Introduction

Organic pollutants (such as dyes1,2,3) and antibiotics [(including aminoglycoside, β-lactam, glycopeptide, and tetracycline (TC)]4 in wastewater have brought great harm to the environment and human health.5 Therefore, the exploration of environmental remediation strategies has attracted more and more attention. Pollutant concentrations can be degraded through conventional adsorption6 and coagulation7 techniques in a short period of time, but pollutants can be enriched rather than eliminated. Meanwhile, chemical oxidation8 requires high costs, while biological degradation9 takes a considerable amount of time, which is still not satisfactory. Compared with all the other methods, due to the mineralization of organics by highly effective active species generated during semiconductor photocatalysis, this has become the most cost-effective and user-friendly alternative for pollutant removal in waste water.10

The selection of semiconductor photocatalysts plays a major role in photocatalysis. Traditional inorganic semiconductor catalysts, such as TiO2 , cannot efficiently utilize the most visible part of solar energy owing to its wide bandwidth (Eg = 3.2 eV).11 Therefore, the search for visible light photocatalysts with stable competence is still the investigation core in the field of photocatalysis. In the past few years, several researchers have explored photocatalysts with appropriate band gaps for quite a number of environmental purposes, like wastewater treatment (such as dyes,12 antibiotics,13 etc.), photocatalytic hydrogen manufacture,14 and photocatalytic conversion of carbon dioxide.15 Among these, graphitized carbon nitride (g-C3N4) is a typical metal-free semiconductor polymer composed of C and N elements, which is easy to prepare, is non-toxic, has high chemical stability and thermal stability, and appropriately accurate metal loading capacity.16,17 Wang's research group in 2009 first used g-C3N4 as a visible light photocatalyst, which can being about the photocatalytic decomposition of aquatic hydrogen and oxygen production, realizing the expansion from a traditional inorganic semiconductor to a polymer semiconductor photocatalyst.18 However, common blocky g-C3N4 presents deficiencies, such as limited specific surface area, photogenerated rapid carrier recombination, insufficient absorption of visible light, and low quantum efficiency, leading to defects in its photocatalytic performance.19 In order to avoid these limitations, various modification strategies have been adopted. In recent years, the construction of heterojunction composites, such as Z-type heterojunction, has attracted the great attention of researchers.20 There is evidence that various semiconductors can form heterostructures with g-C3N4, such as TiO2,21,22 V2O5,23,24 Bi-based materials,25,26,27,28 and quantum dots,29 which are effective in prohibiting the recombination of photo-induced electrons and holes.

As a typical n-type semiconductor, tungsten trioxide (WO3) has received more and more attention due to its narrow band gap (2.6–2.8 eV), which can be a good candidate capable of capturing visible light.30 To increase its catalytic performance, pure WO3 must be combined with various other semiconductors to enhance its visible-light photocatalytic efficiency through precisely controlled in situ hydrolysis and a successive polymerization process.31 For instance, Xiao et al. developed WO3/g-C3N4 composite hollow microspheres, showing enhanced degradation rates against ceftiofur sodium and tetracycline hydrochloride under visible light.32 Pan et al. synthesized a 2D-2D WO3/g-C3N4 S-scheme heterojunction by a straight forward calcination method, and it displayed more significant photocatalytic performance in TC degradation than WO3 and g-C3N4 under visible-light irradiation.33 To summarize, based on the advantages of electronic transfer and light absorption between the heterojunctions, the WO3/g-C3N4 composite catalyst can exhibit excellent photocatalytic activity. The nanotubes present a significant advantage in terms of enhancing charge carrier mobility and surface area, but the synthesis process is too time-consuming for widespread production.34

In this study, we firstly synthesized a porous WO3/g-C3N4 nanotube heterojunction system by a simple one-pot calcination method, and confirmed its unique porous tubular microstructure by field-emission scanning electron microscope (FE-SEM) and transmission electron microscope (TEM) analysis. The degradation of Rhodamine B (RhB) and TC under visible light and simulated sunlight, respectively, were investigated. Finally, the photocatalytic degradation mechanism of the WO3/g-C3N4 composite is proposed, based on the free radical capture experiment.

Experimental

Materials

Urea (CH4N2O; Macklin, 99%), melamine (C3H6N6; Macklin, 99%), ammonium metatungstate hydrate ((NH4)6H2W12O40·xH2O; Macklin, 99.5%) were used without purification. Deionized (DI) water was used in the whole experiment.

Preparation of g-C3N4, WO3 and WO3/g-C3N4 Nanotubes

In a typical synthesis process, the nanotubular g-C3N4(CNNT) is achieved by placing 4 g urea and 0.4 g melamine in a quartz vessel and heating it at 5°C/min to 550°C for 5 h. Similarly, added melamine was calcined to generate bulk g-C3N4 (BCN) for comparison. In the same way, the 2% WO3/CNNT was synthesized by adding 0.0086 g ammonium metatungstate into the raw materials. Then, 1%, 3%, 5%, and 7% WO3 -loaded CNNT were obtained by varying the mass of WO3 under a parallel method (Scheme 1).

Scheme 1
scheme 1

One-pot calcination of tubular WO3/g-C3N4.

Full size image

Characterization

The X-ray diffraction (XRD) patterns of the materials were observed (D8-Advance; Bruker) with Cu Kα radiation (λ = 0.154056 nm). Fourier-transform infrared (FT-IR) spectra in the wave-number region of 400–4000 cm−1were obtained (IRAffinity-1S spectrometer; Shimadzu). The produced samples were examined with a field-emission scanning electron microscope (FE-SEM; SU8010; Hitachi) fitted with an energy dispersive x-ray spectrometer (EDS). Transmission electron microscope (TEM) images were obtained (Model 800; Hitachi), and a high-resolution transmission electron microscope (HRTEM; JEM-1230; JEOL). X-ray photoelectron spectroscopy (XPS) spectra were acquired (ESCALAB 250) using monochromatized Al Kα radiation (hv = 1486.7 eV). UV–vis diffuse reflectance spectroscopy (UV–vis DRS) spectra of the samples were acquired (V-650; JASCO, Japan). UV adsorption was tested (UV-1800; MAPADA). To obtain the BET specific surface areas and pore diameters of the synthetic materials, N2 adsorption and desorption isotherm tests were conducted (Autosorb-iq; Quantachrome Instruments). To better understand the electrochemical properties, a CHI760E electrochemical system (Shanghai, China) was used.

Electrochemical Performance Measurement

The electrochemical performance was evaluated using a three-electrode system with 0.5 M Na2SO4 solution as the electrolyte. The catalyst was coated on 1 cm2 fluoride-doped tin oxide-coated glass, with half the area immersed in the electrolyte to test its photocurrent transient response. To determine where the conduction and valence bands were located, Mott–Schottky curves were measured at 1 kHz. Furthermore, electrochemical impedance spectra (EIS) measurements were used to determine the charge transfer resistance. The counter electrode (cathode) was composed of a Pt wire, and the reference electrode and the working electrode (photo-anode) were made of Ag/AgCl (saturated KCl).

Photocatalytic Performance Measurement

By degrading Rhodamine B (RhB) and TC in aqueous solutions, the photocatalytic activity of different catalysts was investigated. A RhB solution was prepared with 10 mL 1 × 10−4 mol/L of the original solution and 40 mL DI water. The visible-light source was provided using a 300-W Xenon light (Hxuv300; Beijing CEJ Tech., Beijing, China) and an ultraviolet filter (λ ≥ 420 nm). Prior to the photocatalytic reaction, 10 mg of the catalyst was added to the prepared RhB solution for shading and stirring for 60 min to achieve the adsorption–degradation equilibrium. After that, under continuous visible-light irradiation for 15 min, 5 mL of the suspension was collected and centrifuged at 4000 rpm for 3 min to filter the particles at intervals of 5 min. The photocatalytic property of the WO3/CNNT composites was also characterized using reactive free radical trapping experiments. Capture agents, ammonium oxalate (AO, 1 mmol), 1,4-benzoquinone (BQ, 0.1 mmol), and dimethyl sulfoxide (DMSO, 1 mmol), 2-propanol (IPA, 1 mmol), were detected as hydroxyl radical (·OH), hole (h+), superoxide radical (·O2−), and electron (e−) scavengers, respectively.

The photocatalyst (20 mg) was dispersed in a solution of 50 mL TC at a concentration of 40 mg/L. Adsorption equilibrium was reached by magnetized stirring of the suspension for 1 h in darkness, which was then exposed to simulated solar light (without a filter). A 5 mL suspension was extracted every 20 min, and the granular photocatalyst was derived from aqueous solution using a 20,000-rpm centrifuge. With a UV–vis spectrophotometer, the RhB concentrations and the TC concentrations were respectively measured during the photodegradation process by recording the variations in the absorption bands at 554 nm and 357 nm.

Results and Discussion

XRD Analysis

The XRD patterns of tubular g-C3N4 are similar to bulk g-C3N4 with two characteristic peaks at 2θ = 13.5° and 27.8° (shown in Fig. 1a) as the (100) and (002) crystal facets (JCPDS No. 87–1526). The weaker diffraction peak at 2θ = 13.5° corresponded to the in-plane structures of tri-s-triazine units, whereas a more prominent peak at 2θ = 27.8° is associated with the interlayer packing.35 However, the peak strength of CNNT is lower than that of BCN, indicating that the nanomaterial has obvious size dependence, and that tri-s-triazine building blocks have a shorter correlation length of interlayer periodicity.36 There is no discernible peak of WO3 in the composite materials, perhaps because of its low load. By magnifying the XRD figure 2θ = 25°–30°, shown in Fig. 1b, it is clear that, when the WO3 loading increases, the reduction of the (002) crystal facets may be caused by the insertion of loaded WO3 into the lattice of CNNT.

Fig. 1
figure 1

(a) XRD patterns of pure WO3, BCN and CNNT, and of the WO3/CNNT composite samples, (b) magnified XRD patterns of the WO3/CNNT samples.

Full size image

FT-IR Analysis

Figure 2 compares the FT-IR patterns of WO3, CNNT, and 2% WO3/CNNT composites at room temperature. The broad peak located at 450–900 cm−1 is attributed to the stretching vibration of W-O-W,37 note the formation of a g-C3N4 phase in the CNNT and 2% WO3/CNNT. The distinctive peak at 810 cm−1 is from the characteristic breathing vibration of the tri-s-triazine units.38 Meanwhile, typical stretching modes of C-N and C=N groups were in the 1200–1700 cm−1 region. In addition, the peak at 3000–3400 cm−1 can be ascribed to the stretching vibration of -OH.39,40

Fig. 2
figure 2

FT-IR spectra of WO3, CNNT and 2% WO3/CNNT.

Full size image

Morphology

The microstructures of WO3, BCN, CNNT, and 2% WO3/CNNT were revealed by FE-SEM. According to Fig. 3a, the diameter of pure WO3 nanoparticles is between 100 and 200 nm. It can be noted from Fig. 3b that the morphology of BCN is a block structure formed by a nanosheet assembly. As shown in Fig. 3c, the CNNT is a tubular structure with a diameter of about 300 nm. As seen from Fig. 3d, the 2% WO3/CNNT exhibited a porous nanotube structure. During calcination, the mass of NH3 and CO2 were released, resulting in the characteristic porous structure. Figure 3e, f show the EDS spectra of 2% WO3/CNNT. The existence of the Cu element is derived from the use of a Cu sheet as a carrier, clearly demonstrating that there are C, N, W, and O elements in the sample. The presence of impurities can act as the capture site of electrons–holes and lead to reducing the mobility of the charge.41 Figure 3f further ratifies the absence of impurities in the sample. The EDS mapping images of the 2% WO3/CNNT indicate the elemental composition of C, N, W, and O with homogeneous distribution in the composite, as shown in Fig. 3g–k. As can been seen from the TEM image in Fig. 4a, some WO3 nanoparticles are thoroughly bound to the surface of the CNNT. Using HRTEM, the lattice fringes and interface between both components were observed. The HRTEM image of the composites in Fig. 4b reveals a close interface between the two components, supporting the heterojunction formation. The lattice spacing was about 0.202 nm, attributed to the (123) plane of WO3. According to the above results, it can be easily concluded that the WO3 and CNNT are compatible with a heterojunction through robust interface interactions.

Fig. 3
figure 3

FE-SEM micrographs of the samples: (a) WO3; (b) BCN; (c) CNNT; (d) 2%WO3/CNNT; (e) FE-SEM image and (f) EDS spectrum of 2%WO3/CNNT; (g–k) FE-SEM image and the elemental mapping images of 2%WO3/CNNT.

Full size image
Fig. 4
figure 4

(a) TEM and (b) HRTEM photo of 2% WO3/CNNT; inset HRTEM.

Full size image

XPS Surveys

The XPS survey (Fig. 5a) revealed that the 2% WO3/CNNT composite is composed of C, N, W, and O elements, which is consistent with the EDS results. The C 1s spectrum (Fig. 5b) presents two principal peaks located at 284.8 eV and 288.4 eV, respectively, belonging to the sp2 C-C bonds and the sp2-hybridized N-C=N bonding. Furthermore, a weak peak at 286.3 eV is due to incomplete polymerization of terminal amino groups at the edge of heptazine units.42 The N 1s spectrum (Fig. 5c) shows that there are four peaks, located at 398.5 eV, 399.2 eV, 400.9 eV, and 404.4 eV, originating from the C-N=C bond, N-(C)3 bond, C-NH2 bond, and the π-excitation of C-N heterocycles, respectively.37 As shown in Fig. 5d, it is evident from the W 4f spectrum that there are two peaks with binding energies of 35.4 eV and 37.6 eV, attributed to W 4f5/2 and W 4f7/2, respectively.43 The high-resolution O 1s spectra (Fig. 5e) exhibits two peaks, at 530.4 eV and 532.2 eV, with W-O-W bond ascribed to oxygen species in WO3 and the other peak is adsorbed oxygen from g-C3N4.44 Above all, the XPS analysis proves the introduction of WO3 to CNNT in the 2% WO3/CNNT composite.

Fig. 5
figure 5

(a) Total XPS survey of 2% WO3/CNNT sample and the high-resolution spectra of (b) C 1s, (c) N 1s, (d) W 4f, and (e) O 1s.

Full size image

Specific Surface Area and Pore Structure

The nitrogen adsorption–desorption isotherms were studied, and the effects of morphology and doping WO3 nanoparticles on the specific surface area and pore size of carbon nitride were also analyzed. As seen in Fig. 6a, the as-prepared catalyst exhibited a type IV adsorption–desorption isotherm with a H3 hysteresis loop, indicating that the sample has an obvious mesoporous structure.45,46 Compared with the BCN, the CNNT has a more significant BET surface area (83.82 m2/g), which is quite likely due to the fact that the hollow nanotube structure, richly endowed with a hierarchical pore structure obtained from supramolecular precursors, can increase the specific surface. Furthermore, the BET surface area of the 2% WO3/CNNT nanotube composite was investigated and reached 108.82 m2/g, larger than that of a single g-C3N4 and WO3, suggesting that composite materials can improve the surface area, thus providing more reaction and adsorption active sites in the photocatalytic process.47,48 The pore size distribution of the composite belongs to a mesoporous structure, which is about 19.14 nm, much larger than that of the single material. The porous structure of the 2% WO3/CNNT facilitates the complete absorbtion of light illuminating its surface and speeds up the mobility of reactants and products throughout the reaction process.

Fig. 6
figure 6

(a) N2 adsorption–desorption isotherms and (b) corresponding pore size distribution curves of the synthesized WO3, BCN, and 2% WO3/CNNT.

Full size image

Optical Properties

UV–vis DRS analysis was used to study the optical absorption region of the BCN, CNNT, WO3 and 2% WO3/CNNT composite in the range of 300–600 nm. As shown in Fig. 7a, the BCN and CNNT display an absorption edge in the visible light region whereas WO3 shows the strongest absorption at approximately 497 nm. The as-prepared binary composites exhibit a red shift and more significant UV and visible absorption than the BCN and CNNT, which may be attributed to the remarkable porous nanotube structure and coupling between WO3 and g-C3N4. Upon irradiating nanocomposites with visible light, remarkable photocatalytic properties can be achieved by the high absorption in the visible region.

Fig. 7
figure 7

(a) UV–vis diffuse reflectance spectra of WO3, BCN, CNNT, and 2% WO3/CNNT and (b) band gap energies of WO3, BCN, CNNT, and 2% WO3/CNNT.

Full size image

Generally, by using the Tauc plot method, a semiconductor photocatalyst's optical absorption edge (Eg) can be estimated:

$$ \alpha h\nu = A(h\nu - E_{{\text{g}}} )^{n/2} $$
(1)

where the photon energy, light frequency, proportionality and band gap energy are represented by the absorption coefficients of α, h, ν, A, and Eg, respectively. The band gap energy for the CNNT was determined from a plot of (αhv)2 versus hv (n = 4), and the WO3was also calculated from a plot of (αhv)2 versus hv (n = 4). The band gaps of CNNT and WO3 are, respectively, estimated to be 3.01 eV and 2.66 eV (Fig. 7b), nearly the same as those previously reported.49,50 In the composites, the band gap is between individual samples, which can explain the enhancement in light absorption.

Photoelectrochemical Measurements

Mott–Schottky Plots

An investigation of the band edge potentials and heterojunction formation has been conducted using Mott–Schottky plots. From Fig. 8a and b, the positive slope of the plots reveals that both WO3 and CNNT are n-type semiconductors.51 It was found that, for WO3 and CNNT, the bandgap potentials were − 0.02 eV and − 1.2 eV (vs. Ag/AgCl, pH = 7), respectively. The following can be used to analyze the potentials of the conduction band (CB) (compared to a normal hydrogen electrode, NHE pH = 0):

$$ E_{{({\text{NHE}})}} { = }E_{{({\text{Ag}}/{\text{AgCl}})}} + 0.197 $$
(2)

thus, the CB potentials of the WO3 and CNNT are, respectively, estimated to be 0.18 eV and − 1.0 eV (vs. NHE). By further considering their Eg (Fig. 7b), the valence band (VB) potentials of the WO3 and CNNT are, respectively, measured to be 2.84 eV and 2.01 eV (versus NHE) by applying:

$$ E_{{{\text{VB}}}} = \, E_{{\text{g}}} + E_{{{\text{CB}}}} $$
(3)
Fig. 8
figure 8

Mott–Schottky plots: (a) WO3 and (b) CNNT; (c) photocurrent responses and (d) EIS spectra of the samples under visible light.

Full size image

Photocurrent Transient Response Tests

For the purpose of investigating electron transfer during the light-on and light-off modes of the composite system, photocurrent response tests were conducted as shown in Fig. 8c. After the catalysts had been irradiated, the photocurrent dramatically increased. However, the photocurrent rapidly decreased after the light was turned off. There was an obvious improvement in photocurrent intensities in the composites, and the course can be repeated continuously. According to the results, it can be seen that the WO3 promotes the separation of electrons and holes by forming interfaces between the CNNT and the WO3.

EIS Analysis

The EIS tests of the BCN, CNNT, pure WO3 , and 2% WO3/CNNT composite are set out in Fig. 8d. A comparison of the BCN and CNNT samples shows a large arc radius of Nyquist plots; however, for the 2% WO3/CNNT composite, the arc radius clearly decreases, suggesting a better transport of charges. However, the radius of the Nyquist diagram of WO3 sample is small. Therefore, the charge transfer rate in this composite system may not be a critical factor in high degradation efficiency.

Photocatalytic Properties of the Samples

As a representative dye for the photo-oxidation process, the RhB molecule (maximum absorption band at 554 nm) was employed to study the photocatalytic activity of the photocatalysts. As shown in Fig. 9a and b, the photocatalytic results indicated that pure WO3 nanoparticles and BCN have insufficient degradation efficiency of RhB. However, the degradation efficiency of CNNT was significantly improved, indicating that morphology regulation could affect the degradation efficiency of g-C3N4. Additionally, under visible-light irradiation, the WO3 doping significantly affected the WO3/CNNT composite's photocatalyst performance, the 1%, 2%, 3%, and 5% WO3/CNNT exhibiting higher photocatalytic efficiencies in degrading RhB compared with CNNT. This improvement of photocatalytic performance may be due to the enhancement of carrier separation efficiency. Among them, the 2% WO3/CNNT is most prominent, with a high k value of 0.335 min−1, which is 3.4 times greater than CNNT and 83.8 times greater than single WO3. When the proportion of WO3 was further increased to 7%, the photocatalytic activity decreased. It may be that the active site of CNNT is blocked by excessive WO3 nanoparticles, and the photoadsorption sites of CNNT are covered. In Fig. 9c, the cyclic photocatalytic degradation tests were carried out with the 2% WO3/CNNT composites. The RhB degradation rate only slightly declined, from 97.4% to 91.8%, during the course of the four cycles of the 2% WO3/CNNT composite, revealing its excellent stability and reusability. The different scavengers (AO, BQ, DMSO, IPA) were taken to explore the impacts of the main active species during photocatalytic degradation of the RhB. The AO, BQ, DMSO, and IPA have been added for quenching the h+, ·O2−, ·OH, and e− radicals, respectively.52,53 From Fig. 9d, it can be seen that the AO and BQ can more successfully inhibit the degradation process than DMSO and IPA during photocatalysis. The results identified in these responses are that the ·O2− and h+ radicals are the pivotal species in the total degradation process.

Fig. 9
figure 9

(a) Photocatalytic RhB adsorption–degradation curves under visible-light irradiation of the synthesized photocatalysts; (b) pseudo-first-order kinetics plot for the RhB degradation over the as-prepared photocatalysts; (c) four cycles of 2% WO3/CNNT for the degradation rate; and (d) photocatalytic RhB under visible-light irradiation degradation with different reactive species scavengers (AO, BQ, DMSO, IPA).

Full size image

As shown in Fig. 10a, there was no apparent absorbance peak measured after 40 min of irradiation, indicating almost complete degradation of TC. As can be seen in Fig. 10b and c, the photocatalyst degradation rate of TC accorded with the following order: 2% WO3 > CNNT > 1% WO3 > 5% WO3 > 3% WO3 > 7% WO3 > BCN > WO3. Similarly, the k value of 2% WO3/CNNT (0.043 min−1) for TC degradation was about 1.1 times higher than CNNT (0.040 min−1), and 14.3 times higher than pure WO3 (0.003 min−1). It can be seen in Fig. 10d that the synthesized catalysts can be used repeatedly for the treatment of antibiotic wastewater by photocatalysis. In Fig. 10e, the collected photocatalyst after the degradation process was characterized by XRD and compared with the composite photocatalyst before the photodegradation cycles. It can be seen that there are no obvious impurities and that the position of the peak remains unchanged. These complete structures imply that the WO3/CNNT composite photocatalyst possesses excellent stability without apparent inactivation.

Fig. 10
figure 10

(a) UV–vis spectra of TC solutions during photocatalytic treatment using the 2% WO3/CNNT catalyst; (b) adsorption–degradation curves; (c) pseudo-first-order kinetics of samples; (d) photocatalytic TC stability test of the 2% WO3/CNNT sample under simulated solar light; and (e) XRD patterns of before and after photodegradation cycles of the 2% WO3/CNNT sample.

Full size image

Mechanism of Enhanced Photocatalytic Performance

The catalytic mechanism of RhB and TC degradation in the WO3/g-C3N4 nanotube composites were inferred according to the above experimental results and discussion, as shown in Scheme 2. It is important to note that in most semiconductor heterojunction photocatalysts, there may be two general charge transfer routes (traditional band to band transfer and Z-type mechanisms) that will compete with each other. By utilizing the double-charge transfer mechanism (Scheme 2a), with sunlight illuminating the g-C3N4 nanotubes, electron–hole pairs are generated, and the electrons are transferred to the WO3 nanoparticles as electron acceptors, owing to the potential of CB of WO3 being below O2/·O2− (− 0.33 eV) and the ·O2− cannot be gained.54 Therefore, we can surmise that the Z-scheme mechanism would be better matched to explain the charge separation process of the WO3/CNNT nanotubes (Scheme 2b). Meanwhile, the photocatalysts consisting of one-dimensional tubes have higher charge carrier mobility along the longitudinal direction, which is beneficial for electron and hole separation and transfer, and can maximize the composites' exposure to light attributable to the incident light's several reflections. Finally, the obtained ·O2− and h+ can easily react with RhB (TC) molecule. A comparison of the degradation of various pollutants using different g-C3N4-based heterojunction photocatalysts is presented in Table I. Based on these results, the synthesized WO3/g-C3N4 composites can effectively remove RhB and TC from wastewater with outstanding photocatalytic properties.

Scheme 2
scheme 2

The possible photocatalytic mechanism of the WO3/g-C3N4 nanotubes: (a) double-charge transfer mechanism and (b) Z-scheme mechanism.

Full size image
Table I Comparing the degradation performance of other g-C3N4-based heterojunction photocatalysts with previous examples
Full size table

Conclusions

A simple one-pot calcination method was used to successfully synthesize tubular g-C3N4 incorporated with WO3 and examined for photocatalytic activity. For visible-light and sunlight degradation, the 2% WO3/CNNT was believed to have the highest removal rate. Under visible light, the degradation efficiency of RhB solution reached 97.4% in 15 min, and the 84.1% TC solution can be degraded in 40 min under the condition of simulated sunlight. In addition, the reaction rate constant, k, reached 0.335 min−1 and 0.043 min−1, respectively. The selectivity may be the reason for the different degradation efficiency of catalysts for different pollutants in sunlight and visible light. According to the active species trapping experimental results, the ·O2− and h+ radicals played crucial roles in the photocatalytic degradation of RhB in 2% WO3/CNNT. The enhanced photocatalytic activity was due to the formation of the Z-scheme heterojunction between the WO3 and g-C3N4, leading to a high surface area and light absorption capacity of the photocatalyst. The current work paves the way for the development of versatile g-C3N4-based heterojunction materials with a specific structure and an enhanced photocatalytic performance.