A Novel Sustainable Semiconductor/Metal-organic Framework Coated Electret Filter for Simultaneous Removal of PM 2.5 and VOCs

A MIL-125-NH 2 metal-organic framework (MOF) coated electret filter, named E-MOFilter, was previously developed to simultaneously remove PM 2.5 (particulate matter less than 2.5 µ m in aerodynamic diameter) and volatile organic compounds (VOCs, e


INTRODUCTION
The statistical data have shown that people spend 87-90% of their time indoors (Klepeis et al., 2001). However, the concentrations of indoor air pollutants, e.g., fine particulate matter, PM2.5 (particulate matter with an aerodynamic diameter less than 2.5 µm), and volatile organic compounds (VOCs), are often 2-5 times higher than the outdoor concentrations (Huang et al., 2018;U.S. EPA, 2017). Besides, due to the high concentration of ambient PM2.5, particles can infiltrate into indoor environments to pose threat to people's health (Chang et al., 2019;Tang et al., 2018a). Many researches have shown that PM2.5 and VOCs can lead to acute and chronic effects on humans' respiratory and central nervous systems and eventually cause hematological problems and cancer (Brasche and Bischof, 2005;Chen et al., 2019;Sarigiannis et al., 2011;Zhu et al., 2021). Therefore, it is urgent to simultaneously remove PM2.5 and VOCs to enhance indoor air quality.
Recently, metal organic frameworks (MOFs) coated electret filter, named E-MOFilter, has been developed to simultaneously and effectively remove PM and VOC. It had a PM filtration efficiency of 85% and toluene adsorption efficiency of 90% . However, the E-MOFilter reduces VOCs primarily through adsorption, which only captures VOC pollutants. This non-destructive adsorption process is simple but reduction of adsorption due to saturation and desorption by concentration gradient is of great concern (Schreck and Niederberger, 2019). Therefore, to also destroy the VOCs, it is rational to utilize the photocatalytic oxidation (PCO) technology which was discovered by Fujishima andHonda in 1972 (Fujishima andHonda, 1972). The photocatalysts adopt solar energy (or visible lights) and operate at room temperature to degrade VOCs to harmless final products (mostly H2O and CO2). Inspired by the merits of PCO technology, extensive efforts have been made to expand the application of PCO in the mitigation of indoor air pollutants (Mo et al., 2009). In PCO technology, TiO2, is the most widely used catalyst because of its stability, corrosion resistance, and non-toxicity (Augugliaro et al., 1999;Li et al., 2011;Maira et al., 2001;Sun et al., 2012;Xie et al., 2020). However, its high excitation energy, low electron transfer rate to oxygen, and high recombination rate of electron/hole pair limit its photocatalytic performance (Dong et al., 2015). Recently, MOF/semiconductor composites, such as ZIF-8/TiO2, ZIF-8/ZnO, MIL-100(Fe)/TiO2, UiO-66-NH2/TiO2, UiO-66-NH2/g-C3N4, MIL-125(Ti)-NH2/TiO2, MIL-125(Ti)/GO, MIL-125-NH2/g-C3N4, MIL-125-NH2/BiOCl, MIL-125-NH2/Bi2WO6, exhibit significantly improved catalytic performance as compared with the individual components, and have drawn an increasing attention (He et al., 2019;Huang et al., 2019;Li et al., 2018;Wang et al., 2016Wang et al., , 2019Xu et al., 2020;Zeng et al., 2016;Zhang et al., 2018Zhang et al., , 2020Zhu et al., 2023). In these new composites, MOFs provide abundant adsorption sites to pre-concentrate and store a large amount of VOCs, meanwhile, the semiconductor particles serve as photocatalysts to photodegrade the VOC molecules that diffuse from the internal pores of MOF to the surfaces of semiconductors (Qian et al., 2021). For instance, Li et al. (2018) combined NH2-MIL-125 (Ti) and graphene oxide (GO) to form a heterojunction photocatalyst via a microwave heating technique. The prepared 10-GO/NH2-MIL-125(Ti) sample exhibited much enhanced photocatalytic efficiency of gaseous acetaldehyde under visible-light compared to its two pure components. It was concluded that the enhanced photocatalytic performance of the composite was a comprehensive result of the increased visible-light absorbance, more electron carrier density, more efficient charge transfer and less electron-hole recombination rate. Huang et al. (2019) developed a composite photocatalyst TiO2@NH2-MIL-125 via an in-situ solvothermal method for formaldehyde (HCHO) removal under UV irradiation. Their results showed that the synthesized TiO2@NH2-MIL-125 exhibited significantly enhanced photocatalytic performance for HCHO removal (HCHO removal rate of 90%) in comparison with pure TiO2 and NH2-MIL-125, which was owing to the synergistic combination of high adsorption capacity of NH2-MIL-125, high dispersion of TiO2 and efficient interfacial charge transfer between TiO2 and NH2-MIL-125. Zhang et al. (2020) developed a TiO2-UiO-66-NH2 nanocomposites through a simple solvent evaporation method. This nanocomposite photocatalyst demonstrates 72.7% and 70.74% conversion efficiency for toluene and acetaldehyde, respectively, which was about 9.7 and 10.5 times higher than that of UiO-66-NH2, due to the enhanced surface area of the photocatalyst and the suppressed recombination rates of the charge carriers. As demonstrated by these studies, it is obvious that the formation of heterojunction can considerably improve the performance of MOF/semiconductor photocatalyst in VOCs remediation.
In this study, to improve the E-MOFilter towards VOC degradation, Bi2WO6/BiOCl was selected as the semiconductor photocatalyst, termed p-BWO, as it showed the photocatalytic oxidation of toluene with a conversion rate 166-fold higher than that of pristine Bi2WO6 and BiOCl, because the layered and amorphous structures of the nanosheets effectively facilitate the separation of electron-hole pairs (Cao et al., 2018). The p-BWO nanosheets were grafted on the surface of MOF MIL-125-NH2, a representative porous MOF, by using a cationic metal-ion adhesive, poly(ethylenimine) (PEI, 50 wt%). This integration technique differs from the common hydrothermal method (an in situ self-assembly method). The specific details were shown in the later Section 2.1.3. The composite photocatalyst (p-BWO@ MIL-125(Ti)-NH2) was then transported to MERV 13 electret filter media to fabricate the photocatalytic filter, named PE-MOFilter. The MERV 13 electret filter here acted as the substrate for the deposition of the composite photocatalyst, which would promise this photocatalytic filter remains charged and has a high PM filtration efficiency based on prior research (Chang et al., 2019;Li et al., 2020;Tang et al., 2018aTang et al., , 2018bTien et al., 2020). Various characterizations, such as XRD, SEM and UV-vis, were performed to demonstrate the effective synthesis of the composite photocatalyst. The photocatalytic efficiency of toluene, a representative indoor VOC, by the PE-MOFilter was examined. As a simultaneous gas-particle removal filter, the PE-MOFilter was also examined for their PM removal efficiency to confirm the coating of the composite photocatalyst would not cause any degradation for PM removal performances. The ultimate goal of this study is to demonstrate the PE-MOFilter not only captures but effectively photodegrades VOC pollutants via the synergistic action of adsorption and photocatalytic oxidation (PCO).

p-BWO synthesis
The p-BWO nanosheets were synthesized via a hydrothermal method according to Cao et al. (2018). In brief, 165 mg Na2WO4·2H2O and 80 mg octadecyltrimethylammonium chloride (OTAC) as the surfactant were dissolved in 40 mL deionized water (DI water). 486 mg Bi(NO3)3·5H2O (1 mmol) milled powders were added to this solution under stirring. After 60 min of stirring, the mixture was moved into a 100 mL Teflon-lined autoclave and then heated at 140°C for 24 h. Finally, the product was collected by centrifuge and then dried in vacuum oven for 10 hrs. The OTAC used in this study not only aided in reducing the thickness of the nanosheets, but also played a crucial role in imparting photochromic properties by providing a source of chlorine that facilitated the rich crystalline-amorphous boundaries in the crystal structure, thereby boosting the photoactivity of the semiconductor photocatalyst p-BWO nanosheets.

MIL-125(Ti)-NH2 synthesis
The MIL-125(Ti)-NH2 was prepared according to Zhang et al. (2021) as follows: 0.797 mL titanium tetraisopropoxide (TTIP) and 0.651 g 2-aminoterephthalic acid (BDC-NH2) were dissolved in the mixture of dimethylformamide (DMF)/methanol (15 mL/15 mL). Then, the mixture was transferred to a Teflon-lined steel autoclave reactor and placed in an oven at 150°C for 15 h. The obtained yellow products were separated by centrifugation and washed by 30 mL DMF and 30 mL methanol, respectively, for three times. Finally, the samples were dried under 50°C for 12 hr in vacuum. To be brief, MIL-125(Ti)-NH2 was abbreviated as MIL hereafter.

p-BWO@MIL composite photocatalyst preparation
The composite photocatalyst p-BWO@MIL was prepared via a metal-ion adhesive. This synthesis method differs from the common hydrothermal method, in which a simple in situ selfassembly method was used but the stabilization of semiconductors on MOFs was still a challenge because the interfacial contact between MOF and semiconductor was not yet sufficiently strong. The synthesis process of the composite photocatalyst in our method was based on the electrostatic attraction that occurred on the interface of MOF and semiconductor, because two components and the adhesive had the opposite surface charges as demonstrated by their zeta potential values, (Barick et al., 2015;Lv et al., 2016;Zhu et al., 2023). 0.25 g poly(ethylenimine) (PEI, 50 wt%) as a metal-ion adhesive was dissolved in 10 mL DI water to form PEI/DI water solution (solution I) and then a certain amount of p-BWO nanosheets were dispersed in the solution I and then were stirred for 12 hours to form p-BWO/PEI water solution (solution II). After that, solution II was centrifuged to obtain p-BWO/PEI solid. The solid was washed with DI water two times to remove PEI residues and then dispersed in 15 mL DI water again to form p-BWO/PEI water solution (solution III). Subsequently, a certain amount of MIL was added and dispersed in solution III and stirred for 4 hrs. The final yellow product p-BWO@MIL was collected by centrifuge and then dried in a vacuum oven for 10 hrs. The composite photocatalysts are, hereafter, termed p-BWO@MIL-ratio (0.5 to 1.5 of p-BWO/MIL mass ratio), where the mass of MIL was kept at 0.1 g.

PE-MOFilter Fabrication
To fabricate the p-BWO@MIL coated electret filter media, PE-MOFilter, the following properties should be considered in the process. Firstly, the charges of the electret media should not be degraded; secondly, the MOF particles should firmly attach to the electret media with a minimized growth of air resistance; thirdly, the transfer process is simple and cost-efficient. Here, we chose Initial efficiency for 0.3 µm particles at 5 cm s -1 (%) MERV 13 13.1 ± 0.9 0.47 ± 0.02 75 ± 2 0.1 ± 0.003 ~50 4.5 ± 0.1 91.1 ± 0.2 the liquid filtration coating method. The experimental setup and parameters for the composite photocatalyst coating onto electret filter media can be found elsewhere . The PE-MOFilter was fabricated by using the 50 mm diameter flat sheets of a MERV 13 electret filter media for the deposition of the p-BWO@MIL particles for the PM2.5 and toluene removal tests. Table 1 summarizes the specifications of the MERV 13. The coating amount of the composite photocatalyst was controlled at 10 wt% (photocatalyst/media) based on Zhang et al. (2021). Two layers of MERV 13 flat sheets were used to enhance the removal efficiency of both PM2.5 and toluene. In total, three kinds of PE-MOFilters were fabricated and tested, which were MERV 13 electret filter media coated with p-BWO@MIL-0.5, p-BWO@MIL-1, and p-BWO@MIL-1.5 particles.

Characterization of MOF Particles and PE-MOFilters
The scanning electron microscopy (SEM, Hitachi SU-70, Hitachi Corp., Tokyo, Japan), X-ray diffraction (XRD, PANalytical X'Pert Pro, Malvern PANalytical Ltd., Malvern, UK), UV-Visible spectrophotometer (Evolution 220, ThermoFisher) and photoluminescence (PL) spectroscopy (QuantaMaster 400, Photon Technology International) were utilized to characterize the size and morphology, crystal structure, and optical properties, respectively, of the synthesized samples. Besides, nitrogen adsorption-desorption isotherms were collected using a gas sorption analyzer (Autosorb iQ, Quantachrome Instruments Corp., Boynton Beach, FL) at 77 K to characterize the pore size distribution and Brunauer-Emmett-Teller (BET) surface area. The SEM analysis was also conducted for the composite PE-MOFilters to evaluate if the p-BWO@MIL particles were coated uniformly in the filters.

PM and Toluene Removal Efficiency by the PE-MOFilter
After the chemical and physical characterizations of the p-BWO@MIL and its corresponding PE-MOFilter, the filters were evaluated for their PM and VOC removal efficiencies. By comparing the PM and toluene removal results of the PE-MOFilter with that of pristine MERV 13 and MIL, the effects of p-BWO@MIL coating on charge degradation and p-BWO coating on the reduction of toluene adsorption were determined.
The filter holder applied to conduct the experiments was a newly fabricated muti-functional holder, as shown in Fig. 1. It can be used not only in PM filtration but also VOC adsorption and VOC photoreaction tests. This holder had an inner diameter of 4 cm, length of 6 cm and volume of 75 cm 3 . It clamps and seals the filter media in the middle and has two quartz windows on both sides. In the PM filtration tests, 1 and 2 taps were used to measure the upstream and downstream particle concentration to determine the efficiency. The taps 3 and 4 were for the pressure drop measurement. For adsorption tests, all the taps were closed and then filter was challenged by toluene. For photoreaction, all the taps were also closed and the wide opened quartz windows on both sides allowed the majority of p-BWO@MIL being exposed to the light source. This new filter holder significantly made it convenient for performing all the tests in this study. The test methods for the PE-MOFilter against PM2.5 and toluene are discussed in the flowing and the test conditions are summarized in Table 2.

Initial PM removal efficiency test
The initial PM removal efficiency of PE-MOFilter was tested under 5 cm s -1 face velocity by using the same experimental setup as shown in Tang et al. (2018a). In brief, atomization (Model 9302, TSI Inc., Shoreview, MN) and classification (Model 3082, TSI Inc., Shoreview, MN) were utilized to produce monodisperse NaCl particles with sizes between 20-500 nm. The classified monodisperse particles were firstly neutralized to reach Boltzmann distribution to mimic the particles present in the ambient condition before challenging the PE-MOFilters. The upstream, dx)up, and downstream, C(dx)down, concentrations (particle cm -3 ) of particles with diameter of dx were measured by an ultrafine condensation particle counter (UCPC, Model 3776, TSI Inc. Shoreview, MN). Then the initial size-fractioned efficiency, ηPM, can be determined as: For comparison, the initial efficiency of the original and discharged (by IPA vapor) MERV 13 electret media without coating p-BWO@MIL particles was also measured.

Toluene initial efficiency and adsorption capacity test
Toluene (C6H5CH3) as a common and representative harmful indoor VOC Jafari et al., 2018;Kim et al., 2018aKim et al., , 2018bRezaei et al., 2016) was selected to challenge the PEb-MOFilters. The experimental system shown in Fig. 2 was used to find the toluene initial adsorption efficiency and adsorption capacity and photoreaction. The light source was not used in the adsorption process. The line of the dummy holder was used to generate the calibration curve between the toluene concentration, i.e., 0.05 to 50 ppm, and peak area, determined by a gas chromatography (GC, Agilent 7890B, Agilent Technologies Inc., Santa Clara, CA) equipped with a flame ionization detector (FID, Agilent Technologies Inc., Santa Clara, CA). 5 ppm of toluene with the 5 cm s -1 of face velocity was applied in the testing. When the adsorption started, the 3-way valve was switched to dummy line first to confirm if a correct toluene concentration was produced from the dilution of a 100 ppm toluene cylinder. When it was correct, the 5 ppm toluene flow was switched to the filter line and continued challenging the filter until breakthrough (90% Fig. 2. Experimental setup for initial removal efficiency and adsorption capacity test, and photocatalytic efficiency test for toluene. of 5 ppm in this study). The relative humidity (RH) was controlled to be 10% in the adsorption tests by adjusting the flow rates of the compressed air (~5% RH) and the water mists generated by the single jet atomizer (Model 9302, TSI Inc., Shoreview, MN). The equations to determine the toluene initial efficiency and adsorption capacity can be found elsewhere .

Photodegradation of Toluene by the PE-MOFilters
The photocatalytic degradation of gaseous toluene was performed under static flow condition at ambient temperature of 20°C, atmospheric pressure of 1 bar, and room RH (~50%). The static means both valves 1 and 2 were closed after the PE-MOFilter was saturated with the toluene in the breakthrough test. Meanwhile, the UV lights were irradiated the PE-MOFilter on both sides for a certain time for the photoreaction. The light sources were two 25W UV bulbs with a wavelength of 254 nm. They were placed against the two quartz windows to initiate the photocatalytic processes. The power density of the UV light in the reaction chamber was calculated to be 9.5 mW cm -2 . Be noted, the static test conducted here is very relevant to the realistic operation of filter in an air purifier. For example, a residential purifier is normally turned on in the evening and overnight when people are home. During the day, people are out for school and work so the blower of purifier can be off, and UV can be on to conduct photoadaptation for the adsorbed VOCs. Similarly, for the purifiers used in offices or schools, the photodegradation can be launched during the night.
To determine the efficiency of photodegradation for the PE-MOFilter under UV irradiation, this study developed a simple procedure. That was, we compared the toluene mass from the desorption with a 5 cm s -1 air flow blowing under three different conditions. The first condition, called "Blank", was to initiate the toluene adsorption right after the breakthrough (saturated with toluene). The second condition, called "Dark", was to switch off the room light to keep the holder in dark after saturation and then to initiate the toluene adsorption after the sample stayed in dark for 16 hours. The third condition, called "Photoreaction", was similar to the "Dark" condition except the holder was continued to be irradiated with UV lights for 16 hours.
Ideally, the results between Blank and Dark might be very close if there was no leakage. The efficiency of the photodegradation, ηPhoto, was calculated as: where Ct,Photo (mg m -3 ) is the concentration of toluene at the time t (min) in the desorption process of Photoreaction condition, Ct,Blank (mg m -3 ) is the concentration of toluene at the time t in the adsorption process for the Blank condition, Q is the total flowrate (mL min -1 ), mPhoto and mBlank are the toluene desorption mass under Photoreaction and Blank condition, respectively.  Fig. 3(a) demonstrates that MOF MIL-125-NH2 has a morphology of tetragonal plate, having an average length and thickness of ~900 and ~300 nm, respectively, which were in good agreement with literature (Kim et al., 2018a). The p-BWO nanosheets shown in Fig. 3 hundreds of nanometers long and comprised of multiple units of tiny nanoplates with a thickness of ~10 nm. The image of the photocatalyst p-BWO@MIL in Fig. 3(c) shows that the p-BWO nanoplates are well distributed on the surfaces of MOF MIL, which is not only beneficial to the diffusion of toluene from the internal pores of MOF to the surface of p-BWO, but also favorable for the interface charge transfer in the PCO process (He et al., 2019;He and Wang, 2018).

Characterization of Synthesized MIL, p-BWO and p-BWO@MIL-1 Particles
The XRD patterns were measured to investigate the crystallinity of the synthesized p-BWO, MIL and p-BWO@MIL particles. As shown in Fig. 3(d), the XRD characteristic peaks of p-BWO and MIL are clearly displayed in the composite p-BWO@MIL particles, indicating a two-phase composition of MIL and p-BWO (Zhang et al., 2020). The specific surface area and porosity of the synthesized samples were determined by BET analysis using N2 adsorption-desorption tests at 77 K. As shown in Fig. 3(e), the isotherm of MIL-125-NH2 exhibits a typical Type I curve with a sharp increase at low pressure regions, indicating its microporous structures and the corresponding high surface area. The isotherm of p-BWO shows a Type IV curve with a typical hysteresis loop due to the aggregations of p-BWO nanosheets (Zhang et al., 2020), which indicates the non-porous characteristics of p-BWO. This is consistent with its low pore volume and small surface area shown in Table 3. The isotherm of p-BWO@MIL exhibits hybrid Type I/IV, and the hysteresis loops occur at high-pressure zone (P/Po > 0.85), suggesting the presence of coexistence of micropore and mesopores. Table 3 shows the composite photocatalyst p-BWO@MIL shows high pore volume and surface area. As aforementioned, high surface area of the composite photocatalyst is beneficial to pre-concentrate a large amount of toluene molecules, and in turn, is facilitated to the photodegradation process. Meanwhile, it is well acknowledged that gas-solid photocatalytic reactions usually occur on the surfaces of catalysts, thus, it is expected that the photocatalytic efficiency would be greatly enhanced because of the increased photoreaction sites when compared to pure MOFs and pure semiconductor materials (He et al., 2019). Fig. 3(f) shows that the pore size distribution of the prepared samples. It is seen that the original MIL and p-BWO@MIL had the same peak pore size of 0.78 nm, a pore size which is beneficial for toluene adsorption (kinetic diameter, 0.58 nm) (Ma et al., 2019). However, pure p-BWO had the peak pore size of 3.16 nm, which is significantly larger than the toluene molecular size, and unfavorable for toluene capture. In Fig. 3(i), the drastically lowered photoluminescence intensity demonstrated the suppressed recombination of photogenerated carriers in p-BWO. It could be predicted that the synthesized p-BWO nanosheets have excellent photocatalytic activity because of the significantly enhanced charge carrier (h + , e -) separation in PCO process (Cao et al., 2018). The photocolouration response shown in the inset of Fig. 3(i) is a macroscopic manifestation of prompt charge carrier separation. The UV-vis spectra were evaluated to explore the optical properties of the samples. As shown in Fig. 3(g), the pure p-BWO exhibits light absorption (< 450 nm). After the incorporation of MOF MIL, the light absorption extends to the visible light range (< 600 nm), indicating the enhanced light absorption of the composite photocatalyst p-BWO@MIL. Tauc plots in Fig. 3(h) were derived from the UV-vis spectra to calculate the band gaps. As displayed in Fig. 3(h), the band gaps of p-BWO and MIL were determined to be 2.96 eV and 2.62 eV, respectively. It is obvious that the incorporation of MIL narrows down the band gap, which would boost the light utilization during the photocatalytic processes (He et al., 2019).

Performance of PE-MOFilters
The p-BWO@MIL were transported to MERV 13 electret filter media (~10 wt%) to form PE-MOFilters. Fig. 4 shows the SEM images of the depositions of p-BWO@MIL particles on the Table 3. BET analysis of p-BWO, MIL and p-BWO@MIL particles.

Adsorbents
BET surface area (m 2 g -1 ) Total pore volume (cm 3 g -1 ) Micro pore volume (cm 3 g -1 ) Pore diameter fibers of MERV 13 media. It is seen the p-BWO@MIL particles were uniformly deposited on individual fibers and in-depth of the PE-MOFilter (Zhang et al., , 2022.

Initial Efficiency of PE-MOFilters for PMs
To investigate the effects of p-BWO@MIL coating (10 wt%) on the PM removal of fabricated PE-MOFilters, the size-fractioned efficiencies of the original MERV 13 and p-BWO@MIL coated PE-MOFilters were compared in Fig. 5. The efficiency of discharged MERV 13 was also shown to evaluate the efficiency decline due to charge degradation of PE-MOFilters. As seen in Fig. 5, the decline of PM removal efficiency in all particle sizes was less than 10% compared to the original MERV 13. The substantial higher efficiency of the PE-MOFilters than the discharged MERV 13 indicates a significant retention of fiber charges after the coating of the p-BWO@MIL. It is noted that the PM removal efficiencies for MERV 13 electret media coated with p-BWO@MIL-0.5 and p-BWO@MIL-1.5 have the similar trend as p-BWO@MIL-1. Thus, they are not shown here. In addition, the results also showed that the coating of p-BWO@MIL caused a negligible increase of pressure drop. Therefore, the proposed PE-MOFilter here remains the merits of electret media, including high PM removal efficiency and low pressure drop.

Initial Adsorption Efficiency of PE-MOFilters for Toluene
In this section, the toluene removal performance was quantitatively compared amongst PE-MOFilters with different p-BWO to MIL ratios. Fig. 6 compares the initial toluene (5 ppm)  removal efficiency of the original MIL, pure p-BWO and p-BWO@MIL-0.5 (-1.5) coated on MERV 13 under 5 cm s -1 face velocity. It is seen that the initial removal efficiencies of the PE-MOFilters, 80-88%, had a comparable toluene removal performance with the original E-MOFilter of 92%. The efficiency slightly reduced with increasing ratio of p-BWO to MIL, which was in agreement with the BET surface area shown in Table 3. It is noted that the pure p-BWO nanosheets coated MERV 13 had a low toluene initial removal efficiency due to the extremely low surface area and large pore size as shown in Table 3 and Fig. 3(f). It is concluded that with the appropriate amount of p-BWO@MIL coating, there were negligible adverse effects of coating on the surface area and porous structure of MOF and toluene adsorption.  Fig. 7(a) refer to the feeding 5 ppm toluene and monitored downstream toluene concentration of the PE-MOFilter, respectively. The adsorption capacity from greatest to least was original MIL > p-BWO@MIL-0.5 > p-BWO@MIL-1 > p-BWO@MIL-1.5 > p-BWO. Notably, the PE-MOFilters exhibited only slight reduction of adsorption capacity compared to the original E-MOFilter, which was due to their retention of highly porous structure, while p-BWO coated media showed a very low adsorption capacity due to its low surface area. These results are consistent with the order of their corresponding surface area shown in Table 3.

Toluene Adsorption Capacity and Photodegradation of the PE-MOFilters
Figs. 7(b-f) compared the toluene desorption curves amongst Blank, Dark and Photoreaction conditions. The C0 in these figures was the first detected toluene desorption concentration from the conditions of Blank and Cx is the desorption concentration under Dark or Photoreaction conditions. As expected, the results between the Blank and Dark were very close indicating a good leakage free of the photoreactor chamber. It can be seen the order of toluene conversion rate is p-BWO@MIL-1 > p-BWO@MIL-1.5 > p-BWO@MIL-0.5 > p-BWO > MIL. The results indicate that all the composite photocatalysts exhibit higher efficiencies of photodegradation for toluene than the original MIL and p-BWO coated electret media, which can be attributed to the rapid charge transfer at the interface of p-BWO and MIL, as verified by the photoluminescence in previous studies (Xu et al., 2020;Yin et al., 2020aYin et al., , 2020b. In Fig. 7(b), the original E-MOFilter was demonstrated to have the lowest photocatalytic efficiency (3.5%), indicating the poor photocatalytic performance of the pure MIL. Meanwhile, the pure p-BWO coated electret media shows a toluene conversion rate of 25.3% (Fig. 7(c)), suggesting pure semiconductor p-BWO could not serve as an effective photocatalyst towards photodegradation of VOCs. This can be explained by its extremely small surface area and the consequently small number of photoreaction sites. In the contrary, p-BWO@MIL coated PE-MOFilters (shown in Figs. 7(d-f)) exhibited good photocatalytic performances, and the conversion of toluene increased from 45.0% to 68.7% and then reduced to 52.5% for p-BWO to MIL ratio from 0.5 to 1.0 and 1.5, respectively. Based on the results, it is concluded that the ratio of p-BWO to MOF plays a significant role in enhancing the photocatalytic efficiency of the composite photocatalyst. This is because the appropriate amount of MOF facilitates the interfacial charge transfer and suppresses the charge carrier recombination. However, if the amount of MOF exceeds the optimal value, the extra MOF would lower light absorption, diminish the density of excited charge carriers, and impair interfacial charge transfer in the PCO process, which would result in a reduction in the photodegradation efficiency (He et al., 2019). In contrast, semiconductor serves as a photoactive center to photodegrade toluene. If it is higher than ideal value, the micropore of MOFs would be partially blocked and thus the surface area would be reduced (Molavi et al., 2018), meanwhile, the interface charge transfer would also be restricted, causing a reduced photocatalytic efficiency.
To evaluate the structural stability of the prepared samples, SEM and XRD analysis of p-BWO@MIL-1 before and after photocatalytic tests were conducted. As shown in Fig. 8, the morphology and crystal structure remained unchanged after photoreaction process, demonstrating that the UV light will not cause damage to the MOF structure. Additionally, to demonstrate that the UV light will not degrade the charge of the electret filter, the filtration efficiencies of the original and UV light treated MERV 13 electret filter media were examined and compared. Results shown in Fig. 8 indicated that UV light treated MERV 13 electret media has a negligible filtration efficiency change compared with the original one.

CONCLUSIONS
In this study, a representative MOF (MIL-125-NH2) was successfully synthesized and coated with amorphous semiconductor p-BWO nanosheets to form a composite photocatalyst p-BWO@MIL-125-NH2 (p-BWO@MIL). Then the composite photocatalyst was coated to a MERV 13 grade electret filter media to develop a photocatalytic PE-MOFilter for a simultaneous removal of PM2.5 and VOCs. The composite photocatalysts were characterized by SEM, XRD, UV-vis, photoluminescence and BET analysis. The characterization results demonstrated the successful synthesis of the composite photocatalyst p-BWO@MIL, as the p-BWO nanosheets formed an intimate contact on the surface of MOF MIL-125-NH2. The PL spectrum of p-BWO demonstrated this composite photocatalyst a promising photocatalytic activity. The XRD analysis for the crystallinity of p-BWO@MIL indicated a two-phase composition of MIL and p-BWO. The BET surface area of the PE-MOFilter also remained high (1213 cm 2 g -1 ).
The newly fabricated PE-MOFilter was found to have an initial toluene adsorption efficiency of more than 80% and a comparable adsorption capacity to that of the MIL-125-NH2 based E-MOFilter. The PM removal efficiency was found to remain higher than 85% indicating the coating of p-BWO@MIL on the MERV 13 electret media had a negligible degradation of fiber charge based on 10 wt% of coating.
A simple method was developed to evaluate the photocatalytic performance of the PE-MOFilter. The method compares the toluene desorption mass of PE-MOFilter right after saturation to that of PE-MOFilter after 16 hours of UV illumination. It was found the p-BWO@MIL-1 based PE-MOFilter displayed the best photodegradation rate of about 70%.
To conclude, this study is the first to combine the semiconductor/MOF particles and coat them to electret filter media (denoted as PE-MOFilter) for the simultaneous removal of PM and VOC with low pressure drop. We found out the promising photocatalytic performance of the p-BWO@MIL towards toluene removal. The semiconductor/MOF particles could be regenerated by UV