking reflection like that of the graphite [31,32]. Comparing with BCN, the peaks of (1 0 0) and (0 0 2) in bmw-DCN-X weakened with the increase in the pro- portion of NaCl. These results might be attributable to the loss of ordered structures within the framework [28,33]. In addition, bmw-DCN-30 shows similar trend compared with wCN and bmCN-30 (Fig. S1a). Noticeably, all of the peaks in (0 0 2) plane did not shift, which may be associated with potential of doped Na cation to interact in a single layer without manipulating layer spacing [34]. In the present study, the introduction of NaCl and NaOH may have destroyed the g-C3N4 periodic arrangement of interlayer melon strands during the thermal polymerization of dicyandiamide. This is considered to be the main reason for the weakening of diffraction.
The bmw-DCN-X ( × : 10, 20, 30 or 40) has similar FTIR spectra to BCN as revealed in Fig. 1b. Previous studies have described related phenomenon. The strong vibration around 810 cm—1 was assigned to the breathing mode of tri-s-triazine units in g-C3N4 structure [35]. In the range of 1000–1650 cm—1, the main adsorp- tion band stems from C A N and C@N stretching vibrations in g- C3N4 layers, which were generally related to skeletal stretching vibrations of s-triazine/tri-s-triazine rings [36]. The broad peak of 3000–3400 cm—1 was attributed to the stretching vibration of — NH of amino-groups and –OH groups at the edge of g-C3N4 [37]. However, the obvious change was the development of a new peak at 2177 cm—1 in bmw-DCN-X compared with BCN, indi- cating a generation stretching vibration of cyano groups (–C„N) [38]. More so, in the current study, it can be observed in Fig. S1b that the peak of the cyano group exists in the case of NaCl or NaOH. The possible speculation is that sodium ions have induced partial decomposition of tri-s-triazine units. Therefore, the OH— from NaOH released with increased temperature may react with amine groups of urea-derived intermediates during the thermal polymer- ization of g-C3N4 [39].
Compared with the yellow color of pristine BCN, the modified g- C3N4 using the two-step method presented light brown. The mor- phology images of BCN and bmw-DCN-30 were observed through SEM and TEM. As seen from Fig. 2a-d, it is obvious that the BCN presented irregular block morphology, while bmw-DCN-30 was composed of sheet structure. The huge morphological difference between the BCN and bmw-DCN-30 is reflected in the TEM images in Fig. 2e-f. The different morphologies would directly lead to dif- ferent number of exposed active sites [46]. Previous studies have revealed that sodium ions are limited to the electron transport on the carbon nitride layer rather than interlamination [34]. This is so, especially for the sodium doped carbon nitride. A change in morphology from bulk to sheet, apart from favouring the adsorp- tion of CO2, provides better electron transfer conditions for subse- quent reduction of CO2. The EDS mapping of the C, N and Na elements in bmw-DCN-30 indicated that the distribution of the three elements was more uniform (Fig. S2).
To further investigate the surface chemical states and elemental composition of BCN and bmw-DCN-30, X-ray photoelectron spec- troscopy (XPS) measurements were tested. The wide scan XPS sur- vey spectra as shown in Fig. 3b, indicated clearly the existence of C, N and the appearance of Na (only in bmw-DCN-30) elements in the samples. To further demonstrate the introduction of marginal nitrogen defects, narrow scan Na1s, C1s and N1s XPS spectra were plotted and fitted as presented in Fig. 3c, d and e, respectively. The C1s spectra of BCN contains two peaks. The peaks were located at 284.8 and 288.3 eV, which were associated with adventitious hydrocarbon or C A C and sp2 bonded carbon (N — C@N) coordina- tion in the framework of g-C3N4, respectively [40,41]. Compared to BCN, a new peak at 286.4 eV appeared in bmw-DCN-30, corre- sponding to carbon in the form of C — NHx. This particular peak justified the formation of C„N groups (as displayed on FTIR) since cyano groups own similar C1s binding energies to C — NHx [42]. The N1s spectrum of two samples could be deconvolved into three peaks located at 398.9, 400.4 and 401.4 eV. Such peaks might be ascribed to sp2-hybridized nitrogen of C — N@C in tri-s-triazine rings, the tertiary nitrogen of N—(C)3 and NHx groups attached to C in the heptazine framework [43].
Moreover, a distinct peak of Na element was identified in the bmw-DCN-30. The high-resolution XPS spectrum of Na1s showed one peak at 1071.2 eV in Fig. 3d, which was assigned to the binding energy of NaN3 as reported earlier [44]. The specific content of Na, in team of ICP measurement, is presented in Table. S1. Comparison between bmw-DCN-30 and bmCN-30, revealed that the content of Na in bmw-DCN-30 nearly doubled. This observation implies that NaOH may promote Na doping. The Na atoms release the 3 s electrons in the outermost layer and then combine with in-planar N atoms having ionic bonds since the electronegativity of the N atom is stronger than C atom [34].
Considering that the melting point of NaOH was 318℃, at this temperature, the dicyandiamide had been converted to melamine, and the thermal polycondensation reaction (330℃) was about to occur. As the temperature rose, NaOH began to melt, but NaCl still maintained a cubic crystal structure, while g-C3N4 grew along the surface of NaCl. The molten NaOH might make the precursors of g- C3N4 to firmly adhere to NaCl, which ultimately promotes Na doped [28,45]. The N3c peak in the modified sample presented weaker than the pristine sample, which might be related to the for- mation of cyano group and the decrease of crystallinity in g-C3N4 [46]. It is worth noting that the intensity ratios of N2c/C were 0.84 and 1.15, respectively (Table. S2), meaning that N2c ratio on the surface of bmw-DCN-30 was lower. Furthermore, the result of electron paramagnetic resonance (EPR) was performed to reveal the unpaired-electron variation of bmw-DCN-30 caused by the nitrogen vacancies. As shown in Fig. 3a, compared with BCN, the strong Lorentzian curve EPR signal value for bmw-DCN-30 was detected at a g value of 2.0023. Generally, as found in literature, the intensity of spin for bmw-DCN-30 is obviously more remark- able than that for BCN [46,47]. The result indicates that unpaired electron concentration on carbon atoms of the heptane rings increases due to the formation of N vacancy. Therefore, the XPS analysis and EPR result present evidence for the construction of N vacancies in the bmw-DCN-30.
Nitrogen adsorption/desorption isotherms of BCN and bmw- DCN-30 were performed to appraise the specific surface area and pore size distribution. Two samples emerged as obvious IV adsorption/desorption isotherm. As shown in Fig. 4a, the specific surface areas of BCN and bmw-DCN-30 were 5.941 and 12.986 m2·g—1, respectively. The corresponding pore-size distribution of BCN and bmw-DCN-30 were 0.0355 cm3/g and 0.0760 cm3/g, respectively. Combined with the data analysis, the specific surface area and pore volume of modified g-C3N4 were not significantly improved com- pared with original g-C3N4. As suggested in previous studies, this result may be due to the large size of NaCl cubic template [48].
Although the specific surface area had not been greatly improved, it did not mean that the difference in CO2 absorption capacity was not significant. The CO2 temperature-programmed desorption (TPD) was regarded as a key characterization to esti- mate CO2 absorption capacity. The desorption peaks of BCN and bmw-DCN-30 are shown in Fig. 4b. Based on the figure, BCN had only one weak desorption peak located at 85℃, which may be attributed to physical adsorption on the sample surface [49]. Inter- estingly, apart from physical adsorption, bmw-DCN-30 possessed chemical adsorption located around 155℃ [50]. Compared to
According to the existing literature reports, after Na doping, the atomic orbital of Na may be situated in the energy gap of g-C3N4. As well, orbital hybridization may be carried out between the dopant orbital and the molecular orbital of carbon nitride, resulting in the alteration of valance band and conduction band [55,56]. Hence, in the current study, Mott-Schottky measurements were used to further measure the specific conduction band of the pre- pared samples and then account for its valance band through for- mula derivation (EVB = Eg + ECB). As displayed in Fig. 6a, the slope values of Mott-Schottky plots were all positive, signifying that the several samples were n-type semiconductors [57]. The flat band potentials of BCN and bmw-DCN-30 as estimated at baseline (Y-axis = 0) and finding the intercept were around — 0.59 and — 0.85 V vs. Ag/AgCl. The scale of the normal hydrogen elec- trode (NHE) was obtained using the following formula transformation: of basic sites on the photocatalyst increased markedly. Similar with Lewis acid, CO2 was commonly adsorbed on the basic sites. Mean- while, Na atoms and N defects might play the role of basic position, which greatly improves the ability of the photocatalyst to absorb and activate CO2 [47,51]. The disappearance of some N atoms in the heprazine ring unit resulted in the accumulation of localized electrons. However, Na atoms existed in the in-planes without direct interaction with CO2, as opposed to metal oxides, which mainly promoted electron transfer [52]. Therefore, in the process of thermal polymerization of NaCl and NaOH, it is particularly meaningful that some basic sites created through the introduction of appropriate N defects improve the activity of the photocatalyst. The UV — vis DRS spectra were implemented to explore the optical absorption properties of a series of samples, as shown in Fig. 5. The figure reveals that the study samples could be as photo- catalysts with typical semiconductor optical absorption character- istics [48]. According to Tauc formula (ahm) = A(hm — Eg)n/2, the energy gap of BCN, wCN and bmCN-30 were about 2.43 eV, 2.44 eV and 2.50 eV, respectively. The Eg of g-C3N4 modified by two steps decreased gradually following the increase in amount of NaCl, which were 2.39 eV, 2.34 eV, 2.33 eV and 2.30 eV, respec- tively. Owing to the decrease in size and thickness of g-C3N4, the quantum confinement effect was enhanced, which enlarged the energy gap [53]. On the other hand, the introduction of abundant cyano group defects and Na-doping may enhance optical absorp- tion [54,55]. Generally, the synthesis strategy has no significant effect on expansion of the photoresponse. conductor, since CB potential (ECB) is 0.2 V negative more than Vfb, the calculated ECB of BCN and bmw-DCN-30 were — 0.59 and — 0.85 eV vs. RHE, respectively. Correspondingly, EVB of BCN and bmw-DCN-30 obtained were 1.84 and 1.48 eV through the con- version of calculation Equation. Band structure alignments for all samples are shown in Fig. 6b. The results show that the ECB of mod- ified g-C3N4 is more negative than that of BCN, which means that the conduction band is more favorable for CO2 reduction [59].
To get to the bottom of understand the relationship between the charge carrier transport behaviors and prepared products, elec- trochemical impedance spectroscopy (EIS) measurements were conducted without lamp on. As shown in Fig. 6c, bmw-DCN-30 possessed the smallest arc radius among the series of photocata- lyst, which illustrated the minimum resistance to charge motion in bmwDCN-30 [60]. The excellent conductivity is beneficial to the transfer of photoinduced carriers. In addition, the transient photocurrent density response measurements were performed by switching on the light source at an interval of 20 s. As demon- strated in Fig. 6d, the photocurrent density response of bmw- DCN-30 was larger than that of BCN, which demonstrated that bmw-DCN-30 has a vastly improved efficiency in electron-hole separation and charge transfer as compared with BCN. Interest- ingly, with the increase of NaCl content, both the transient pho- tocurrent density response and electrochemical impedance spectroscopy exhibited similar characteristics. The impedance gradually decreased and the photocurrent density increased. Sev- eral factors could be linked to these results. For instance, Na ions on the heprazine ring cavity may have accelerated the migration of photogenerated carriers to the surface. As well, potential of flaky structure of the modified carbon nitride to favor is more favorable to Na doping which only transports electrons on the surface [34]. Furthermore, the cyano groups can promote electron delocaliza- tion in g-C3N4, boosting the separation of photoinduced electrons-holes [61]. However, for bmw-DCN40, it is completely opposite, which may be due to the excessive introduction of N defects, resulting in more serious carrier recombination, including the decrease of crystallinity [28]. It illustrates once again that only a moderate amount of N defects can guarantee the maximum improvement of activity.
The yields of carbon monoxide and methane were used to rep- resent the CO2 reduction performance of the as-prepared photo- catalysis in a high purity CO2 gas closed-loop system. The reaction gas-phase products were pushed by the carrier gas into gas chromatography for detection at 2 h intervals. The external standard method was used for quantitative analysis of the products. As seen in Fig. 7, among these samples, the pristine BCN exhi- bit tiny activity with CO (2.1 lmol·g—1·h—1) and CH4 (0.5 lmol·g—1·h—1) generation rates. After the introduction of N defects and Na ions, the activity of g-C3N4 was obviously improved.
Apparently, the bmw-DCN-30 appeared at the maximum forma- tion rate of CO (30.6 lmol·g—1·h—1) and CH4 (5.4 lmol·g—1·h—1). For the bmw-DCN-30, the apparent quantum yield of CO and CH4 were estimated to be 0.103% and 0.071% respectively, both at 420 nm. The bmw-DCN-30 was almost 15 times as much as BCN in terms of CO production, and the yield of CH4 was 11 times.
Hence, one can see that, the synthesis strategy is extraordinarily significant for the improvement of photocatalytic reduction of CO2. Moreover, the photocatalytic performance of wCN and bmCN-30 samples were investigated under the same condition and results presented in Fig. S4. The results indicate that, compared with BCN, the activity of the two samples were improved to some extent as compared with BCN, but not as high as that of bmw-DCN- 30. In order to further determine the carbon source of the products, control experiments divided into two groups (a-b) were carried out without CO2 (replaced by Ar gas): (a) H2O + TEOA + Ar + visble light, (b) H2O + TEOA + Ar + bmw-DCN-30 + visble light. The CO and CH4 were barely generated in the above experiments (Fig. S7), indicating that the carbon source of products almost orig- inates from the used CO2. Based on the results of the comparison between a series of activity data and the support of characteriza- tion analysis, link the following factors to the enhanced photocat- alytic efficiency of modified g-C3N4: (1) an appropriate amount of N defects that provides enough active sites. (2) Na doping that enhance carrier transport rate. (3) The flaky structure that favors for electron transport on the surface.
The photocatalytic reduction of CO2 over bmw-DCN-30 was carried out in 4 cycles to verify the stability. As displayed in Fig. 8b, the third and fourth time activity decreased a little, which could be considered relatively stable. Furthermore, XRD patterns and FTIR spectrum were measured to demonstrate the stability of bmw-DCN-30. As shown in Fig. S6, there was no significant change in XRD patterns and FTIR spectrum of the bmw-DCN-30, before and after the reaction are observed, which indicates again that the photocatalyst possess fine stability.

4. Conclusion

In summary, the sodium doped flaky g-C3N4 with nitrogen defects was prepared through NaCl and NaOH assisted thermal treatment of dicyandiamide. The CO2 reduction activity of the modified carbon nitride was significantly high than that of the pris- tine g-C3N4, resulting from the multiple effects of N defects, laky structure and Na doping. The physical property of NaCl is easily washed away and thus, can support the synthesis of flake morphol- ogy. Thus further Na doping and N defects were introduced to modify carbon nitride. The modified carbon nitride possessed stronger adsorption capacity, a narrower energy gap and better electrochemical performance. In this way, the mechanism that defects are utilized for adsorption and activation, flaky structure and sodium concentration on the surface accelerates electron transport, thus improving the efficiency of photocatalytic reduc- tion of CO2. Compared with the previous work[48,53,54], this study further verified the adsorption of CO2 using nitrogen defects, which was verified by XPS, TPD and EPR. Hence, the design of element doped carbon nitride materials with N defects may have broad application prospects in the field of CO2 conversion and other pho- tocatalytic applications.

References

[1] K. Li, B.S. Peng, T.Y. Peng, Recent Advances in Heterogeneous Photocatalytic CO2 Conversion to Solar Fuels, ACS Catal 6 (11) (2016) 7485–7527.
[2] M. Majdoub, Z. Anfar, A. Amedlous, Emerging Chemical Functionalization of g- C3N4: Covalent/Noncovalent Modifications and Applications, ACS Nano 14 (10) (2020) 12390–12469.
[3] A. Kumar, A. Kumar, V. Krishnan, Perovskite Oxide Based Materials for Energy and Environment-Oriented Photocatalysis, ACS Catal 10 (17) (2020) 10253– 10315.
[4] L. Chen, X. Dai, X. Li, J. Wang, H. Chen, X. Hu, H. Lin, Y. He, Y. Wu, M. Fan, A novel Bi2S3/KTa0.75Nb0.25O3 nanocomposite with high efficiency for photocatalytic and piezocatalytic N2 fixation, J. Mater. Chem. A. (2021).
[5] W. Zhang, P. Xing, J. Zhang, L. Chen, J. Yang, X. Hu, L. Zhao, Y. Wu, Y. He, Facile preparation of novel nickel sulfide modified KNbO3 heterojunction composite and its enhanced performance in photocatalytic Sodium hydroxide nitrogen fixation, J. Colloid Interface Sci. 590 (2021) 548–560.
[6] L. Chen, W. Zhang, J. Wang, X. Li, Y. Li, X. Hu, L. Zhao, Y. Wu, Y. He, High piezo/ photocatalytic efficiency of Ag/Bi5O7I nanocomposite using mechanical and solar energy for N2 fixation and methyl orange degradation, Green, Energy Environ. (2021).
[7] M. Marszewski, S.W. Cao, J.G. Yu, M. Jaroniec, Semiconductor-based photocatalytic CO2 conversion, Mater. Horiz. 2 (3) (2015) 261–278.
[8] Y.Q. Zhu, C. Romain, C.K. Williams, Sustainable polymers from renewable resources, Nature 540 (7633) (2016) 354–362.
[9] J. Fu, J. Yu, C. Jiang, B. Cheng, g-C3N4-Based Heterostructured Photocatalysts, Adv. Energy Mater. 8 (3) (2018).
[10] W.-J. Ong, K.P.Y. Shak, 2D/2D Heterostructured Photocatalysts: An Emerging Platform for Artificial Photosynthesis, Sol. RRL. 4 (8) (2020).
[11] S.F. Ng, M.Y.L. Lau, W.J. Ong, Engineering Layered Double Hydroxide-Based Photocatalysts Toward Artificial Photosynthesis: State-of-the-Art Progress and Prospects, Sol, RRL, 2021.
[12] L. Biswal, S. Nayak, K. Parida, Recent progress on strategies for the preparation of 2D/2D MXene/g-C3N4 nanocomposites for photocatalytic energy and environmental applications, Catalysis Sci Tech. 11 (4) (2021) 1222–1248.
[13] W.J. Ong, L.K. Putri, A.R. Mohamed, Rational Design of Carbon-Based 2D Nanostructures for Enhanced Photocatalytic CO2 Reduction: A Dimensionality Perspective, Chemistry 26 (44) (2020) 9710–9748.
[14] X.C. Wang, S. Blechert, M. Antonietti, Polymeric Graphitic Carbon Nitride for Heterogeneous Photocatalysis, ACS Catal 2 (8) (2012) 1596–1606.
[15] J.J. Zhu, P. Xiao, H.L. Li, S.A.C. Carabineiro, Graphitic Carbon Nitride: Synthesis, Properties, and Applications in Catalysis, ACS Appl. Mater. Inter. 6 (19) (2014) 16449–16465.
[16] K. Wang, Q. Li, B.S. Liu, B. Cheng, W.K. Ho, J.G. Yu, Sulfur-doped g-C3N4 with enhanced photocatalytic CO2-reduction performance, Appl. Catal. B. 176 (2015) 44–52.
[17] C. Ye, J.X. Li, H.L. Wu, X.B. Li, B. Chen, C.H. Tung, L.Z. Wu, Enhanced Charge Separation Efficiency Accelerates Hydrogen Evolution from Water of Carbon Nitride and 3,4,9,10-Perylene-tetracarboxylic Dianhydride Composite Photocatalyst, ACS Appl. Mater. Inter. 10 (4) (2018) 3515–3521.
[18] I. Papailias, N. Todorova, T. Giannakopoulou, N. Ioannidis, N. Boukos, C.P. Athanasekou, D. Dimotikali, C. Trapalis, Chemical vs thermal exfoliation of g- C3N4 for NOx removal under visible light irradiation, Appl. Catal. B. 239 (2018) 16–26.
[19] Y. Wang, L. Guo, Y. Zeng, H. Guo, S. Wan, M. Ou, S. Zhang, Q. Zhong, Amino- Assisted NH2-UiO-66 Anchored on Porous g-C3N4 for Enhanced Visible-Light- Driven CO2 Reduction, ACS. Appl. Mater. Inter. 11 (34) (2019) 30673–30681.
[20] D.R. Paul, S. Gautam, P. Panchal, S.P. Nehra, P. Choudhary, A. Sharma, ZnO- Modified g-C3N4: A Potential Photocatalyst for Environmental Application, ACS Omega 5 (8) (2020) 3828–3838.
[21] H. Shi, G. Chen, C. Zhang, Z. Zou, Polymeric g-C3N4 Coupled with NaNbO3 Nanowires toward Enhanced Photocatalytic Reduction of CO2 into Renewable Fuel, ACS Catal 4 (10) (2014) 3637–3643.
[22] R. Acharya, K. Parida, A review on TiO2/g-C3N4 visible-light- responsive photocatalysts for sustainable energy generation and environmental remediation, J. Environ Chem. Eng. 8 (4) (2020).
[23] M. Que, W. Cai, J. Chen, L. Zhu, Y. Yang, Recent advances in g-C3N4 composites within four types of heterojunctions for photocatalytic CO2 reduction, Nanoscale 13 (14) (2021) 6692–6712.
[24] C. Lv, Y. Qian, C. Yan, Y. Ding, Y. Liu, G. Chen, G. Yu, Defect Engineering Metal- Free Polymeric Carbon Nitride Electrocatalyst for Effective Nitrogen Fixation under Ambient Conditions, Angew. Chem. Int. Ed. Engl. 57 (32) (2018) 10246– 10250.
[25] P. Niu, L.C. Yin, Y.Q. Yang, G. Liu, H.M. Cheng, Increasing the visible light absorption of graphitic carbon nitride (melon) photocatalysts by homogeneous self-modification with nitrogen vacancies, Adv. Mater. 26 (47) (2014) 8046–8052.
[26] J. Chen, Z. Mao, L. Zhang, D. Wang, R. Xu, L. Bie, B.D. Fahlman, Nitrogen- Deficient Graphitic Carbon Nitride with Enhanced Performance for Lithium Ion Battery Anodes, ACS Nano 11 (12) (2017) 12650–12657.
[27] G.G. Liu, G.X. Zhao, W. Zhou, Y.Y. Liu, H. Pang, H.B. Zhang, D. Hao, X.G. Meng, P. Li, T. Kako, J.H. Ye, In Situ Bond Modulation of Graphitic Carbon Nitride to Construct p-n Homojunctions for Enhanced Photocatalytic Hydrogen Production, Adv. Funct. Mater. 26 (37) (2016) 6822–6829.
[28] H.J. Yu, R. Shi, Y.X. Zhao, T. Bian, Y.F. Zhao, C. Zhou, G.I.N. Waterhouse, L.Z. Wu, C.H. Tung, T.R. Zhang, Alkali-Assisted Synthesis of Nitrogen Deficient Graphitic Carbon Nitride with Tunable Band Structures for Efficient Visible-Light-Driven Hydrogen Evolution, Adv. Mater. 29 (16) (2017).
[29] M. Luo, Y. Bai, R. Sun, Z. Wang, W. Sun, P. Lin, X. Dai, K. Sun, Enhanced Performance of Lithium-Sulfur Batteries with Co-Doped g-C3N4 Nanosheet- Based Separator, Ind. Eng. Chem. Res. 60 (3) (2021) 1231–1240.
[30] J. Jiang, S. Cao, C. Hu, C. Chen, A comparison study of alkali metal-doped g-C3N4 for visible-light photocatalytic hydrogen evolution, Chinese J. Catal. 38 (12) (2017) 1981–1989.
[31] M.J. Bojdys, J.O. Muller, M. Antonietti, A. Thomas, Ionothermal synthesis of crystalline, condensed, graphitic carbon nitride, Chemistry 14 (27) (2008) 8177–8182.
[32] Z.Y. Fang, D. Li, R.J. Chen, Y.Y. Huang, B.F. Luo, W.D. Shi, Multiple Doped Carbon Nitrides with Accelerated Interfacial Charge/Mass Transportation for Boosting Photocatalytic Hydrogen Evolution, ACS Appl. Mater. Inter. 11 (25) (2019) 22255–22263.
[33] D. Zhang, Y.L. Guo, Z.K. Zhao, Porous defect-modified graphitic carbon nitride via a facile one-step approach with significantly enhanced photocatalytic hydrogen evolution under visible light irradiation, Appl. Catal. B. 226 (2018) 1–9.
[34] T. Xiong, W.L. Cen, Y.X. Zhang, F. Dong, Bridging the g-C3N4 Interlayers for Enhanced Photocatalysis, ACS Catal 6 (4) (2016) 2462–2472.
[35] J.R. Holst, E.G. Gillan, From triazines to heptazines: deciphering the local structure of amorphous nitrogen-rich carbon nitride materials, J. Am. Chem. Soc. 130 (23) (2008) 7373–7379.
[36] Y.C. Zhao, Z. Liu, W.G. Chu, L. Song, Z.X. Zhang, D.L. Yu, Y.J. Tian, S.S. Xie, L.F. Sun, Large-scale synthesis of nitrogen-rich carbon nitride microfibers by using graphitic carbon nitride as precursor, Adv. Mater. 20 (9) (2008) 1777.
[37] L. Luo, A.F. Zhang, M.J. Janik, C.S. Song, X.W. Guo, Facile fabrication of metal- free urchin-like g-C3N4 with superior photocatalytic activity, RSC Adv 6 (97) (2016) 94496–94501.
[38] W. Lei, D. Portehault, R. Dimova, M. Antoniettit, Boron Carbon Nitride Nanostructures from Salt Melts: Tunable Water-Soluble Phosphors, J. Am. Chem. Soc. 133 (18) (2011) 7121–7127.
[39] Y.X. Li, H. Xu, S.X. Ouyang, D. Lu, X. Wang, D.F. Wang, J.H. Ye, In situ surface alkalinized g-C3N4 toward enhancement of photocatalytic H2 evolution under visible-light irradiation, J. Mater. Chem. A. 4 (8) (2016) 2943–2950.
[40] Y. Xu, Y.Y. Gong, H. Ren, W.B. Liu, L.Y. Niu, C. Li, X.J. Liu, In situ structural modification of graphitic carbon nitride by alkali halides and influence on photocatalytic activity, RSC Adv 7 (52) (2017) 32592–32600.
[41] Y.M. He, Y. Wang, L.H. Zhang, B.T. Teng, M.H. Fan, High-efficiency conversion of CO2 to fuel over ZnO/g-C3N4 photocatalyst, Appl. Catal. B. 168 (2015) 1–8.
[42] A. Vinu, Two-dimensional hexagonally-ordered mesoporous carbon nitrides with tunable pore diameter, surface area and nitrogen content, Adv. Funct. Mater. 18 (5) (2008) 816–827.
[43] X.M. Bu, Y. Bu, S.W. Yang, F. Sun, L.F. Tian, Z. Peng, P. He, J. Sun, T. Huang, X.Y. Wang, G.Q. Ding, J.H. Yang, X.M. Xie, Graphitic carbon nitride nanoribbon for enhanced visible-light photocatalytic H2 production, RSC Adv 6 (113) (2016) 112210–112214.
[44] F. Guo, J.L. Chen, M.W. Zhang, B.F. Gao, B.Z. Lin, Y.L. Chen, Deprotonation of g- C3N4 with Na ions for efficient nonsacrificial water splitting under visible light, J. Mater. Chem. A. 4 (28) (2016) 10806–10809.
[45] H.L. Gao, S.C. Yan, J.J. Wang, Y.A. Huang, P. Wang, Z.S. Li, Z.G. Zou, Towards efficient solar hydrogen production by intercalated carbon nitride photocatalyst, Phys. Chem. Chem. Phys. 15 (41) (2013) 18077–18084.
[46] S.P. Wan, M. Ou, Y.N. Wang, Y.Q. Zeng, Y.H. Xiong, F.J. Song, J. Ding, W. Cai, S.L. Zhang, Q. Zhong, Protonic acid-assisted universal synthesis of defect abundant multifunction carbon nitride semiconductor for highly-efficient visible light photocatalytic applications, Appl. Catal. B. 258 (2019).
[47] W. Ma, N. Wang, Y. Guo, L.Q. Yang, M.F. Lv, X. Tang, S.T. Li, Enhanced photoreduction CO2 activity on g-C3N4: By synergistic effect of nitrogen defective-enriched and porous structure, and mechanism insights, Chem. Eng. J. 388 (2020).
[48] H.X. Ma, Y.F. Jia, G.Q. Zhu, F.C. Zhang, S. Rhee, B. Lee, C.L. Liu, Study of cyano and hydroxyl groups modification on the properties of porous carbon nitride synthesized by using a salt assistant method, Appl. Surf. Sci. 507 (2020).
[49] Z.F. Jiang, H.L. Sun, T.Q. Wang, B. Wang, W. Wei, H.M. Li, S.Q. Yuan, T.C. An, H.J. Zhao, J.G. Yu, P.K. Wong, Nature-based catalyst for visible-light-driven photocatalytic CO2 reduction, Energ. Environ. Sci. 11 (9) (2018) 2382–2389.
[50] J. Hu, J. Ding, Q. Zhong, Ultrathin 2D Ti3C2 MXene Co-catalyst anchored on porous g-C3N4 for enhanced photocatalytic CO2 reduction under visible-light irradiation, J. Colloid Interface Sci. 582 (2021) 647–657.
[51] Z.X. Sun, J.M.T.A. Fischer, Q.A. Li, J. Hu, Q.J. Tang, H.Q. Wang, Z.B. Wu, M. Hankel, D.J. Searles, L.Z. Wang, Enhanced CO2 photocatalytic reduction on alkali-decorated graphitic carbon nitride, Appl. Catal. B. 216 (2017) 146–155.
[52] K. Wang, J.L. Fu, Y. Zheng, Insights into photocatalytic CO2 reduction on g- C3N4: Strategy of simultaneous B, K co-doping and enhancement by N vacancies, Appl. Catal. B. 254 (2019) 270–282.
[53] X.Y. Qian, X.Q. Meng, J.W. Sun, L.L. Jiang, Y.N. Wang, J.L. Zhang, X.M. Hu, M.Shalom, J.W. Zhu, Salt-Assisted Synthesis of 3D Porous g-C3N4 as a Bifunctional Photo and Electrocatalyst, ACS Appl. Mater. Inter. 11 (30) (2019) 27226–27232.
[54] F. Yang, D.Z. Liu, Y.X. Li, L.J. Cheng, J.H. Ye, Salt-template-assisted construction of honeycomb-like structured g-C3N4 with tunable band structure for enhanced photocatalytic H2 production, Appl. Catal. B. 240 (2019) 64–71.
[55] Z. Shu, Y. Wang, W.B. Wang, J. Zhou, T.T. Li, X.Q. Liu, Y.G. Tan, Z.L. Zhao, A green one-pot approach for mesoporous g-C3N4 nanosheets with in situ sodium doping for enhanced photocatalytic hydrogen evolution, Int. J. Hydrog. Energy. 44 (2) (2019) 748–756.
[56] J. Zhang, S.Z. Hu, Y.J. Wang, A convenient method to prepare a novel alkali metal sodium doped carbon nitride photocatalyst with a tunable band structure, RSC Adv 4 (108) (2014) 62912–62919.
[57] K.Q. Li, Y.Q. Jiang, Y.D. Li, Z. Wang, X. Liu, P. Wang, D.B. Xia, R.Q. Fan, K.F. Lin, Y. L. Yang, In situ preparation of graphitic carbon nitride bonded with cyano groups for enhanced photocatalytic activity, Int. J. Hydrog. Energy. 45 (16) (2020) 9683–9694.
[58] Y.A. Wang, Y.Q. Zeng, S.P. Wan, W. Cai, F.J. Song, S.L. Zhang, Q. Zhong, In Situ Fabrication of 3D Octahedral g-C3N4/BiFeWOx Double-Heterojunction for Highly Selective CO2 Photoreduction to CO Under Visible Light, Chemcatchem 10 (20) (2018) 4578–4585.
[59] Y. Xie, Y.F. Zhuo, S.W. Liu, Y.N. Lin, D.R. Zuo, X. Wu, C.H. Li, P.K. Wong, Ternary g-C3N4/ZnNCN@ZIF-8 Hybrid Photocatalysts with Robust Interfacial Interactions and Enhanced CO2 Reduction Performance, Sol. RRL. 4 (8) (2020).
[60] X.L. Yang, F.F. Qian, G.J. Zou, M.L. Li, J.R. Lu, Y.M. Li, M.T. Bao, Facile fabrication of acidified g-C3N4/g-C3N4 hybrids with enhanced photocatalysis performance under visible light irradiation, Appl. Catal. B. 193 (2016) 22–35.
[61] Z.F. Chen, S.C. Lu, Q.L. Wu, F. He, N.Q. Zhao, C.N.A. He, C.S. Shi, Salt-assisted synthesis of 3D open porous g-C3N4 decorated with cyano groups for photocatalytic hydrogen evolution, Nanoscale 10 (6) (2018) 3008–3013.