Adsorption coupling photocatalytic removal of gaseous n-hexane by phosphorus-doped g-C3N4/TiO2/Zn(OAc)2-ACF composites

VOCs emission reduction in the petroleum and petrochemical industry is a hot and difficult topic at present. The single method may not be able to meet the actual treatment status. Therefore, the adsorption coupled photocatalytic degradation technology was used to remove VOCs. Phosphorus-doped carbon nitride (PCN) and PCN/TiO2 were prepared by hydrothermal synthesis and sol–gel method, and then PCN/TiO2/Zn(OAc)2-ACF composites were prepared by ultrasonic impregnation on zinc acetate modified activated carbon fibers (Zn(OAc)2-ACF). The removal efficiency of n-hexane by composite materials was explored in a self-made reactor, and the factors affecting removal efficiency, removal mechanism, and possible ways of degradation were investigated. The results showed that under the optimum reaction conditions (initial concentration of n-hexane 200 mg/m3, space velocity 1000 h−1, light intensity 24 W, mass fraction of doped PCN 6%, loading twice, calcination temperature 450 °C), PCN/TiO2/Zn(OAc)2-ACF composite has the highest removal efficiency of n-hexane (90.2%). The adsorption capacity of the composites after doping the P element was 215.3 mg/g, which did not enhance the adsorption performance compared with that before doping, but the removal rate of n-hexane was higher. This showed that doping P element was helpful to enhance the photocatalytic activity of the composites.


Introduction
Volatile organic compounds (VOCs) are common gaseous environmental pollutants, mainly in the petroleum and petrochemical industry, printing, automobile exhaust, coating manufacturing, and other industries. They mainly include alkanes, olefins, alkynes, benzene series, alcohols, aldehydes, ethers, ketones, acids, esters, halogenated hydrocarbons, etc. (Dobslaw and Ortlinghaus 2020;Sekar et al. 2019;Zhang et al. 2017). Alkanes are the main composition of VOCs emitted artificially, accounting for about 40% of the anthropogenic emissions (Wei et al. 2019;Ziemann 2011). N-hexane (C 6 H 6 ) is a widely used alkane in industry and the most representative nonpolar solvent, mainly from biogas plants. N-hexane is highly fat-soluble and easy to accumulate in organisms. In the absence of protective devices, accumulation occurs in humans exposed to environments as low as 1 ppm Fedtke and Bolt 1986). Short-term exposure to n-hexane can cause irritation, dizziness, headache, and nausea. Longterm exposure can lead to chronic poisoning symptoms such as headache, dizziness, numbness of limbs, and permanent nerve injury (Guo et al. 2022;Yang et al. 2019). In addition, it can promote the generation of secondary organic aerosols, resulting in photochemical smog, seriously affecting air quality and human health, so it is urgent to deal with it (Thanh Truc et al. 2019;Yu et al. 2020).
Several VOCs control techniques including photocatalytic oxidation (PCO), adsorption, membrane filtration, Responsible Editor: Angeles Blanco EnCheng Sun and HaiDi Wei are co-first authors, and they have made equal contributions to this work. biotechnology (such as biofiltration and biotrickling filters), and catalytic combustion have been developed and used to treat alkane (Dobslaw and Ortlinghaus 2020;Zhao et al. 2019;Ziemann 2011). Generally, the treatment of n-hexane by membrane separation is often accompanied by high cost and low efficiency. The energy consumption of catalytic combustion is very high. The biological method has the advantages of high removal rate, simple equipment, low operation cost, and less secondary pollution (Dobslaw and Ortlinghaus 2020). Mokhtari et al. studied the effect of rhamnolipid biosurfactant on the filtration of n-hexane vapor in a laboratory-scale biofilter. Studies have shown that adding appropriate concentration of biosurfactant (rhamnolipid) in biofilter can significantly improve the removal rate of n-hexane ). Abdolahnejad et al. used a hybrid system composed of TiO 2 membrane scoria particles filled photoreactor and biofilter to improve the removal rate of n-hexane in biofilter. The experimental results show that the hybrid system has higher removal efficiency than single biological filter process . However, microorganisms have high requirements for growth environment and are sensitive to temperature and humidity changes. Therefore, there is a great challenge to design novel catalysts with the dual effects of photodegradation and adsorption for removal (Tian et al. 2017).
Adsorption technology is the most effective and has been widely used. Activated carbon, activated alumina, silica gel, and zeolite are commonly used adsorbents (Guo et al. 2008;Shi et al. 2008). Activated carbon fiber (ACF) is an effective VOCs adsorbent. ACF has many advantages such as high surface adsorption reactivity, uniform micropore structure, good reproducibility, and large specific surface area, but it is prone to secondary pollution (Das et al. 2004;Lin et al. 2012;Liu et al. 2014;Miyamoto et al. 2005;Yi et al. 2008). To improve the adsorption effect and selectivity, it is often necessary to adjust the pore structure of ACF or modify its surface characteristics. At present, the commonly used modification methods are surface oxidation-reduction, supported metal, and metal oxide (Yi et al. 2008). Bi et al. (2021) modified activated carbon fiber (ACF) with Zn(NO 3 ) 2 , ZnCl 2 , and Zn(OAc) 2 , respectively, and then loaded TiO 2 on the modified ACF. The experiment showed that TiO 2 /Zn(OAc) 2 -ACF had the best toluene degradation performance.
Photocatalytic technology has the advantages of high efficiency, energy-saving, safety, low cost, and mild reaction conditions, which is widely used to treat sewage and gas pollution (C. Hou et al. 2021;Mamaghani et al. 2017). However, photocatalytic oxidation technology has the disadvantages of catalyst deactivation, easy recombination of photogenerated electrons and holes, and easy agglomeration of carriers. The commonly used catalysts include TiO 2 , g-C 3 N 4 , bismuth-based materials, graphene, and its composites (Fu et al. 2019;Wang et al. 2019;Xie et al. 2019). Among many catalysts, TiO 2 has attracted much attention in recent years due to its outstanding advantages such as low energy consumption, simple operation, wide application range, and no secondary pollution. However, due to the wide bandgap (3.2 eV), pure TiO 2 can only absorb ultraviolet light with a short wavelength, and the utilization of solar energy is poor (Raja et al. 2020). Therefore, TiO 2 needs to be modified to improve its visible light response. There are many modification methods for TiO 2 , such as noble metal modification, semiconductor composite, dye sensitization, and transition metal ion doping . For example, Tian et al. deposited TINFs in ACF felt to prepare ultra-long TiO 2 nanofibers/activated carbon fibers (TiNF/ACF) porous composites with good adsorption performance for toluene (Tian et al. 2017). Shi et al. prepared Fe (III) and Ho (III) codoped TiO 2 thin films with good degradation performance for methyl orange in water by sol-gel-adsorption method (Shi 2009).  (H:TiO 2-x ) nanoparticles by controlled reduction method to improve the utilization of visible light and hydrogen production performance (Sinhamahapatra et al. 2018). Graphitic carbon nitride (g-C 3 N 4 ) is a non-metallic semiconductor catalyst with excellent performance. The raw materials are cheap and easy to get (amino nitrile, urea, melamine, dicyandiamide), narrow bandgap (2.7 eV), good thermal and chemical stability (Wen et al. 2015;Huang et al. 2019). In recent years, g-C 3 N 4 has become a research hotspot of the photocatalyst. However, the photocatalytic activity of g-C 3 N 4 was low because of its small specific surface area and the easy recombination of photogenerated electrons and holes (Fang et al. 2018;Humayun et al. 2018). Accordingly, various methods have been reported to improve its photocatalytic performance, such as doping metal/nonmetal elements, constructing heterostructures, or optimizing morphology (Gao et al. 2019). Recently, some studies have shown that doping P in g-C3N4 can significantly improve its photocatalytic performance. (Gao et al. 2019;Humayun et al. 2018;Li et al. 2016;Liu et al. 2018;Wu et al. 2020). For example, the phosphorus (P)-doped g-C 3 N 4 synthesized by a simple sintering method showed an obvious red shift at the absorption edge (Wu et al. 2020). Li et al. (2016) prepared P (x%)-g-C 3 N 4 /TiO 2 composites showed enhanced light absorption and photocatalytic properties in the visible light region, and had high photocatalytic degradation activity for methyl blue (MB).
Another difficult problem of VOCs photocatalytic oxidation degradation technology is that powder photocatalytic materials are difficult to apply and recover due to the fluidity of gas. To solve the above problems, the powder material needs to be loaded on ACF. Previous studies have shown that the strong adsorption performance of ACF cannot only enrich target pollutants, capture intermediate toxic products, promote the photocatalytic performance of nano-TiO 2 , but also provide support for the renewable performance of TiO 2 photocatalytic materials. In addition, metal oxide-based catalysts loaded on ACF can provide a large number of binding sites on the surface of the adsorbent, which significantly improves the adsorption performance of ACF. Moreover, the introduction of catalyst reduces the blockage of ACF microporous structure, promotes the in situ regeneration of adsorbent, and reduces the risk of adsorbent failure due to the increase of adsorption concentration (Guo et al. 2008;Shi et al. 2008;Tran Thi and Lee 2017).
Under the guidance of the above strategies, TiO 2 modified by the P-doped g-C 3 N 4 photocatalyst was prepared by the sol-gel method in this work. PCN/TiO 2 /Zn(OAc) 2 -ACF composites were prepared by ultrasonic impregnation on zinc acetate modified ACF. Then, by exploring the effects of different factors (calcination temperature, impregnation times, light intensity, initial concentration of n-hexane, etc.) on PCN/TiO 2 /Zn(OAc) 2 -ACF's degradation of n-hexane, the adsorption and photocatalytic mechanism of the catalyst were further studied.

Preparation of ACF modified by zinc acetate (Zn(OAc) 2 -ACF)
Firstly, ACF was pretreated. The ACF with the shape of 2 cm * 2 cm was soaked and washed three times in an 8% ethanol solution to remove organic matter and impurities. Then the product was boiled in deionized water for 2 h. In this process, deionized water was changed every half hour. The pretreated ACF was taken out and dried in an oven at 105 °C to obtain the pretreated ACF. Then the pretreated ACF was put into a 100-mL beaker containing 50 mL Zn(OAc) 2 (0.01 mol/L) solution. Then, in order to make the solution uniformly and fully soak ACF, the beaker was placed in the vacuum chamber for 10 min, and the ultrasonic-vacuum impregnation process was repeated twice. Finally, the impregnated ACF was dried at 105 °C and marked as Zn(OAc) 2 -ACF.

Preparation of PCN material
Phosphorus-doped carbon nitride (PCN) was prepared by hydrothermal synthesis (Fig. 1a). Melamine (C 3 H 6 N 6 ) (1.26 g) was added into a 200-mL beaker containing 60 mL dimethyl sulfoxide (DMSO) to stir evenly, denoted as solution A. Cyanuric acid (C 3 H 3 N 3 O 3 ) (1.29 g) was added to DMSO (50 mL), and magnetically stirred to a uniform mixing and denoted as solution B. Slowly drop solution B into solution A with uniform stirring speed and continue stirring for 2 h. The obtained suspension was centrifuged in a highspeed centrifuge and washed three times with anhydrous ethanol and water, respectively.
The centrifugal product was put into the hydrothermal synthesis reactor, adding 60 mL water and 0.01 mol (3.92 g) sodium phosphate, stirring for 2 h, and then heating at 180 °C for 8 h. The precursor was obtained by centrifugal drying of the material. The precursor was heated at 520 °C for 4 h in a tube furnace protected by nitrogen to obtain phosphorus-doped carbon nitride (PCN).

Characterization
The elemental composition and valence state of the surface of the composites were analyzed by an X-ray photoelectron spectrometer (XPS, Thermo Scientific, K-Alpha). The crystal structure and grain size of the composites were determined by X-ray diffraction spectra (XRD, PANalytical B.V., AXIOS-Petro) with power 2200 W, Cu-Kα radiation (λ = 1.54187 Å), scanning angle 10 ~ 80°, step 0.02°/s. The functional groups of the samples were characterized by Fourier transform infrared spectrometer (FTIR, Nicolet, Nexus). The size and morphology of the samples were observed by scanning electron microscope (SEM, Hitachi, S4800) with a scanning voltage of 50 kV. TEM was performed by using a JEM-2100UHR transmission electron microscope with analysis voltage of 200 kV. Adsorption was carried out at 77.4 K with N 2 as adsorbate, and desorption occurred at 300 K. The specific surface area, pore size, and pore volume distribution of the samples were measured by the analysis of Brunauer-Emmett-Teller (BET, ASAP3020, Mike Instruments Co.). The photoluminescence spectra of samples were measured by F97Pro fluorescence spectrophotometer (Semerfeld, China). I-T photocurrent analysis can be used to determine the light response of materials. Samples/conductive glass is used as working electrode, platinum wire as a counter electrode, and saturated calomel electrode as a reference electrode. The samples were subjected to visible light irradiation (300-W Xe lamp, Institute of Perfect Light Sources, Beijing) and a photocurrent test at a bias voltage of 0.0 V. The optical absorption properties of the composites were measured by ultraviolet-visible diffuse reflectance spectrophotometer (UV-vis DRS, TU-1901, Beijing General Instrument Co.).

N-hexane removal experiment
The composites were fixed in the quartz tube reactor. Then, the self-assembly experimental device was operated under normal temperature and pressure, RH (relative humidity) = 50%, stable airflow, and no pollutants. Then a dark reaction was carried out to make the composite completely absorb n-hexane (the initial concentration of 200 mg/m 3 ). After reaching the adsorption equilibrium, the low-pressure mercury lamp was turned on for photocatalytic reaction. During the experiment, the inlet and outlet gas of n-hexane were sampled regularly by activated carbon sampling tube. And the effects of mass fraction of PCN, calcination temperature of composites, loading times, light source, light intensity, initial concentration of n-hexane, and space velocity of n-hexane on the removal performance of composites were investigated.
The samples were resolved with carbon disulfide and injected into the gas chromatography (GC, SP-3420A) to determine the concentration of n-hexane. The removal efficiency of n-hexane (R) was calculated by Eq. (1): where C 0 (mg/m 3 ) and C (mg/m 3 ) represent the concentration of n-hexane at import and export, respectively.

Characterization analysis
To further explore the chemical composition of the sample, the PCN is analyzed by XPS. XPS spectra (Fig. 2a) shows that PCN consists of four elements: C, N, O, and P. The peaks at 284.51 and 287.86 in high-resolution C 1 s XPS spectra (Fig. 2b) correspond to C-C and N-C = N. The high-resolution N 1 s XPS spectrum (Fig. 2c) consists of three peaks, which can be assigned to C-N = C (398.29 eV), N-(C) 3 (399.92 eV), and N-H (400.94 eV). The peak at 531.80 in high-resolution O1s XPS spectra (Fig. 2d) corresponds to O-H. The peak at 132.20 in high-resolution P2p XPS spectra (Fig. 2e) corresponds to P-N (Z. Li et al. 2016). It is proved that P was successfully doped into the lattice of g-C 3 N 4 .
XRD and FTIR spectra are conducted to collect the crystalline and framework structure of the pristine and composite samples; the results are shown in Fig. 3a and b. According to the PDF card (PDF#21-1272), the anatase phase characteristic diffraction peaks of TiO 2 composites supported on modified ACFs with different zinc salts are observed at 2θ = 25.37°, 38.60°, 48.07°, 54.10°, 54.86°, and 62.58°, corresponding to 101, 004, 200, 105, 211, and 204 crystal planes, respectively. The g-C 3 N 4 sample presents typical peaks at 13.10° and 27.50°, which are assigned to (1 0 0) and (0 0 2) planes (JCPDS 87e1526). The weak peak at 13.10° results from the in-planar structure of g-C 3 N 4 , while the peak at 27.50° could be ascribed to the stacking aromatic structure (Chen et al. 2020;Hou et al. 2012). The characteristic peak of the PCN sample is consistent with that of carbon nitride and has high peak strength, indicating that the crystallinity is increased. With the PCN ratio increasing, TiO 2 diffraction peaks at 25.4°and 27.5° are shifting towards the high angle indicating the chemical interaction between PCN and TiO 2 , which further confirms the PCN/TiO 2 heterojunction formation. FT-IR spectra of PCN/TiO 2 /Zn(OAc) 2 -ACF composites are shown in Fig. 3b. The spectra show that all the   3 a XRD patterns of pure g-C 3 N 4 , PCN, and 2%, 4%, 6%, and 8% mol PCN/TiO 2 /Zn(OAc) 2 -ACF catalysts; b FTIR spectra of TiO 2 , PCN, and PCN/TiO 2 /Zn(OAc) 2 -ACF catalysts spectra of the samples are the combination of characteristic peaks of TiO 2 and g-C 3 N 4 . Specifically, the peak at about 470 cm −1 is the characteristic peak of anatase TiO 2 , which is attributed to the tensile vibration of Ti-O-Ti. The sharp absorption peak at 806 cm −1 is attributed to the out-of-plane bending vibration of tri-s-triazine units of g-C 3 N 4 , and the strong absorption bands between 1240 and 1640 cm −1 are attributed to the typical C-N and C = N stretching vibrations of the tris-s-triazine ring. The broad bands ranging between 3000 and 3600 cm −1 are attributed to the stretching vibration of O-H groups in water molecules adsorbed on the material surface (Du et al. 2020). As shown in Fig. 3b, all photocatalysts exhibited a typical IR pattern of g-C 3 N 4 , which indicated that PCN has been loaded onto TiO 2 /Zn(OAc) 2 -ACF composites.
As shown in Fig. 4a, the surface of ACF modified by zinc acetate is slightly rough and owns micropores and mesopores, which provide attachment sites for photocatalytic materials and improve the adsorption performance. The pure g-C 3 N 4 monomer was a nano-scale tubular structure (Fig. 4b). From the SEM and TEM images of the PCN monomer in Fig. 4c, d, it can be seen that PCN do not change this structure. The diameter of PCN is 98-243 nm, and the smaller diameter of PCN is conducive to the transmission of electrons and holes. At the same time, it increases the overall specific surface area and increases the active sites of contact with pollutants and photocatalytic reactions. It can be seen from Fig. 4c, e that PCN has a massive structure and large particle size, while the PCN/TiO 2 composite increases the specific surface area. Figure 4f and g are the TEM images of PCN/TiO 2 composites. It can be observed that the composites are composed of black granular TiO 2 and gray lamellar PCN. It is proved that TiO 2 particles are uniformly distributed in the PCN tubular nanosheets, and the corresponding heterojunction (001) lattice and TiO 2 anatase (101) lattice are obtained by calculating the lattice spacing. However, since PCN is a semiconductor, the PCN lattice is not measured by TEM. The SEM images of PCN/TiO 2 / Zn(OAc) 2 -ACF composites are shown in Fig. 4h, i. It can be observed that PCN/TiO 2 is successfully loaded on the surface of Zn(OAc) 2 -ACF.
The N 2 absorption desorption measurements are carried out, and the results are plotted in Fig. 5. All samples show type IV isotherms, indicating the presence of mesopores. The specific surface area, pore diameter, and pore volume are shown in Table 1. The specific surface area of Zn(OAc) 2 -ACF is large, which can reach 1220.77 m 2 /g, which is because the void on the surface of ACF is not filled by catalyst particles as yet in the SEM image. And the pore diameter and specific surface area of PCN/TiO 2 /Zn(OAc) 2 -ACF decreased with the increase of photocatalyst loading (Geng et al. 2019). The pore size and specific surface area of 6% PCN/TiO 2 /Zn(OAc) 2 -ACF composite were 1.81 nm and 863.96 m 2 /g, respectively. The above results are consistent with the nitrogen adsorption-desorption curve.
In order to study the separation, migration, and recombination properties of photogenerated electron-hole pairs, photoluminescence (PL) and photocurrent response are characterized. The photoluminescence (PL) spectra of the bulk g-C 3 N 4 , PCN and PCN/TiO 2 are shown in Fig. 6a. The bulk g-C 3 N 4 and PCN have strong characteristic emission peaks at 465 nm, and the peak intensity of PCN is lower than that   . 6 a PL spectra of bulk g-C 3 N 4 , PCN, and PCN/TiO 2 ; b photocurrent response curves of 2%, 4%, 6%, 8% PCN/TiO 2 /Zn(OAc) 2 -ACF composts of pure g-C 3 N 4 . This shows that the doping of appropriate P in g-C 3 N 4 greatly inhibits the recombination rate of charge carriers, thus promoting the accumulation of photoinduced electrons and holes in the conduction and valence bands of g-C 3 N 4 to produce high-concentration reactive oxygen species (e.g., ·OH and ·O 2 − ). The photoluminescence intensity of PCN/TiO 2 is significantly weakened, indicating that the recombination of photocarriers is inhibited by the heterojunction between TiO 2 and g-C 3 N 4 . The photocurrent response of PCN/TiO 2 /Zn(OAc) 2 -ACF with different mass fractions is shown in Fig. 6b. With the deposition and sculpture-reduction process, the transient photocurrents of the samples exhibit an obvious enhancement under the continuously illumination on and off. The photocurrent response of PCN/TiO 2 /Zn(OAc) 2 -ACF is significantly enhanced compared with bulk g-C 3 N 4 . The 6% PCN/TiO 2 /Zn(OAc) 2 -ACF has the strongest photocurrent response, where the increased photocurrent can be mainly attributed to the efficient photogenerated separation and transfer, which is beneficial to the photocatalysis and corresponds to the results.
The UV-vis diffraction spectra of the composites were characterized to study their light absorption properties and determine their band gap. The results in Fig. 7a show that the pure g-C 3 N 4 sample has a strong absorption band in the ultraviolet-visible region, while the anatase TiO 2 sample has weak visible light response. The absorption edge of TiO 2 and g-C 3 N 4 are about 380 nm and 470 nm, respectively. Kubelka-Munk equation (Brill and Li 2016) is used to calculate the optical gap, and the equation is shown as Eq. (2): where K, S, and R represent absorption coefficient, reflection coefficient, and diffuse reflectivity (%), respectively. The energy band gap of the TiO 2 , g-C 3 N 4 , and PCN can be calculated by the curve of the plot of [F(R) × hν] n/2 vs hν (h is Planck constant (eV·s) and v is light frequency). Substitute n = 1 into the formula to get the bandgap widths of TiO 2 , g-C 3 N 4 , and PCN which is 3.18 eV, 2.75 eV, and 2.61 eV (inset in Fig. 6a), respectively. The band gap difference between PCN and g-C 3 N 4 is caused by P doping in PCN photocatalyst. The spectra of PCN/TiO 2 and PCN/TiO 2 / Zn(OAc) 2 -ACF samples have a red shift compared with those of TiO 2 samples. This can be attributed to the synergistic reaction between PCN and TiO 2 , which reduces the band gap energy and changes the optical properties. On the other hand, some chemical bonds formed between the two semiconductors may result in enhanced optical properties. In order to further study the band structure of the composites, the valence band edges (VB) of TiO 2 , g-C 3 N 4 , and PCN are detected by VB-XPS, as shown in Fig. 7b. When the VB edge of anatase TiO 2 appears at 2.54 eV, the CB value is − 0.64 eV. The VB edge of g-C 3 N 4 is located at 1.98 eV, and its CB is located at − 0.77 eV. The VB edge of PCN is 1.78 eV and its CB is − 0.83 eV. Since the CB of anatase TiO 2 is lower than that of PCN, there is sufficient Gibbs free energy to induce the electrons to inject from PCN into TiO 2 , which is conducive to separating the electron-hole pairs.

Photocatalytic performance
When 200 mg/m 3 n-hexane is used as the target pollutant, the effect of PCN mass fraction on the removal performance of the composite is investigated (Fig. 9a). The n-hexane degradation efficiency of PCN/TiO 2 /Zn(OAc) 2 -ACF composite is higher than that of pure TiO 2 /Zn(OAc) 2 -ACF. And with the increase of PCN mass fraction, the degradation performance increases firstly and then decreases. When the mass fraction of PCN is 6%, the removal rate of n-hexane reaches the best, which is because that the appropriate amount of phosphorus doping can enhance the visible light response of the material and improve the charge separation and transfer rate. In the reaction process, the n-hexane removal rate of composites can reach 100% within the first 30 min which is mainly due to the adsorption of ACF. With the gradual saturation of ACF, the n-hexane removal rate of the composite begins to decline. At the same time, the PCN/TiO 2 photocatalyst degrades the n-hexane gas adsorbed by ACF which can increase the adsorption sites of ACF and maintain the degradation performance of the composites. After about 180 min of reaction, the removal rate tends to be stable, which is because the n-hexane adsorbed by ACF in unit time reaches a dynamic equilibrium with the n-hexane degraded by photocatalyst. After 360 min of reaction, the n-hexane removal rates of pure TiO 2 and 2%, 4%, 6%, and 8% PCN/ TiO 2 /Zn(OAc) 2 -ACF are 65.86%, 76.50%, 78.35%, 90.15%, and 84.50%, respectively.
When 200 mg/m 3 n-hexane is used as the target pollutant, 6% PCN/TiO 2 /Zn(OAc) 2 -ACF composites with calcination temperatures of 350 °C, 450 °C, and 550 °C are placed in the reactor. And n-hexane is continuously introduced to investigate the effect of calcination temperature on the n-hexane  (Fig. 9b). With the increase in calcination temperature, the removal of n-hexane by the composite increases firstly and then decreases. At 360 min, the removal rates of n-hexane by 6% PCN/TiO 2 / Zn(OAc) 2 -ACF calcined at 350 °C, 450 °C, and 550 °C are 84.50%, 90.15%, and 80.50%, respectively. It can be seen that the removal rate of the material is the highest when the calcination temperature is 450 °C. BET analysis shows that with the increase of calcination temperature promotes the transformation of TiO 2 to the anatase phase and promotes the effective combination of PCN and TiO 2 , which improves photocatalytic activity. However, the decomposition reaction of PCN occurs at high temperatures, which leads to the decrease of its content in the composite. The grain size of TiO 2 will gradually increase at high temperatures and even appear sintering phenomenon. The specific surface area decreases and the catalytic activity is inhibited.
When 200 mg/m 3 n-hexane is used as the target pollutant, 6% PCN/TiO 2 /Zn(OAc) 2 -ACF composites which is dipped once, twice, and three times are put into the reactor, respectively. The effect of impregnation times on the n-hexane removal performance of the composites is shown Fig. 9c. With the increase in impregnation times, the removal efficiency of n-hexane by the composites increases firstly and then decreases. At 360 min, the removal rates of PCN/TiO 2 / Zn(OAc) 2 -ACF-1, PCN/TiO 2 /Zn(OAc) 2 -ACF-2, and PCN/ TiO 2 /Zn(OAc) 2 -ACF-3 are 83.75%, 90.15%, and 86.25%, respectively. It can be seen that PCN/TiO 2 /Zn(OAc) 2 -ACF-2 has the highest n-hexane removal rate. This is because the amount of PCN/TiO 2 loaded on ACF increases with the increase in impregnation times. Then the photocatalytic performance of the composite increases and the removal rate increases. However, after twice impregnation, the specific surface area and average pore volume of ACF decrease sharply due to excessive loading and agglomeration, so the adsorption performance of the composites for n-hexane decreases sharply. The n-hexane contact with PCN/TiO 2 also decreases and the photocatalytic degradation rate of n-hexane by PCN/TiO 2 decreases. It also shows that there is a synergistic effect between the adsorption performance of ACF and the photocatalytic performance of PCN/TiO 2 .
The n-hexane with different initial concentrations is continuously injected, and 6% PCN/TiO 2 /Zn(OAc) 2 -ACF composite is put into the reactor. The effect of initial n-hexane concentration on the n-hexane removal performance of the composite is shown Fig. 10a. With the increase of initial concentration, the removal efficiency of n-hexane gradually decreases. At 360 min, the removal rates of 100 mg/ m 3 , 200 mg/m 3 , 500 mg/m 3 , and 1000 mg/m 3 are 93.25%, 90.15%, 81.25%, and 59.48%, respectively. This is because the number of n-hexane molecules passing through the composite increases in unit time with the increase of initial concentration. However, the adsorption sites and photocatalytic sites of the composites are limited, and increasing the concentration of hexane concentrations increases probability to pass through the package without contact. Therefore, Fig. 10 Effect on removal rate of n-hexane by a initial concentration of n-hexane, b space velocities, d light intensity; c comparison of photocatalytic and dark adsorption effects on n-hexane removal n-hexane removal performance decreases with the increase of initial concentration. As can be seen from Fig. 10b, the degradation rates of n-hexane of PCN/TiO 2 /Zn(OAc) 2 -ACF under the air velocity of gasses of 1000 h −1 , 2000 h −1 , 4000 h −1 , and 6000 h −1 are 90.15%, 84.56%, 77.25%, and 66.15%, respectively. This indicates that the removal rate decreases with the increase of the flow rate, and when the airflow velocity is 1000 h −1 , the removal efficiency of n-hexane is the highest. This is due to the decrease in the residence time of n-hexane molecules on ACF as the flow rate increases, thereby reducing the degradation efficiency of PCN/TiO 2 photocatalysts.
The 6% PCN/TiO 2 /Zn(OAc) 2 -ACF composite is put into a self-made reactor and the n-hexane gas is injected into the reactor (Fig. 10c). The removal efficiency of n-hexane gas by the composites is compared under light and dark conditions. With the progress of the experiment, the photocatalytic removal performance of n-hexane will decrease with the passage of time regardless of whether the light is used. Under dark conditions, ACF gradually tends to be saturated and the degradation performance of composite materials for n-hexane begins to decline sharply which decreases to 0% in 210 min with the increase of reaction time. However, the removal rate of n-hexane by composite materials slowly decreases under light conditions. When the reaction time is about 180 min, the removal rate tends to be stable, about 90.15%. The adsorption performance of the composite increases the residence time of n-hexane, which improves the photocatalytic degradation performance of the composite for n-hexane. At the same time, PCN/TiO 2 catalyzes the degradation of n-hexane adsorbed by ACF, thereby increasing the free adsorption sites in ACF. The experiments show that the PCN/TiO 2 /Zn(OAc) 2 -ACF composite combines the adsorption advantage of ACF with the photocatalytic degradation advantage of PCN/TiO 2 , which improves the performance of hexane degradation.
It can be seen in Fig. 10d that the n-hexane degradation rates of 6% PCN/TiO 2 /Zn(OAc) 2 -ACF under 8 W, 16 W, and 24 W light intensity are 76.25%, 83.25%, and 90.15%, respectively. The results show that the degradation efficiency of n-hexane increases gradually but not exponentially with the increase of light intensity. This is because the photon energy determines the activation energy of the reaction but is restricted by the limited number of, but is limited by the number of active sites.
The 6% PCN/TiO 2 /Zn(OAc) 2 -ACF composite is used for the cyclic experiment of photocatalytic removal of n-hexane to evaluate its photocatalytic stability. After each reaction, the composites are regenerated by calcination at high temperature (450 °C). After five cycles, the removal rate of n-hexane decreases slightly from 90.15 to 87.34%, indicating that the catalyst has excellent stability (Fig. 11a). The XRD pattern of the composite material after the reaction does not change significantly (Fig. 11b), indicating that the composite material has excellent stability.

Adsorption capacity of n-hexane on PCN/TiO 2 / Zn(OAc) 2 -ACF composites
In order to analyze the adsorption performance of PCN/ TiO 2 /Zn(OAc) 2 -ACF composite for n-hexane, the adsorption capacity and removal rate of ACF, PCN/TiO 2 , and 6% PCN/TiO 2 /Zn(OAc) 2 -ACF composite for n-hexane gas are described in Fig. 12. The specific operation steps of the experiment are as follows: (i) the composite material is put into the adsorption coupled photocatalytic device, in Fig. 11 a Cyclic experiment of PCN/TiO 2 /Zn(OAc) 2 -ACF; b XRD patterns of 6% PCN/TiO 2 /Zn(OAc) 2 -ACF samples before and after five cycles which the stable dynamic gas is bubbling out by the bubble method; (ii) the adsorption experiment of composite materials is carried out in a lightless closed environment; (iii) under light conditions, the photocatalytic degradation of n-hexane is carried out to calculate the real-time removal rate. The experimental results show that the adsorption capacities of ACF, PCN/TiO 2 , and PCN/TiO 2 /Zn(OAc) 2 -ACF composites for n-hexane gas are 235.5 mg/g, 16.88 mg/g, and 215.3 mg/g, respectively. Although the specific surface area of 6% PCN/TiO 2 /Zn(OAc) 2 -ACF composites (911.56 m 2 /g) is much smaller than that of ACF (1220.77 m 2 /g), the adsorption capacity of 6% PCN/ TiO 2 /Zn(OAc) 2 -ACF composites did not decrease significantly. This is mainly due to the existence of Ti-O or -OH bonds which provide a large number of active sites for n-hexane adsorption. Under UV irradiation, the removal efficiency of n-hexane by ACF decreases rapidly in the first 4 h. With the extension of adsorption time, the adsorption equilibrium of PCN/TiO 2 /Zn(OAc) 2 -ACF composite is achieved. ACF provides abundant active sites and crisscross channels for adsorption. Interestingly, the PCN/ TiO 2 /Zn(OAc) 2 -ACF composite exhibits strong n-hexane adsorption coupled photocatalytic synergy, rather than a simple linear combination of PCN/TiO 2 photocatalytic degradation and ACF adsorption.

Mechanism of photocatalysis
The synergistic mechanism of adsorption coupling photocatalysis of PCN/TiO 2 /Zn(OAC) 2 -ACF composite on n-hexane is shown in Fig. 13. The removal of n-hexane is mainly due to the synergistic effect of ACF adsorption and photocatalytic degradation of photocatalyst PCN/TiO 2 . The specific steps are as follows.
Firstly, n-hexane molecules are adsorbed on the surface of the composite, especially the active sites of ACF. It is well known that ACF is a large number of crisscross fiber rods with ravines and bulges. And the modification by zinc acetate makes its surface rougher and has more active sites. The adsorbed n-hexane on the composite promotes the close contact between n-hexane and PCN/TiO 2 , which accelerates the photocatalytic reaction. In addition, ACF has no photocatalytic activity as a non-semiconductor material, but it significantly enhances the photodegradation performance of Schematic diagram of the mechanism of removing n-hexane from PCN/TiO 2 / Zn(OAc) 2 -ACF composites the composites by preventing electron-hole recombination, improving the photocatalytic adsorption performance, and accelerating the adsorption of n-hexane on the surface of the composites. ACF in the composites has a large electron-hole pair capacity, which can accept the transferred electrons in PCN/TiO 2 , thereby inhibiting the recombination of electron-hole pairs (consistent with the PL spectrum) (Fig. 6a).
Based on the above results, it can be concluded that the degradation mechanism of n-hexane in PCN/TiO 2 / Zn(OAc) 2 -ACF composites (Fig. 13) is an S-type heterojunction mechanism. There are three main reasons. Firstly, the VB (2.54 eV) potential of TiO 2 is higher than the standard oxidation potential of OH − /·OH (2.4 eV vs NHE), and the VB (1.78 eV) potential of PCN is lower than the standard oxidation potential of OH − /·OH (2.4 eV vs NHE), so only the h + on the valence band of TiO 2 can produce ·OH. At the same time, because the CB (− 0.83 eV) potential of PCN is more negative than the standard oxidation potential of O 2 /·O − 2 (− 0.33 eV vs NHE), the photoinduced electrons on PCN can directly produce ·O − 2 (Long et al. 2018). Besides, due to the high positions of CB and VB of PCN, the work function is smaller than that of TiO 2 . When these two semiconductors are in close contact, electrons in PCN spontaneously diffuse to TiO 2 , forming an electron depletion layer and electron accumulation layer near the interface of PCN and TiO 2 , respectively. TiO 2 is negatively charged and PCN is positively charged. And form the internal electric field from PCN to TiO 2 . This internal electric field accelerates the transfer of photogenerated electrons from TiO 2 to PCN. Secondly, the Fermi level of TiO 2 surface is lower than that of PCN. When TiO 2 and PCN contact, their Fermi energy should be at the same level. This causes the Fermi (7) 38 ⋅ OH + C 6 H 14 → 6CO 2 + 26H 2 O levels of TiO 2 and PCN to move upward and downward, respectively. Band bending promotes the recombination of photogenerated electrons in CB of TiO 2 in the interface area and holes in VB of PCN. Thirdly, the photogenerated electrons in TiO 2 CB and holes in PCN VB tend to recombine at the interface under the Coulomb attraction between holes and electrons. In conclusion, the useless electrons and holes are eliminated by recombination, while the strong photogenerated electrons in CB of PCN and holes in VB of TiO 2 are retained to participate in the photocatalytic reaction Zhang et al. 2022).
Finally, the empty activated sites on ACF can continue to adsorb n-hexane molecules to realize in situ regeneration of ACF.

Conclusions
In summary, PCN/TiO 2 /Zn(OAc) 2 -ACF was designed firstly and prepared for the dynamic degradation of n-hexane. XPS, XRD, and SEM proved that P element was successfully doped into g-C 3 N 4 , and PCN and TiO 2 were successfully composited and successfully loaded onto ACF. UV-vis showed that P-doped g-C 3 N 4 resulted in band gap narrowing and spectral response expansion, and the light response of TiO 2 was improved after PCN was combined with TiO 2 . At the same time, PCN and TiO 2 formed an S-type heterojunction, which promoted the separation and transfer of photogenerated electron-hole pairs and extended the life of charge carriers, which was confirmed by electrochemical analysis and PL spectra. Due to the synergistic effect between PCN/TiO 2 and ACF, PCN/TiO 2 /Zn(OAc) 2 -ACF composite has excellent photodegradation and adsorption properties for n-hexane. Under the optimal reaction conditions (initial concentration of n-hexane 200 mg/m 3 , space velocity 1000 h −1 , light intensity 24 W, mass fraction of doped PCN 6%, loading twice, calcination temperature 450 °C), the adsorption capacity of the composite was 215.3 mg/g. In addition, ACF significantly enhanced the photocatalytic activity of the composites because it could hinder the recombination of electron-hole pairs, enhance the photocatalytic adsorption capacity, and accelerate the adsorption of n-hexane. Under the optimum reaction conditions, 6% PCN/TiO 2 /Zn(OAc) 2 -ACF composite showed excellent n-hexane degradation performance with a removal rate of 90.15%, which was 1.37 times higher than that of pure TiO 2 / Zn(OAc) 2 -ACF. The composites exhibit excellent stability and reusability.