Visible-light-activated N-doped CQDs/g-C₃N₄/Bi₂WO₆ nanocomposites with different component arrangements for the promoted degradation of hazardous vapors

1. Introduction

Graphitic carbon nitride (g-C₃N₄, CN) sheets have been frequently utilized as a semiconductor photocatalyst for the neutralization of hazardous pollutants due to their outstanding optical and electronic properties, satisfactory flexibility, and excellent thermal and chemical stability [[1], [2], [3]]; in addition, this material is activated under visible-light illumination and fabricated by the facile pyrolysis of nitrogen-containing organic reagents [1]. However, the practicable utilization of pristine CN is hindered by the rapid recombination rates of charge carriers, thereby leading to unsatisfactory photocatalytic performance [3].

The disadvantage of CN can be overcome by its modification with other semiconductors to afford semiconductor/semiconductor heterojunction photocatalysts [1]. Fortunately, CN can be combined easily with other semiconductors to develop heterojunction structures with strong interactions owing to its flexible two-dimensional (2D) geometric structure [3]. Meanwhile, bismuth tungstate (Bi₂WO₆, BWO) demonstrates promise as an agent for coupling with 2D CN to afford visible-light-responsive heterojunction materials with promoted photocatalytic performances [[4], [5], [6]]. The abovementioned semiconductor photocatalyst exhibits a high charge mobility and a large specific surface area due to its lamellar structure [5,7]; in addition, this material exhibits good thermal and chemical stability due to its accumulated layers of (WO₄)^2- sheets and perovskite-like (Bi₂O₂)²⁺ sheets [6,8]. Moreover, a Z-scheme visible-light-responsive CN/BWO heterojunction exhibits superior photocatalytic ability compared with that of pristine CN or BWO for the degradation of aqueous rhodamine B under visible-light illumination [5].

Nevertheless, CN and BWO in the CN/BWO dual nanocomposites exhibit medium band gaps of ~2.70 and 2.73 eV, respectively, still requiring the promotion of light-absorption capability in the visible region [3,5,9]. Photosensitization techniques by the combination of light-harvesting materials can be employed to better use visible light in CN/BWO heterojunctions. Typically, synthetic dye materials and metallic semiconductor quantum dots have been employed as photosensitizers for various photocatalysts [10]. However, these conventional photosensitization materials are typically hindered by the low photochemical stability of dye materials or highly toxic effects of heavy metals in metallic semiconductor quantum dots [10,11]. Non-metallic carbon quantum dots (CQDs), which are fluorescent quasi-spherical carbon nanoparticles with a graphitic core structure, are emerging alternatives due to their excellent properties, including broad optical absorption due to photoluminescence emission, low toxicity, photochemical stability, cost-effectiveness, tunable surface functionalization, fluorescence properties, and facile synthesis [11,12]. Additionally, CQDs are easily dispersed in water owing to their hydrophilic properties, facilitating their processes in an aqueous solution for combination with various semiconductor photocatalysts [11]. More recently, some researchers have reported that doping CQDs with nitrogen can induce the promotion of electron mobility and reactive sites [[13], [14], [15]]. These superior features of N-doped CQDs (NCQDs) prompted their combination with the CN/BWO dual nanocomposites for the promotion of the photocatalytic ability toward the treatment of hazardous vapors.

Moreover, notably, the arrangement of components in the multi-composite photocatalysts may affect their efficiencies because different photocatalyst structures can lead to different charge-transfer routes. Hence, ternary nanocomposite photocatalysts comprising NCQDs, CN, and BWO (referred to as NCQDs/CN/BWO) with two component arrangements were developed, and their photocatalytic abilities for the degradation of two model hazardous vapors, i.e., n-undecane (nUD) and m-xylene (mXYL), respectively, under visible-light irradiation were investigated. These target contaminants were examined because of their abundance in the atmospheric air and hazardous nature [[16], [17], [18]]. For comparison, five individual catalysts (i.e., CN, BWO, NCQDs/BWO, NCQDs/CN, and CN/BWO, respectively) were also examined for the decomposition of the hazardous vapors. In addition, plausible reaction mechanisms over two types of NCQDs/CN/BWO photocatalysts were suggested on the basis of the band energy diagram and hydroxyl radical production analyses; furthermore, the photochemical stability of NCQDs/CN/BWO was inspected by its consecutive use for the degradation of model pollutants.

2. Experimental

2.1. Material synthesis

The NCQDs sample was constructed by a hydrothermal process. Briefly, citric acid (C₆H₈O₇, 99.5 %, 3.15 g) and ethanediamine (C₂H₈N₂, 99.0 %, 1.0 mL) were delivered to deionized water (DW, 30 ml) under stirring. Next, the solution was transferred into an autoclave vessel and heated at 180 °C for 6 h. After cooling, the mixture was subjected to centrifugation at 13,000 rpm for 10 min to afford NCQDs solution as the supernatant. Subsequently, the supernatant NCQDs suspension was dried at 80 °C for 20 h, affording the NCQDs powder. In addition, unmodified CQDs were prepared following this same procedure but without the addition of the nitrogen precursor (ethanediamine).

The sheet CN powder was fabricated via a thermal polymerization. For synthesizing bulk CN, melamine (5.0 g) was delivered to an alumina vessel, placed in a muffle furnace, and heated at 500 °C for 0.5 h and then at 550 °C for 3 h. After cooling, the light yellow bulk CN was ground to obtain a fine powder. Subsequently, bulk CN was subjected to calcination at 500 °C for 2 h, cooled, and ground to provide a CN sheet sample.

BWO sample was constructed using a hydrothermal process. Briefly, bismuth(III) nitrate pentahydrate (Bi(NO₃)₃⸳5H₂O, ≥ 98.0 %, 1 g) was first added to acetic acid (35 mL) under stirring (Solution A). Second, sodium tungstate dihydrate (Na₂WO₄⸳2H₂O, 99 %, 0.35 g) was added to DW under stirring (Solution B). Solution B was slowly dropped into Solution A. Subsequently, this mixture was delivered to an autoclave unit (110 mL) and heated at 170 °C for 24 h. After cooling, the suspended solution was subjected to filtration, washed with DW/ethanol, and dried at 85 °C for 20 h, affording BWO powder.

CN/BWO samples were fabricated using sonication-assisted wet impregnation. In brief, BWO powder (600 mg) and a designated amount of CN powder were delivered to methanol (120 mL), after which this solution was subjected to sonication for 12 h. After placing this suspension in a fume hood for 15 h under stirring, it was subjected to vacuum filtration, washed with DW/ethanol, and dried at 85 °C for 20 h. CN/BWO samples fabricated with CN/(BWO + CN) weight ratios of 0.1, 0.3, 0.5, and 0.7 were referred to as CN/BWO-0.1, CN/BWO-0.3, CN/BWO-0.5, and CN/BWO-0.7, respectively.

Two NCQDs/CN/BWO photocatalysts with different combinations were developed by the combination of NCQDs/BWO and CN for type I and CN/BWO-0.5 and NCQDs for type II. First, NCQDs/BWO was fabricated by the same procedure as that used to fabricate CN/BWO, except that NCQDs powder was substituted for CN powder; in addition, NCQDs/CN was constructed by the same procedure as that used to construct NCQDs/BWO, except that CN powder was substituted for BWO powder. As-synthesized NCQDs/BWO powder (609 mg) and CN powder (591 mg) were added into methanol (120 mL) for the synthesis of type I NCQDs/CN/BWO with an NCQDs/(BWO + CN + NCQDs) weight percentage of 1, while CN/BWO powder (1182 mg) and NCQDs powder (12 mg) were added into methanol (120 mL) for the synthesis of type II NCQDs/CN/BWO with an NCQDs/(BWO + CN + NCQDs) weight percentage of 1. The subsequent procedure for the synthesis of these samples was the same as that used for synthesizing CN/BWO. In addition, type I NCQDs/CN/BWO photocatalysts with different NCQDs/(BWO + CN + NCQDs) weight percentages were synthesized. The photocatalysts with weight percentages of 0.5, 1, 2, and 4 were referred to as NCQDs/CN/BWO-0.5, NCQDs/CN/BWO-1, NCQDs/CN/BWO-2, and NCQDs/CN/BWO-4, respectively.

The physical and chemical properties of the synthesized catalysts were inspected using X-ray diffraction (XRD), transmission electron microscopy (TEM), high-resolution transmission electron microscopy (HRTEM) equipped with energy-dispersive X-ray spectroscopy (EDS), X-ray photoelectron spectroscopy (XPS), UV-visible (UV-vis) spectroscopy, Fourier transform infrared (FTIR) spectroscopy, electron spin resonance (ESR) spectroscopy, photoluminescence (PL) emission spectroscopy, and nitrogen adsorption-desorption technique. Details for the material characterization have been described in the Supporting Material.

2.2. Photocatalytic decomposition tests

Photocatalytic decomposition tests were performed in the presence of visible light using the developed materials. The schematic of the experimental set-up is shown in Scheme S1. Ultra-pure air was directed either to a Pyrex vessel wrapped with a heating tape or to temperature-controlled water-containing impingers for humidification. Standard compounds were automatically infused into the heated Pyrex vessel and mixed with a dry air solution, after which this mixture was mixed with humidified air in an empty canister and then directed into a photocatalyst-coated Pyrex reactor equipped with a visible-light source. When observing similar concentrations of target compounds in the upstream and downstream air of the reactor, the visible-light source was activated to start photocatalytic decomposition tests. Gas samples were periodically collected in the input and output ports of the reactor. Photocatalytic decomposition tests were carried out in triplicate to afford steady results. Additionally, a representative material (i.e., NCQDs/CN/BWO-1) was consecutively employed five times for the photocatalytic decompositions of the pollutants to examine its photochemical stability. Operational conditions and control protocols for the photocatalytic decomposition tests have been provided in the Supporting Material.

Air samples were obtained using a tubular trap packed with Tenax GC powder. After treating this trap by using a pretreatment unit (Perkin Elmer ATD 350), organic vapors were analyzed by using a gas chromatography/mass spectrometry (GC/MS, Perkin Elmer Clarus SQ 8) system. For the pretreatment of traps, the desorption unit were heated at 270 °C to transfer organic vapors into a concentrating trap, which was cooled at ―28 °C. Next, the concentrated organic vapors were heated at 270 °C and transferred to the GC/MS system. The GC temperature was first programmed to 45 °C, held for 5 min, and then increased to 230 °C at 6 °C min^―1. For the quality control of organic vapor analyses, laboratory blank and spiked traps were analyzed daily. The method detection limits for nUD and mXYL were 3.6 and 5.1 ppb, respectively.

2.3. Hydroxyl radicals and photoelectrochemical properties

Details for determining the OH^· and photoelectrochemical properties have been provided in the Supporting Material.

3. Results and discussion

3.1. Material properties

Fig. 1 exhibits the TEM image, size distribution, and HRTEM image with the selected-area electron diffraction (SAED) pattern of NCQDs. NCQDs revealed a pseudo-spheroidal architecture in the TEM image, with an NCQDs size of 3.3 (average value) ±1.9 nm (standard deviation) (Fig. 1(a)); this value was commensurable with the average size range of 2.5-9.4 nm reported previously [6,19,20]. The lattice distance of NCQDs was 0.320 nm, corresponding to the (002) plane of graphitic carbon (Fig. 1(b)) [19]; consistently, the SAED pattern of NCQDs (inset) confirmed the diffraction ring of the (002) crystal plane.

The UV-vis spectra of CQDs and NCQDs suspended in deionized water and their corresponding photographs with and without UV light irradiation at 365 nm were compared to demonstrate that N was doped into CQDs. As shown in Fig. 2, NCQDs exhibited a light blue hue for the suspension subjected to UV illumination, whereas the unmodified CQDs solution was nearly transparent; under the dark, NCQDs and unmodified CQDs displayed a shiny brown hue and transparency, respectively. The color difference between NCQDs and unmodified CQDs was attributable to the presence or absence of N in CQDs, indicating that the NCQDs sample contained N. In addition, according to Fig. 2, NCQDs exhibited distinctly an enhanced light-harvesting ability relative to unmodified CQDs, suggesting their promoted optical properties. Two peaks observed at ~239 and 340 nm corresponded to the C = C and C = O transitions in NCQDs, respectively [15,20]. In addition, NCQDs subjected to visible-light illumination with different wavelengths between 500 and 700 nm exhibited upconversion PL emissions with the maximum intensities observed at 425-500 nm (Fig. S1). These results demonstrated that NCQDs incorporated into the CN/BWO heterostructure can promote the light-harvesting ability, thereby enhancing its photocatalytic capability for the degradation of hazardous vapors.

The XRD patterns of CN, BWO, CN/BWO-0.5, and NCQDs/CN/BWO-1 are illustrated in Fig. 3. Pristine CN revealed two distinct diffraction peaks at 12.8° and 27.4°, which reflect the (100) and (002) crystal planes of CN, respectively [5,21]. Pristine BWO revealed several distinct diffraction peaks at 28.3°, 32.8°, 47.1°, 56.0°, 58.5°, and 68.8°, reflecting the (131), (200), (202), (133), (262), and (400) crystal planes, respectively, of the orthorhombic phase of BWO (JCPDS 39-0256) [5,6]. BWO-based nanocomposites (i.e., CN/BWO-0.5 and NCQDs/CN/BWO-1) exhibited diffraction peaks similar to those of BWO. In addition, these materials revealed diffraction peaks indexed to pristine CN, demonstrating that CN is successfully incorporated with BWO. Unfortunately, peaks corresponding to NCQDs were not observed in the XRD pattern of NCQDs/CN/BWO-1, possibly due to the insufficient amount of NCQDs in this sample.

Fig. 4 illustrates the HRTEM images, in addition to the EDS elemental mapping, of type I and type II NCQDs/CN/BWO-1 samples. These materials revealed highly characterized crystalline properties of the three components of NCQDs, CN, and BWO, respectively. Notably, the type I sample exhibited NCQDs positioned on BWO surface and at interfaces between BWO and CN, while type II sample displayed NCQDs positioned on BWO and CN surfaces. The lattice fringes of 0.315 and 0.320 nm were assigned to the (131) facet of BWO and the (002) facet of NCQDs, respectively [19,21,22]; in addition, the HRTEM mapping results exhibited C, N, O, Bi, and W elements. Overall, the HRTEM analysis of both types of NCQDs/CN/BWO-1 indicated the successful incorporation of the three components of NCQDs, CN, and BOW, respectively.

The XPS analysis was carried out to investigate the elemental composition and chemical states on the NCQDs/CN/BWO-1 surface. In the survey spectrum of this sample, binding energy peaks corresponding to Bi, W, C, N, and O elements were observed (Fig. S2a). Two band energy peaks observed at 159.0 and 164.3 eV correspondes to Bi 4f_7/2 and Bi 4f_5/2, respectively (Fig. S2b) [5,23]. One peak observed at 35.3 eV was contributed to W 4f_7/2, while two peaks observed at 37.5 and 39.7 eV were associated with W 4f_5/2 (Fig. S2c) [24]. The XPS spectra of C 1s were deconvoluted into three binding energy peaks observed at 284.5, 286.3, 287.7, and 289.5 eV (Fig. S2d), corresponding to adventitious carbon, C—O, C=O, and sp³ C configurations, respectively [25]. In the XPS spectra of N 1s, three deconvoluted peaks at 399.1, 400.1, and 402.2 eV were associated with triazine ring, pyrrolic-N (N-(C)₃), and graphitic-N (C—N—H), respectively (Fig. S2e) [25,26]. The O 1s XPS spectra revealed two peaks at 529.9 and 532.0 eV, which were contributed to the lattice O and O—H bond, respectively (Fig. S2f) [27].

FTIR spectra were recorded to further examine the chemical structures of BWO, CN, and NCQDs/CN/BWO-1 (Fig. S3). BWO revealed two characteristic absorption signals at 568 and 1324 cm^-1, which were assigned to the stretching modes of Bi-O and W—O—W bonds, respectively [28]. In the FTIR spectrum of CN, five characteristic signals were observed at 1240, 1319, 1408, 1563, and 1637 cm^-1, indexing to the stretching modes of CN heterocycles, while two signals at 809 and 3173 cm^-1 were contributed to the vibration mode of a triazine unit and stretching mode of the O—H bond, respectively [23]. NCQDs/CN/BWO-1 exhibited FTIR peaks featured by those of both BWO and CN, indicative of the successful combination of the two components in this sample. However, peaks corresponding to NCQDs were not observed in the FTIR spectrum of NCQDs/CN/BWO-1, probably due to the low amount of NCQDs in this sample.

The optical properties of the as-synthesized photocatalysts were investigated by UV-vis spectroscopy. Compared with those of pristine BWO and CN/BWO-0.5, the absorption edge of NCQDs/CN/BWO-1 was more red-shifted, and it exhibited an enhanced light-absorption capability in the visible region (Fig. 5). In addition, the respective band gaps of BWO, CN, CN/BWO-0.5, and NCQDs/CN/BWO-1 were 2.83, 2.73, 2.80, and 2.75 eV, respectively (Fig. S4). The enhanced optical properties of NCQDs/CN/BWO-1 was attributable mainly to the UV light-harvesting property (Fig. 2) and upconversion light emission of NCQDs (Fig. S1).

3.2. Photocatalytic removal of target pollutants

Photocatalytic efficiencies for the removal of of nUD and mXYL by CN, BWO, NCQDs/BWO, NCQDs/CN, CN/BWO-0.5, and NCQDs/CN/BWO-1 under visible-light illumination were evaluated. In the absence of either visible light or a photocatalyst, negligible removal efficiencies were observed for the target pollutants. In contrast, NCQDs/CN/BWO-1 irradiated by visible light displayed the highest removal efficiency for nUD and mXYL, followed by CN/BWO-0.5, NCQDs/BWO, NCQDs/CN, CN, and BWO (Fig. 6). The removal efficiency of nUD by NCQDs/CN/BWO-1 was 57.8 (average) ± 0.8 % (standard deviation), while those observed for CN/BWO-0.5, NCQDs/BWO, NCQDs/CN, CN, and BWO were 35.6 ± 0.9 %, 29.2 ± 0.4 %, 26.1 ± 0.3 %, 17.6 ± 0.9 %, and 13.4 ± 1.2 %, respectively; additionally, the removal efficiency of mXYL by NCQDs/CN/BWO-1 was 60.6 ± 1.1 %, while those observed for CN/BWO-0.5, NCQDs/BWO, NCQDs/CN, CN, and BWO were 30.7 ± 0.5 %, 25.2 ± 0.2 %, 20.7 ± 0.4 %, 11.0 ± 0.8 %, and 6.6 ± 1.2 %, respectively. The order of photocatalysts for the pollutant removal efficiency agreed with that of photocatalysts for the photocurrent response (Fig. S5). The promoted pollutant removal efficiencies over NCQDs/CN/BWO-1 were ascribable to the Z-scheme charge transmission at the interfaces between CN and BWO in this photocatalyst, enhancing the charge-separation efficiency (Detailed discussion has been provided in Section 3.4). In addition, the visible-light absorption efficiency of NCQDs/CN/BWO-1 was greater than those of CN/BWO-0.5 and BWO (Fig. 5), likely due to the upconversion PL behavior of NCQDs, suggesting that this optical property of NCQDs is also a crucial factor in the catalytic performance. Compared to those of BWO, the higher photocatalytic efficiency of NCQDs/CN/BWO-1 was also ascribed to the higher specific surface area (SSA) (Table S1). Notably, the SSA of NCQDs/CN/BWO-1 was less than that of CN, although its performance was greater than that of the latter; a probable justification for this might be that the effect of charge-separation ability on the photocatalytic efficiency outweighed that of the SSA.

The effect of the CN/(BWO + CN) weight ratio in the CN/BWO dual nanocomposites on the pollutant removal efficiency under visible-light illumination was investigated. The pollutant removal efficiencies of the CN/BWO composites decreased in the order of: CN/BWO-0.5 > CN/BWO-0.3 > CN/BWO-0.7 > CN/BWO-0.1 (Fig. 7). Hence, with the increase in the CN/(BWO + CN) weight ratio from 0.1 to 0.5, the pollutant removal efficiency increased; however, with the further increase in the weight ratio to 0.7, the pollutant removal efficiency decreased. This pattern agreed with the photocurrent responses of the investigated samples (Fig. S6). Hence, these results indicated the existence of an optimal CN/(BWO + CN) weight ratio.

In addition, the effect of the NCQDs/(BWO + CN + NCQDs) weight percentage of the NCQDs/CN/BWO ternary nanocomposites, which were developed by the incorporation of NCQDs into CN/BWO-0.5, on the pollutant removal efficiency under visible-light illumination was evaluated. With the increase in the weight percentage from 0.5 % to 1 %, the removal efficiencies of nUD and mXYL increased from 38.3 ± 0.7%-57.8 ± 0.8 % and from 39.0 ± 0.8%-60.6% ± 1.1 %, respectively. With the increase in the weight percentage to 4 %, the removal efficiencies gradually decreased gradually to 35.3 ± 1.9 % and 38.9 ± 1.2 %, respectively, indicative of the existence of an optimal amount of NCQDs (Fig. 8). This result was in accordance with that observed for photocurrent responses (Fig. S7).

Fig. 9 shows the pollutant removal efficiencies for NCQDs/CN/BWO samples with different component arrangements under visible-light illumination. The pollutant removal efficiencies of nUD and mXYL by type I NCQDs/CN/BWO were greater than those of type II NCQDs/CN/BWO. Consistently, the photocurrent response of NCQDs/CN/BWO type I was higher than that of NCQDs/CN/BWO type II (Fig. S8). The higher pollutant removal efficiency of type I NCQDs/CN/BWO was attributed to the higher charge-separation efficiency, as supported by lower PL emission intensity compared with that of type II NCQDs/CN/BWO (Fig. S9). The difference in the charge-separation efficiencies between the two photocatalysts was ascribable to the role of NCQDs in these photocatalysts. The NCQDs in type I NCQDs/CN/BWO can act as a charge mediator at the heterojunctions between CN and BWO, thereby promoting the mobility of charges at the interfaces. In addition, they can act as photosensitizers on the CN or BWO surface, enhancing the visible-light harvesting of type I NCQDs/CN/BWO. In contrast, NCQDs in type II NCQDs/CN/BWO can only act as photosensitizers.

3.3. Photochemical stability

Photochemical stability is an essential factor for estimating the practical applications for environmental remediation [29,30]. To examine the stability of the freshly developed photocatalyst, NCQDs/CN/BWO-1 was consecutively used in five runs for the decomposition of nUD and mXYL under visible-light illumination. The photocatalytic performance did not distinctively differ during the successive runs, affording >98 % of its initial efficiency after the final run (Fig. 10). The slight difference in the pollutant decomposition efficiency during the consecutive runs was attributed to the ineludible systematic errors. The photocurrent responses of NCQDs/CN/BWO-1 measured prior to and after the consecutive test did not exhibit any apparent difference (Fig. S10). Moreover, the XRD spectra of fresh and used NCQDs/CN/BWO-1 did not exhibit any distinct difference in their crystal phases prior to and after the consecutive test (Fig. S11). These results demonstrated that NCQDs/CN/BWO-1 possesses good photochemical stability, corresponding to the promoted charge-separation capability due to the role of NCQDs as a charge mediator as well as the Z-scheme process proposed in the Section 3.4. As such, the constructed photocatalyst can be a highly contending candidate for the control of hazardous vapors with outstanding photochemical stability.

3.4. Photocatalytic mechanism

Fig. 11 presents the potential mechanisms of the two types of NCQDs/CN/BWO photocatalysts for the decomposition of hazardous vapors under visible-light illumination. The band structures of CN and BWO in type I NCQDs/CN/BWO were established on the basis of the converted Tauc plots (Fig. S4) and valence-band (VB) XPS spectra (Fig. S12). The band gaps of CN and BWO were 2.73 and 2.83 eV, respectively, while their corresponding VB values were +1.73 and +3.07 eV. To examine the charge-transfer pathway of type I NCQDs/CN/BWO, hydroxyl radical ($OH^{\bullet}$) populations generated via CN, BWO, and NCQDs/CN/BWO-1 were determined on the basis of the estimated fluorescence emission strengths of 2-hydroxyterephthalic acid (Fig. S13). With respect to all samples, the $OH^{\bullet}$ population generated during the photocatalytic process increased with the illumination time. Additionally, the $OH^{\bullet}$ population of NCQDs/CN/BWO-1 was greater than those of CN and BWO over the entire course of the illumination time. Presuming that the movement of charges in NCQDs/CN/BWO-1 follows a heterojunction-type II pathway, positive holes (h⁺) would be transferred from the VB of BWO to the VB of CN due to the higher energy level of the VB of BWO (+3.12 eV) than that of CN (+1.73 eV). In this case, h⁺ in the VB of CN could not form $OH^{\bullet}$ because of more positive energy level of the OH^-/$OH^{\bullet}$ reaction potential (+1.90 eV) than the CN VB value [31]. Accordingly, the greater $OH^{\bullet}$ population on the NCQDs/CN/BWO-1 surface revealed that the movement of charges in this photocatalyst does not follow the heterojunction-type II pathway.

Instead, a Z-scheme charge pathway was proposed for the removal of vaporous nUD and mXYL over the type I NCQDs/CN/BWO photocatalyst under visible-light illumination (Fig. 11(a)). In this case, the high $OH^{\bullet}$ population on the NCQDs/CN/BWO-1 surface was attributable to the role of h⁺ present in the VB of BWO, oxidizing water vapor to form $OH^{\bullet}$. Upon illuminating, CN and BWO are excited to generate holes and electrons in their VB and conduction band (CB), respectively. The retained holes in the VB of BWO are consumed to oxidize hydroxyl ion (OH^-), generating $OH^{\bullet}$ owing to the more positive VB edge (+3.07 eV) of BWO than the OH^-/$OH^{\bullet}$ reaction potential. Almost synchronously, the electrons in the CB of BWO would transfer to NCQDs, which serve as a charge mediator, due to the high electron affinity of NCQDs, and then transfer to VB of CN to couple with holes. The electrons in the CB of CN react with O₂ to afford $O_{2}^{\bullet -}$O₂. In addition, NCQDs enhanced the visible-light harvesting ability of NCQDs/CN/BWO-1 due to their upconversion fluorescence (Fig. S1), highlighting the multiple features of NCQDs in NCQDs/CN/BWO-1. Thereupon, the active $OH^{\bullet}$ and $O_{2}^{\bullet -}$ oxidize nUD and mXYL to generate innoxious CO₂ and H₂O, and certain intermediates, thus promoting the photocatalytic ability of Z-scheme NCQDs/CN/BWO-1. In addition, Fig. S14 exhibits the ESR spin-trappings of CN, BWO, and NCQDs/CN/BWO-1 for 5,5-dimethyl-1-pyrroline N-oxide (DMPO)- $O_{2}^{\bullet -}$ and DMPO-$OH^{\bullet}$. The strong $O_{2}^{\bullet -}$ peaks for CN and NCQDs/CN/BWO-1 confirmed the role of photoinduced electrons in the photocatalytic mechanism over these catalysts, while the strong $OH^{\bullet}$ peaks for BWO and NCQDs/CN/BWO-1 confirmed the role positive holes in the photocatalytic mechanism over these samples; consequently, these findings were assigned to the Z-scheme mechanism in the removal of vaporous nUD and mXYL over the type I NCQDs/CN/BWO photocatalyst under visible-light illumination. Similar to the type I NCQDs/CN/BWO, type II NCQDs/CN/BWO also followed the Z-scheme charge-transfer mechanism (Figu. 11(b)). However, NCQDs in type II NCQDs/CN/BWO exhibited only upconversion fluorescence, not serving as a charge mediator because of their absence at interfaces between CN and BWO.

4. Conclusion

In this study, the modification of a CN/BWO dual nanocomposite with NCQDs is confirmed to boost the photocatalytic ability for the treatment of hazardous vapors under visible-light illumination. Another major finding is that the arrangement of components in multi-composite photocatalysts affects their photocatalytic efficiencies, due to different charge-transfer routes. Specifically, the type I NCQDs/CN/BWO photocatalyst with dual properties of NCQDs (i.e., charge mediation and upconversion fluorescence, respectively) revealed higher removal efficiencies for nUD and mXYL than those of the selected counterpart photocatalysts. Moreover, this photocatalyst was superior to the type II NCQDs/CN/BWO photocatalyst with only the upconversion fluorescence of NCQDs for the removal of pollutants. In addition, the type I NCQDs/CN/BWO photocatalyst exhibited prominent photochemical stability, suggesting that it is a favorable long-term photocatalyst for environmental decontamination. Hence, the enhanced photocatalytic activity for the ternary nanocomposite photocatalysts of NCQDs-modified CN/BWO for the removal of model compounds can advance the development of upcoming decontamination techniques of hazardous vapors, and the component arrangement is a crucial factor for the development of multi-composite catalysts for efficient performance.

Acknowledgements

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIP) (Nos. 2016R1A2B4009122 and 2017R1A4A1015628).

Appendix A. Supplementary data

Supplementary material related to this article can be found, inthe online version, at doi: https://doi.org/10.1016/j.jmst.2019.09.026.

---