High-Performance 3D Semiconductor-Based Heterostructure Photocatalyst in Sustained Control of Harmful Gas-Phase Pollutants Under Visible- Light Illumination


 Herein, a highly efficient three-dimensional (3D) semiconductor-based heterostructure photocatalyst (i.e., WO3–g-C3N4 monolithic architecture; WOCNM) was developed by immobilizing a WO3–g-C3N4 heterostructure powder on a melamine foam (MF) framework. Subsequently, the sustained control of two harmful model gas-phase pollutants (i.e., n-butanol and o-xylene) over WOCNM and selected monolithic counterparts (i.e., MF-supported WO3 monolith and MF-supported g-C3N4 monolith) was investigated under visible-light irradiation. WOCNM exhibited higher photocatalytic capabilities in the sustained control of the two model pollutants than those of individual WO3 and g-C3N4 monoliths because the WO3–g-C3N4 heterojunction enhanced its charge-separation ability. Notably, WOCNM exhibited highly efficient photocatalytic capabilities in the sustained control of n-butanol (up to 97%) and o-xylene (up to 86%). Moreover, no noticeable changes were observed in the WOCNM photocatalytic capability after the final run of successive applications. The fresh and successively used WOCNMs were nearly identical, and the photocatalyst powder was not observed in the reaction chamber after its successive application. As a result, WOCNM was a highly efficient and stable 3D heterostructure photocatalyst for the sustained control of gas-phase n-butanol and o-xylene, without significant catalyst powder loss. Promisingly, this study will expedite the future development of 3D photocatalysts for the sustained control of harmful gas-phase pollutants.


Introduction
Air quality is crucial because it is directly related to human health and well-being [1][2][3][4][5]. Particularly, the inhalation exposure to n-butanol and o-xylene, which are prevalently present in various ambient environments, has been linked to a range of noxious health outcomes, including symptoms associated with carcinogenic diseases [6][7][8]. Accordingly, techniques need to be developed to control the aforementioned gas-phase pollutants to protect the human health and well-being.
The oxidative photocatalytic process using semiconductor catalysts has been attractively used to control a range of gas-phase pollutants, because it predominantly converts gas-phase pollutants to innoxious products under normal pressure and temperature environments [9][10][11][12][13][14][15]. Graphic carbon nitride (g-C 3 N 4 ), which is a popular visible-light-active semiconductor photocatalyst, exhibits numerous attractive advantages in the oxidative photocatalytic processes, such as outstanding chemical resistance, eminent electronic properties, and a exible two-dimensional morphology. However, g-C 3 N 4 affords a low separation of photoinduced carriers, which induces a low quantum yield and unsatisfactory photocatalytic activity [16][17][18]. This disadvantage can be overcome by hybridizing the aforementioned semiconductor with other semiconductors [19][20][21][22][23][24]. Tungsten trioxide (WO 3 ) is a prospect semiconductor that can be appropriately hybridized with g-C 3 N 4 to yield a highly e cient WO 3 -g-C 3 N 4 heterostructure for the control of harmful gas-phase pollutants because of its greatly positive valence band (VB) position and harmonic electronic structure, which affords a great oxidation potential [26][27][28].
It is further noteworthy that three-dimensional (3D) architecture photocatalysts established by immobilizing them onto a supporting substance is highly suitable for the sustained photocatalytic control of gas-phase pollutants because of the minimal loss of catalyst powders, which are easily blown away with the puri ed air. Therefore, herein, a WO 3 -g-C 3 N 4 heterostructure was immobilized on a melamine foam (MF) framework to yield a highly e cient 3D WO 3 -g-C 3 N 4 monolithic system (WOCNM) for the sustained photocatalytic control of harmful gas-phase n-butanol (nBT) and o-xylene (oXYL) under visiblelight illumination. Additionally, two 3D counterparts (i.e., MF-supported WO 3 monolith (WOM) and MFsupported g-C 3 N 4 monolith (CNM)) were evaluated toward the sustained photocatalytic control of the model gas-phase pollutants. Photocatalyst stability is another crucial engineering aspect for its practical applicability to pollutant control. To examine its stability, WOCNM was successively utilized for the photocatalytic control of the model pollutants. Thereafter, a systematic photocatalytic process over WOCNM was presented based on the electronic structures and active species tests.

Photocatalyst fabrication
To fabricate the WO 3 powder, Na 2 WO 4 ·2H 2 O (3.0 g) and citric acid (1.8 g) were delivered to pure water (120 mL), and the solution was stirred for 15 min. Subsequently, HCl (6 M) was dropped slowly into the above solution until its pH reached 1, accompanied by the formation of a yellow precipitate. After stirring for 0.5 h, the resultant solution was placed in an autoclave (150 mL), which was thermally treated at 150°C for 20 h. The hydrothermally treated suspension was cooled and subjected to centrifugation. The precipitate was cleaned with pure water and conditioned at 80°C for 15 h to afford a yellow WO 3 sample.
Furthermore, a bulk g-C 3 N 4 powder sample was fabricated by grinding melamine (5 g) and then calcining at 500°C for 5 h; subsequently, a sheet g-C 3 N 4 photocatalyst was obtained aftercalcining the asfabricated bulk g-C 3 N 4 powder at 550°C for 1.5 h.
The WO 3 -g-C 3 N 4 heterostructure was fabricated via the electrostatic assembly of the as-fabricated WO 3 and g-C 3 N 4 powders. Brie y, lactic acid (20 mL) was delivered to deionized water (80 mL); then, this solution was stirred for 0.5 h. The g-C 3 N 4 powder (0.4 g) was delivered to the solution. The dispersion was sonicated for 2 h and stirred for 30 min to afford a homogeneous solution (Solution A). Additionally, WO 3 powder (1.6 g) was delivered to deionized water (10 mL); the dispersion was stirred for 0.5 h to yield Solution B; this solution was dropped slowly into Solution A. Subsequently, the mixture was stirred for 2 h. The mixture was centrifuged, and the resultant solid was cleaned with pure water and dried at 80°C for 15 h to yield a WO 3 -g-C 3 N 4 powder.
The 3D WOCNM samples were constructed using a facile immobilization process. In summary, the assynthesized WO 3 -g-C 3 N 4 powder (0.7 g) was delivered to deionized water (100 mL), and the mixture was sonicated for 10 min and then stirred for 0.5 h. An MF piece (5 mg) was then dipped into the solution for 10 min. Finally, the solution-absorbed MF was dried at 85°C for 10 h to provide the 3D WOCNM sample.
Hereafter, the 3D WOCNM heterostructures fabricated using 0.1, 0.3, 0.6, and 1.6 g of WO 3 powder are denoted as WOCNM-0.1, WOCNM-0.3, WOCNM-0.6, and WOCNM-1.6, respectively. Moreover, the 3D WOM and 3D CNM samples were manufactured following the same method used to synthesize the 3D WOCNM but without the addition of g-C 3 N 4 and WO 3 powders, respectively. The sample characterizations are discussed in the Supplementary Material. Figure S1 illustrates the photocatalytic system employed to assess the catalyst performance toward the sustained control of gas-phase nBT and oXYL under visible light. Ultra-puri ed air (99.9999%) furnished by a pressurized-air cylinder was separated into two air-stream lines: one was guided to an evaporation unit, where the liquid-phase target chemicals were simultaneously introduced and evaporated for synthesizing the standard gas, and the other was delivered to a humidi cation unit. The two air-stream lines were merged and mixed in a heated empty chamber. The mixed gas was then transferred to a steady-stream stainless steel (SS) vessel (53 mL volume), where the test monolith was placed. Simulated sunlight was supplied using a xenon lamp (150 W) through a quartz window. The major photocatalytic conditions that control the model pollutants include the amount of each catalyst system (40 mg), light strength (9.5 mW cm −2 ), feed concentration of the model pollutants (1 ppm), and airstream owrate (60 mL min −1 ). To investigate the effect of the model pollutants' feed concentration on the photocatalytic capability of a representative photocatalyst (i.e., WOCNM-0.6), three other feed concentrations (i.e., 0.1, 0.5, and 2 ppm) were also investigated.

Sustained photocatalytic control of gas-phase pollutants
The preliminary control protocols comprised system puri cation, humidity equilibrium, quasi-adsorption equilibrium, and photolytic processes. The experimental system was puri ed by allowing ultra-pure air to ow through the system overnight under illumination. The humidity and quasi-adsorption equilibria were identi ed on the basis of the similarities in relative humidity and pollutant concentrations, respectively, between the input and output airstreams of the reaction vessel. The photolytic process of the model compounds was assessed via a vacant SS chamber upon irradiation. After completing the control protocols, sustained photocatalytic pollutant control tests were initiated by allowing the standard gas to ow through the reaction vessel and activating the xenon lamp. The photochemical stability of WOCNM-0.6 was assessed by continuously running it ve times for the control of the model pollutants.

Measurement of the model gas-phase pollutants
The model gas-phase pollutants were periodically measured at the inlet and outlet portions of the reaction vessel using an adsorbent tube [29]. The model pollutants (nBT and oXYL) adsorbed on the sampling tube were determined using a pretreatment unit/Perkin Elmer gas chromatography/mass spectrometry (GC/MS) system. Speci cally, the sampling adsorbent tube was heated at 270°C for 8 min and cooled to -28°C. The concentrated chemicals were then heated at 270°C and directed to the GC/MS unit.
To control the quality of the chemical determinations, blank and standard samples were daily investigated. Furthermore, the daily variation of the GC/MS signals were identi ed and controlled. The measurement threshold values determined by investigating the standard samples were 3.4 and 3.9 for nBT and oXYL, respectively.   [31]. The XRD pattern of CNM revealed a weak peak at 12.7° and a prominent peak at 27.5°, corresponding to the (100) and (002) phases of g-C 3 N 4 , respectively [21,22]. Additionally, the XRD pattern of WOCNM-0.6 revealed major re ection peaks that were associated with both WO 3 and g-C 3 N 4 , indicating the successful hybridization of the two components.
The surface properties of pure MF, WOM, CNM, and WOCNM-0.6 were additionally surveyed using FTIR ( Figure 5). The pure MF spectrum revealed a broad absorption peak at 3325 cm −1 and a distinct peak at 806 cm −1 , which were ascribed to the O-H bond and triazine ring, respectively [33]. The WOM spectrum displayed a broad absorption peak positioned at 450-900 cm −1 , which matched the vibration form of the W-O-W group [25,26]. The CNM spectrum displayed several peaks at 1237-1627 cm −1 , which were indexed to the C-N bond of g-C 3 N 4 [17]. Furthermore, the WOCNM-0.6 spectrum exhibited the characteristic absorption peaks of both WO 3 and g-C 3 N 4 , verifying the successful hybridization of the two components in this sample. reported for g-C 3 N 4 powder in previous studies [29,34]; and the WOM sample revealed high visible-light absorption capability with a bandgap of 2.45 eV, which is comparable to that reported for WO 3 powders in previous studies [24,26,35]. Moreover, the bandgaps of the four WOCNM samples (i.e., WOCNM-0.1, WOCNM-0.3, WOCNM-0.6, and WOCNM-1.6), which were determined to be intermediates between those of WOM and CNM, increased from 2.48 to 2.65 eV with increasing WO 3 loading amount.

Photocatalytic capabilities of 3D samples
The photocatalytic capabilities of the three 3D samples (i.e., WOM, CNM, and WOCNM) in the sustained control of gas-phase nBT and oXYL were assessed using a steady-stream photocatalytic system under simulated sunlight irradiation. The preliminary experiments performed in the absence of either a catalyst or light revealed no noticeable control of the model pollutants. Contrarily, the illuminated photocatalysts exhibited distinctive sustained removal of the model pollutants. As shown in Figure 7, WOCNM revealed higher photocatalytic capabilities in the sustained control of nBT and oXYL than those of the individual WO 3 and g-C 3 N 4 monoliths. The average sustained photocatalytic control capabilities of nBT and oXYL over WOCNM-0.6 were 81.7% ± 0.6% and 62.7% ± 0.5%, respectively, over WOM were 19.4% ± 0.4% and 8.1% ± 0.3%, respectively, and over CNM were 23.9% ± 0.7% and 18.2% ± 0.2%, respectively. The enhanced photocatalytic capabilities of WOCNM-0.6 relative to those of WOM and CNM are ascribable to the enhancement of its charge-separation ability due to the WO 3 -g-C 3 N 4 heterostructure. This assertion was con rmed by the lower PL peak intensity of WOCNM-0.6 relative to those of WOM and CNM (Figure S3), because a low PL peak intensity denotes a low charge recombination rate [36,37].
Notably, WOCNM exhibited highly e cient photocatalytic capabilities in the sustained control of nBT (up to 97%) and oXYL (up to 86%) when 0.1 ppm of the model pollutants' feed concentration was utilized ( Figure 8). When their feed concentrations were decreased from 2 to 0.1 ppm, the sustained photocatalytic control capabilities of nBT and oXYL gradually increased from 47-97% and from 33-86%, respectively. This trend is similar to that observed in previous studies wherein the photocatalytic control capabilities of other gas-phase pollutants were investigated [29,38]. The feed concentration dependency is ascribed to the low competition between pollutant molecules for reactive spaces on the catalyst surface at low concentrations.
The effect of the amounts of WO 3 powder loaded into g-C 3 N 4 in WOCNM on its photocatalytic capability toward controlling nBT and oXYL was also evaluated. The photocatalytic capabilities of WOCNM in the sustained control of the model pollutants were in the following order: WOCNM-0.1 < WOCNM-0.3 < WOCNM-1.6 < WOCNM-0.6 ( Figure 7). This suggests that an optimal WO 3 amount exists for the WOCNM synthesis. This pattern of the WOCNM catalysts in the sustained photocatalytic control of the model pollutants is ascribable to their charge-separation e ciencies, which were veri ed by their PL peak intensities ( Figure S3).
Photochemical stability analysis of WOCNM-0.6 was performed by continuously running it ve times under simulated sunlight irradiation to assess its practical applicability in gas-phase pollutant control. Figure 9 displays that no noticeable changes were observed in the photocatalytic ability of the material for the target compounds control after the nal run; the material retained >98% of its original control capability. The photochemical stability of WOCNM-0.6 was also demonstrated via XRD analyses ( Figure  10): the XRD patterns of pristine and spent samples exhibited considerable differences. The photochemical stability of WOCNM-0.6 is ascribable to its promoted charge-transfer ability at the WO 3 -g-C 3 N 4 interfaces, which is explained in detail in section 3.4.

Optical characteristics of MF and MF-supported 3D photocatalysts
Pure MF, WOM, CNM, and WOCNM-0.6 displayed white, dark-yellow, light-yellow, and yellow colors, respectively ( Figure S4). According to the optical examination, the powder samples were well xed on the MF framework in the three MF-supported 3D photocatalysts (i.e., WOM, CNM, and WOCNM-0.6); the fresh and successively used WOCNM-0.6 were nearly identical. Additionally, the photocatalyst powder was not observed in the reaction chamber after continuous applications of WOCNM-0.6. Consequently, no noticeable catalyst powder loss occurred during the photocatalytic application of the MF-supported materials. The results indicate that when MF-supported 3D architecture photocatalysts are used in the sustained photocatalytic control of gas-phase pollutants, no signi cant catalyst powder loss occurs.

Systematic photocatalytic process
A systematic photocatalytic process for controlling gas-phase pollutants over a WOCNM system was presented based on the electron diagrams of WO 3 and g-C 3 N 4 and their redox potentials ( Figure 11).
According to the Tauc plots ( Figure S2) and XPS data ( Figure S5), the bandgaps of WO 3 and g-C 3 N 4 equaled 2.45 and 2.72 eV, respectively, their VB energies equaled +2.63 and +1.52 eV, respectively. The charge-transfer pathways in WOCNM were determined by analyzing the hydroxyl radical (OH • ) populations over WOM, CNM, and WOCNM-0.6 ( Figure S6). In all the samples, the OH • population afforded during the photocatalytic process increased with the illumination time. Particularly, WOCNM-0.6 afforded the greatest OH • production over the illumination period. The greater OH • production over WOCNM-0.6 than that over WOM is due to the lower hole-electron recombination rate in the former. If the charge-transfer pathway in WOCNM-0.6 corresponds to the heterojunction type II process, then the photoinduced holes in the WO 3 VB (+2.63 eV) might move to the g-C 3 N 4 VB (+1.52 eV) because of the lower VB potential of the latter. The transferred holes in g-C 3 N 4 VB could not yield OH • because of their lower potential than the OH − /OH • conversion energy (+1.90 eV) [39].
Therefore, the greater OH • production over WOCNM-0.6 than that over WOM is attributable to the Zscheme charge-transfer process. Thus, upon illumination of WOCNM-0.6, the photoinduced holes in the WO 4 VB oxidize OH − to OH • because of its greater positive potential (+2.63 eV) than that of OH − /OH • .
Simultaneously, electrons in the WO 3  The Z-scheme charge-transfer process over WOCNM was further proven via active species experiments.
Therefore, the active species experiment con rms that the photocatalytic progress over WOCNM systems occurs on the basis of the Z-scheme charge-transfer process.

Conclusions
We investigated the photocatalytic capabilities of WOCNM, which was developed by immobilizing a WO 3 -g-C 3 N 4 heterostructure into an MF framework, and two monolithic counterparts (i.e., WOM and CNM) in the sustained control of nBT and oXYL using a steady-stream photocatalytic system under visible light. The WOCNM sample displayed highly e cient photocatalytic performance and outstanding photochemical stability in the sustained control of nBT and oXYL. The enhanced photocatalytic capabilities of WOCNM relative to those of WOM and CNM were ascribable to the enhancement of its charge-separation ability due to the WO 3 -g-C 3 N 4 heterostructure. This assertion was con rmed by the lower PL peak intensity of WOCNM relative to those of WOM and CNM, because a low PL peak intensity denotes a low charge recombination rate. In addition, no signi cant catalyst powder loss was observed for the WOCNM sample successively used in the steady-stream photocatalytic system. Positively, WOCNM was a highly e cient 3D heterostructure photocatalyst for the sustained control of gas-phase nBT and oXYL, without signi cant catalyst powder loss.

DECLARATION OF COMPETING INTEREST
The authors declare that no con icting actual or potential interests exist.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. SupplementaryMaterial.docx