Al2O3-Based Hollow Fiber Membranes Functionalized by Nitrogen-Doped Titanium Dioxide for Photocatalytic Degradation of Ammonia Gas

In recent years, reactive ammonia (NH3) has emerged as a major source of indoor air pollution. In this study, Al2O3-based hollow fiber membranes functionalized with nitrogen-doped titanium dioxide were produced and successfully applied for efficient heterogeneous photocatalytic NH3 gas degradation. Al2O3 hollow fiber membranes were prepared using the phase inversion process. A dip-coating technique was used to deposit titanium dioxide (TiO2) and nitrogen-doped titanium dioxide (N-TiO2) thin films on well-cleaned Al2O3-based hollow fiber membranes. All heterogeneous photocatalytic degradation tests of NH3 gas were performed with both UV and visible light irradiation at room temperature. The nitrogen doping effects on the NH3 heterogeneous photocatalytic degradation capacity of TiO2 were investigated, and the effect of the number of membranes (30, 36, 42, and 48 membranes) of the prototype lab-scale photocatalytic membrane reactor, with a modular design, on the performances in different light conditions was also elucidated. Moreover, under ultraviolet and visible light, the initial concentration of gaseous NH3 was reduced to zero after only fifteen minutes in a prototype lab-scale stage with a photocatalytic membrane reactor based on an N-TiO2 photocatalyst. The number of Al2O3-based hollow fiber membranes functionalized with N-TiO2 photocatalysts increases the capacity for NH3 heterogeneous photocatalytic degradation.


Introduction
Ammonia is an inorganic gas pollutant with the molecular formula NH 3 that is commonly found in low concentrations in the indoor environment. The environmental effects of odor emissions from livestock structures were also identified as critical issues [1]. For such low concentrations of NH 3 , adsorption and ventilation techniques are not costeffective [2]. Other alternative methods, such as oxidation processes combining ultraviolet irradiation with ozone [1], and a reactor consisting of a set of two-stage-in-series biotrickling filters, an influent gas supply system, and a liquid recirculation system [3], have recently been proposed. A new inexpensive approach is required to eliminate this pollutant and maintain a clean indoor environment. In order to make progress in this field, it is necessary to investigate a new photocatalytic membrane reactor and demonstrate that there is a substantial environmental benefit to be gained from reducing the gaseous NH 3 pollutant.
Titanium dioxide (TiO 2 ) and nitrogen-doped titanium dioxide (N-TiO 2 ) materials have gained prominence as the most researched semiconductor materials for photocatalytic purposes, including their use in devices for photocatalytic degradation of gaseous ammonia under various light sources [4][5][6][7][8][9][10][11]. Geng et al. in 2008 investigated the photocatalytic degradation of gaseous NH 3 by using nano-TiO 2 photocatalyst supported on latex paint film under UV irradiation [5]. From the kinetic point of view, it was found that the photocatalytic degradation of ammonia follows a pseudo-first order reaction [5][6][7][8]. Wang et al. reported that the removal percentage of NH 3 after about a 9 h photocatalytic reaction under visible light irradiation reached 53.1% using the optimal TiO 2 thin film doped with iron (III) [7]. After that, in 2013 Zendehzaban et al. coated TiO 2 on light expanded clay aggregate granules (LECA)-which is a porous and lightweight support-and found that more than 85% of NH 3 was photocatalytically removed within 300 min of the process under UV irradiation [8]. Li et al. reported in 2018 that a hierarchical-structured composite, ultrafine TiO 2 encapsulated in a nitrogen-doped porous carbon framework, can be used as a photocatalyst to degrade ammonia gas [10]. This is an interesting case in which the photocatalytic activity was excellent with 100% efficiency; it is represented by ultrafine TiO 2 encapsulated in a nitrogen-doped porous carbon framework [10]. The authors attributed the superior photocatalytic performance of this photocatalytic material to its large surface area and abundant pore structure [10].
However, despite various studies on heterogeneous photocatalytic NH 3 gas degradation through the use of different photocatalytic materials [1][2][3][4][5][6][7][8][9][10][11][12][13] and relatively few tentative attempts to investigate alternative ways to obtain a compact photocatalytic membrane reactor for indoor pollution abatement [11], to the best of our knowledge, there has been no case applied to the heterogeneous photocatalytic degradation process using Al 2 O 3 -based hollow fiber membranes functionalized by N-TiO 2 in a novel photocatalytic membrane reactor to completely reduce the NH 3 pollutant concentration.
In this study, Al 2 O 3 -based hollow fiber membranes functionalized by N-TiO 2 were investigated as an advanced photocatalyst for heterogeneous photocatalytic degradation of an NH 3 gas pollutant in a novel prototype lab-scale photocatalytic membrane reactor, with a modular design, based on 30, 36, 42, and 48 membranes under different light sources such as UV and visible light irradiation. This paper reveals for the first time, to the authors' knowledge, the potential of the developed prototype lab-scale photocatalytic membrane reactor with Al 2 O 3 -based hollow fiber membranes functionalized by N-TiO 2 as a promising and attractive candidate to remove NH 3 pollutants and then maintain a clean indoor environment. In addition, comparative studies of the NH 3 photocatalytic degradation processes were carried out using the undoped TiO 2 photocatalysts.

Materials
The materials and solvents used in the phase inversion process to prepare the Al 2 O 3 hollow fiber membrane and the dip-coating deposition of TiO 2 films were used exactly as received without further purification and are detailed in the Supplementary Material (See Table S1).

Preparation of Al 2 O 3 Hollow Fiber Membranes
Al 2 O 3 hollow fiber membranes were produced in two steps: (a) phase inversion spinning and (b) high-temperature sintering. In brief, a polymer solution including 33.5 wt% of NMP as a solvent and 6 wt% of PVP as a binder were mixed together. Al 2 O 3 powder was added into the polymer solution with 0.5 wt% of PVP as a dispersing agent. To determine the composition of the suspension, we used previous literature results [14,15]. In detail, the prepared suspension of Al 2 O 3 powder (60%) and additives (40%) was stirred (150 rpm) for one day before being degassed with a vacuum pump (IDP3, Varian, Palo Alto, CA, USA) for about one hour. The mixed and degassed suspension was extruded through an iron nozzle at a pressure of three bars. The air gap was set at 10 cm. Deionized water was used as a coagulant at room temperature. The phase transition process was completed within one day. Finally, Al 2 O 3 hollow fiber membranes were dried at 100 • C for 12 h before being sintered at 1300 • C for 3 h.

Preparation of Undoped TiO 2 and N-TiO 2 Photocatalyst Powders
The sol-gel method was used to produce the photocatalysts under consideration. As a titanium precursor, titanium (IV) isopropoxide (TTIP) was used. A total of 48 g of TTIP were mixed with 150 mL of deionized water and stirred for 1 h at room temperature using magnetic stirrers. A filtered white slurry was dried at 110 • C for 12 h to remove the residual water. Finally, the obtained photocatalyst powder was calcined for 3 h at 400 • C. Urea was used as a nitrogen source in the synthesis of N-TiO 2 [12]. In this case, 50 g of urea were previously dissolved in deionized water and stirred at room temperature for 1 h. Dropwise additions of TTIP were made, and then the solution was stirred for another hour. The calcination procedure used to prepare N-TiO 2 was the same as that used to prepare undoped TiO 2 . The powders appear to be different colors at first glance (see Figure S1 in the Supplementary Materials).

Preparation of Al 2 O 3 -Based Hollow Fiber Membranes Functionalized by Undoped TiO 2 and N-TiO 2 Photocatalysts
Undoped TiO 2 and N-TiO 2 photocatalysts were applied on Al 2 O 3 -based hollow fiber membranes by a dip-coating process based on tetraethyl orthosilicate (TEOS) solution as a silica-based binder [16]. In brief, 21 g of the photocatalyst under consideration were dispersed in 78 g of ethanol with 8 g of silica-based binder solution. The suspension was stirred for 1 h at room temperature to obtain a homogenous photocatalyst coating solution. The Al 2 O 3 -based hollow fiber membranes were cleaned in acetone under supersonic conditions and dried at 100 • C before immersing them in the photocatalyst coating solution for 1 h. Then, after this immersion time, the Al 2 O 3 -based hollow fiber membranes functionalized by undoped TiO 2 and N-TiO 2 photocatalysts were washed with common deionized water and dried at 150 • C for 12 h.

Photocatalytic Degradation of Gaseous Ammonia
A novel prototype lab-scale photocatalytic membrane reactor, with a modular design, was specifically designed to evaluate the NH 3 heterogeneous photocatalytic degradation capacity of the Al 2 O 3 -based hollow fiber membranes functionalized by undoped TiO 2 and N-TiO 2 photocatalysts in the application field of air-purifying filters and indoor pollution abatement. The reactor was designed to also include a brushless DC fan to push the gas toward the cylindrical photocatalytic membrane chamber. The acrylic chamber was designed to contain a different number of functionalized membranes in order to study the geometrical effects on the NH 3 heterogeneous photocatalytic degradation performances under different light sources such as UV and visible light irradiation. The modular photocatalytic membrane reactors were made of 30, 36, 42, and 48 Al 2 O 3 -based hollow fiber membranes functionalized by undoped TiO 2 and N-TiO 2 photocatalysts, with membrane surface areas of 8.430 × 10 −3 , 1.012 × 10 −2 , 1.180 × 10 −2 , and 1.349 × 10 −2 m 2 , respectively. The portable prototype weighs less than four hundred grams.
The photocatalytic membrane reactor was supplied with NH 3 gas (5%, He balanced) via 1/4 and 1/8 inch stainless steel tubes, with the flow rate controlled by a mass flow meter (MFC). The photocatalytic membrane reactor is described in the Supplementary Material (see Figure S2). The lab-scale photocatalytic membrane reactor was closed in an emetic black box to prevent light from entering. As UV and visible light sources, a UV lamp and a Xenon lamp were used.
There are two successive steps in the experimental protocol for measuring the photocatalytic degradation characteristics of gaseous ammonia. Figure 1a illustrates the first step. First, the gaseous gas mixture containing NH 3 was injected continuously through the experimental equipment until the stabilization of the process. The process was considered stabilized when two consecutive measurements of intensity peaks differed by less than one percent, according to gas chromatography (GC, iGC-200A, DS Science, Gwangju, Korea) measurements. The NH 3 sampling was repeated every five minutes for at least one hour. Figure 1 illustrates the second step (b). The second step is to switch the system from an open situation (see Figure 1a) to an internal circulation situation (see Figure 1b), in which the entire device changes into a batch-type reactor operating in an unsteady state. The lamp installed in the dark box was turned on at the same time as the change in circulation type. The NH 3 decomposition rate was calculated by dividing the initial concentration (C 0 ) by the concentration at time t (C t ).
pollution abatement. The reactor was designed to also include a brushless DC fan to push the gas toward the cylindrical photocatalytic membrane chamber. The acrylic chamber was designed to contain a different number of functionalized membranes in order to study the geometrical effects on the NH3 heterogeneous photocatalytic degradation performances under different light sources such as UV and visible light irradiation. The modular photocatalytic membrane reactors were made of 30, 36, 42, and 48 Al2O3-based hollow fiber membranes functionalized by undoped TiO2 and N-TiO2 photocatalysts, with membrane surface areas of 8.430 × 10 −3 , 1.012 × 10 −2 , 1.180 × 10 −2 , and 1.349 × 10 −2 m 2 , respectively. The portable prototype weighs less than four hundred grams.
The photocatalytic membrane reactor was supplied with NH3 gas (5%, He balanced) via 1/4 and 1/8 inch stainless steel tubes, with the flow rate controlled by a mass flow meter (MFC). The photocatalytic membrane reactor is described in the Supplementary Material (see Figure  S2). The lab-scale photocatalytic membrane reactor was closed in an emetic black box to prevent light from entering. As UV and visible light sources, a UV lamp and a Xenon lamp were used.
There are two successive steps in the experimental protocol for measuring the photocatalytic degradation characteristics of gaseous ammonia. Figure 1a illustrates the first step. First, the gaseous gas mixture containing NH3 was injected continuously through the experimental equipment until the stabilization of the process. The process was considered stabilized when two consecutive measurements of intensity peaks differed by less than one percent, according to gas chromatography (GC, iGC-200A, DS Science, Gwangju, Korea) measurements. The NH3 sampling was repeated every five minutes for at least one hour. Figure 1 illustrates the second step (b). The second step is to switch the system from an open situation (see Figure 1a) to an internal circulation situation (see Figure 1b), in which the entire device changes into a batchtype reactor operating in an unsteady state. The lamp installed in the dark box was turned on at the same time as the change in circulation type. The NH3 decomposition rate was calculated by dividing the initial concentration (C0) by the concentration at time t (Ct).

Characterization of Al 2 O 3 Hollow Fiber Membranes
The Al 2 O 3 hollow fiber membranes had a 0.3 mm thick wall. The cross-sections obtained by fracturing the Al 2 O 3 hollow fiber membrane after the high-temperature sintering process (1300 • C; 3 h) are shown in Figure S3 in the Supplementary Material. There are two types of pore structure: finger-like pores in the inner edge and sponge-like pores in the outer edge. These findings are consistent with previous research [14,15].

Characterization of Undoped TiO 2 and N-TiO 2 Photocatalyst Powders
The XRD patterns of undoped TiO 2 and N-TiO 2 photocatalyst powders are shown in Figure 2. The XRD patterns in both cases show diffraction peaks that are well indexed to a tetragonal crystal structure (anatase TiO 2 structure) as described in JCPDS card No. 21-1272. In addition, the crystallite size of the undoped TiO 2 and N-TiO 2 photocatalyst powders was calculated using the Debye-Scherrer equation [17].

Characterization of Al2O3 Hollow Fiber Membranes
The Al2O3 hollow fiber membranes had a 0.3 mm thick wall. The cross-sections obtained by fracturing the Al2O3 hollow fiber membrane after the high-temperature sintering process (1300 °C; 3 h) are shown in Figure S3 in the Supplementary Material. There are two types of pore structure: finger-like pores in the inner edge and sponge-like pores in the outer edge. These findings are consistent with previous research [14,15].

Characterization of Undoped TiO2 and N-TiO2 Photocatalyst Powders
The XRD patterns of undoped TiO2 and N-TiO2 photocatalyst powders are shown in Figure 2. The XRD patterns in both cases show diffraction peaks that are well indexed to a tetragonal crystal structure (anatase TiO2 structure) as described in JCPDS card No. 21-1272. In addition, the crystallite size of the undoped TiO2 and N-TiO2 photocatalyst powders was calculated using the Debye-Scherrer equation [17]. Taking into account the first most intense peak, the average crystallite size of the undoped TiO2 photocatalyst powders obtained from the Scherrer formula is 6.9 nm. The average crystallite size of N-TiO2 is 7.1 nm. With the N dopant, the average crystallite size of the TiO2 particles increased, in apparent disaccord with expectations [17].
TEM was used to examine the microstructures of the undoped TiO2 and N-TiO2 photocatalyst powders. The undoped TiO2 was structured as monodispersed spherical TiO2 particles, as shown in Figure 3a. According to the low magnification TEM image ( Figure  3b,c), the average diameter of the individual spherical TiO2 particles was around 10 nm, which is consistent with previous studies [12,18]. The high-resolution TEM images in  Taking into account the first most intense peak, the average crystallite size of the undoped TiO 2 photocatalyst powders obtained from the Scherrer formula is 6.9 nm. The average crystallite size of N-TiO 2 is 7.1 nm. With the N dopant, the average crystallite size of the TiO 2 particles increased, in apparent disaccord with expectations [17].
TEM was used to examine the microstructures of the undoped TiO 2 and N-TiO 2 photocatalyst powders. The undoped TiO 2 was structured as monodispersed spherical TiO 2 particles, as shown in Figure 3a. According to the low magnification TEM image (Figure 3b,c), the average diameter of the individual spherical TiO 2 particles was around 10 nm, which is consistent with previous studies [12,18]. The high-resolution TEM images in Figure 3e,f show that spherical TiO 2 particles are mixed with polygonal particles in the N-TiO 2 photocatalyst powder. The size of the N-TiO 2 nanoparticles is shown in Figure 3f.
The XPS data of undoped TiO 2 and N-TiO 2 powders are shown in Figure 4, which allows us to determine the oxidation states of various elements present in photocatalysts. It is important to note that the N 1s peak in the 390-410 eV range is correctly present only in the N-TiO 2 photocatalyst sample. The N 1s spectrum of the N-TiO 2 photocatalyst, in particular, has a binding energy peak around 399 eV. According to Yalçın et al. [18], the peak represents anionic N substitutionally incorporated in TiO 2 through O-Ti-N linkages. Furthermore, it can be seen that the O 1s peak decreased with the N-doping of the TiO 2 photocatalyst.  The XPS data of undoped TiO2 and N-TiO2 powders are shown in Figure 4, which allows us to determine the oxidation states of various elements present in photocatalysts. It is important to note that the N1s peak in the 390-410 eV range is correctly present only in the N-TiO2 photocatalyst sample. The N1s spectrum of the N-TiO2 photocatalyst, in particular, has a binding energy peak around 399 eV. According to Yalçın et al. [18], the peak represents anionic N substitutionally incorporated in TiO2 through O-Ti-N linkages. Furthermore, it can be seen that the O1s peak decreased with the N-doping of the TiO2 photocatalyst.  The XPS data of undoped TiO2 and N-TiO2 powders are shown in Figure 4, which allows us to determine the oxidation states of various elements present in photocatalysts. It is important to note that the N1s peak in the 390-410 eV range is correctly present only in the N-TiO2 photocatalyst sample. The N1s spectrum of the N-TiO2 photocatalyst, in particular, has a binding energy peak around 399 eV. According to Yalçın et al. [18], the peak represents anionic N substitutionally incorporated in TiO2 through O-Ti-N linkages. Furthermore, it can be seen that the O1s peak decreased with the N-doping of the TiO2 photocatalyst.   Table 1 summarizes the band gap energy, element analysis (at. percent), and BET surface area (m 2 g −1 ) results. The photocatalyst's BET surface area was determined using nitrogen adsorption methods. After N-doping, the apparent BET surface area of undoped TiO2 increased from 50 ± 2 to 57 ± 2 m 2 g −1 [19]. These findings agree completely with those of Bakre et al. [12], where the BET surface area of the N-TiO2 sample prepared with urea as a nitrogen source was approximately 42 m 2 g −1 .  Table 1 summarizes the band gap energy, element analysis (at. percent), and BET surface area (m 2 g −1 ) results. The photocatalyst's BET surface area was determined using nitrogen adsorption methods. After N-doping, the apparent BET surface area of undoped TiO 2 increased from 50 ± 2 to 57 ± 2 m 2 g −1 [19]. These findings agree completely with those of Bakre et al. [12], where the BET surface area of the N-TiO 2 sample prepared with urea as a nitrogen source was approximately 42 m 2 g −1 . The diffuse reflectance spectrophotometer was used to study the absorbance properties of undoped TiO 2 and N-TiO 2 photocatalyst powders, and the results are shown in Figure 5a. Both photocatalysts' absorption spectra show strong absorption in the UV region, with the absorption edge at about 400 nm. When compared to the undoped TiO 2 photocatalyst, the absorption band of the N-TiO 2 photocatalyst is significantly red-shifted. Different authors attribute the observed redshift in the absorption edge to luminescence from localized surface states caused by the recombination of photogenerated electron-hole pairs [12,20,21]. Additionally, the bandgaps are presented in Figure 5b. The bandgap energies for undoped TiO 2 and N-TiO 2 photocatalyst were 3.17 eV and 2.36 eV, respectively. The bandgap energies obtained in this work show that, as expected, doping TiO 2 with N resulted in a decrease in the bandgap.  The samples' photoluminescence (PL) emission was also measured in order to better understand the behavior of holes and light-generated electrons. Figure 6 shows the PL spectra of undoped TiO 2 and N-TiO 2 photocatalyst powders at the wavelengths ranging from 400 to 800 nm. The emission spectra of the two photocatalysts had similar shapes. The emission peaks for undoped TiO 2 and N-TiO 2 photocatalyst powders were~520 nm and 590 nm, respectively. In detail, for undoped and N-doped TiO 2 , the PL spectra showed a red band at about 590 nm and a green band at around 520 nm, respectively. Defects in N-doped TiO 2 samples can be assigned to the green PL band [22]. The intensity of the green band decreases with the N doping sample when compared to undoped TIO 2 . Furthermore, the PL intensity of the N-TiO 2 photocatalyst is lower than that of the undoped TiO 2 sample, indicating that these samples have improved charge transfer and effective separation of electron-hole pairs. The PL intensity is known to be directly related to the electron-hole recombination rate, and a low electron-hole recombination rate results in a low PL intensity and then a high photocatalytic performance [23]. Finally, a weaker feature at a wavelength of about 610 nm is usually ascribed to oxygen vacancies and defects on the N doping sample [24,25].

Characterization of Al2O3-Based Hollow Fiber Membranes Functionalized by Undoped TiO2 and N-TiO2 Photocatalysts
SEM analysis detected the photocatalyst when the surfaces of the hollow fiber membrane before and after TiO2 photocatalytic coating are compared (see Supplementary Material, Figure S4). Furthermore, because a form such as a low-thickness coating layer does not appear in the SEM cross-sections of the hollow fiber membrane, the amount of photocatalyst coated on the surface of the hollow fiber membrane is presumed small (see Supplementary Material, Figure S5). The EDS mapping image and EDS line scanning analysis, on the other hand, confirmed the presence of photocatalysts effectively deposited on the Al2O3-based hollow fiber membrane surfaces. Images of EDS mapping of the surface of Al2O3-based hollow fiber membranes functionalized by N-TiO2 photocatalysts are included in the Supplementary Material (see Figure S6). The EDS line scanning analysis of Al2O3-based hollow fiber membranes functionalized by N-TiO2 photocatalysts is shown

Characterization of Al 2 O 3 -Based Hollow Fiber Membranes Functionalized by Undoped TiO 2 and N-TiO 2 Photocatalysts
SEM analysis detected the photocatalyst when the surfaces of the hollow fiber membrane before and after TiO 2 photocatalytic coating are compared (see Supplementary Material, Figure S4). Furthermore, because a form such as a low-thickness coating layer does not appear in the SEM cross-sections of the hollow fiber membrane, the amount of photocatalyst coated on the surface of the hollow fiber membrane is presumed small (see Supplementary Material, Figure S5). The EDS mapping image and EDS line scanning analysis, on the other hand, confirmed the presence of photocatalysts effectively deposited on the Al 2 O 3 -based hollow fiber membrane surfaces. Images of EDS mapping of the surface of Al 2 O 3 -based hollow fiber membranes functionalized by N-TiO 2 photocatalysts are included in the Supplementary Material (see Figure S6). The EDS line scanning analysis of Al 2 O 3 -based hollow fiber membranes functionalized by N-TiO 2 photocatalysts is shown in Figure 7. According to the EDS line scanning analysis in Figure 7, the thickness of the N-TiO 2 photocatalyst coating is in the range of about 1-2 µm.

Photocatalytic Degradation Test of Gaseous Ammonia
The NH3 heterogeneous photocatalytic degradation capacity of Al2O3-based hollow fiber membranes functionalized by undoped TiO2 and N-TiO2 photocatalysts was studied under different light conditions and photocatalytic membrane reactor configurations at room temperature.
Under ultraviolet light, Figure 8a illustrates the NH3 heterogeneous photocatalytic

Photocatalytic Degradation Test of Gaseous Ammonia
The NH 3 heterogeneous photocatalytic degradation capacity of Al 2 O 3 -based hollow fiber membranes functionalized by undoped TiO 2 and N-TiO 2 photocatalysts was studied under different light conditions and photocatalytic membrane reactor configurations at room temperature.
Under ultraviolet light, Figure 8a illustrates the NH 3 heterogeneous photocatalytic degradation capacity of a photocatalytic membrane reactor consisting of 48 Al 2 O 3 -based hollow fiber membranes functionalized by undoped TiO 2 and N-TiO 2 photocatalysts. Under ultraviolet light, both photocatalytic membrane reactors demonstrated good photocatalytic performance, but the initial concentration of gaseous NH 3 was reduced to zero after only fifteen minutes in the case of Al 2 O 3 -based hollow fiber membranes functionalized by N-TiO 2 photocatalysts. By contrast, in the case of undoped TiO 2 photocatalysts, the gaseous NH 3 completely disappeared after about twenty minutes of ultraviolet light.

Photocatalytic Degradation Test of Gaseous Ammonia
The NH3 heterogeneous photocatalytic degradation capacity of Al2O3-based hollow fiber membranes functionalized by undoped TiO2 and N-TiO2 photocatalysts was studied under different light conditions and photocatalytic membrane reactor configurations at room temperature.
Under ultraviolet light, Figure 8a illustrates the NH3 heterogeneous photocatalytic degradation capacity of a photocatalytic membrane reactor consisting of 48 Al2O3-based hollow fiber membranes functionalized by undoped TiO2 and N-TiO2 photocatalysts. Under ultraviolet light, both photocatalytic membrane reactors demonstrated good photocatalytic performance, but the initial concentration of gaseous NH3 was reduced to zero after only fifteen minutes in the case of Al2O3-based hollow fiber membranes functionalized by N-TiO2 photocatalysts. By contrast, in the case of undoped TiO2 photocatalysts, the gaseous NH3 completely disappeared after about twenty minutes of ultraviolet light.  The behavior of the Al 2 O 3 -based hollow fiber membranes functionalized by undoped TiO 2 and N-TiO 2 photocatalysts under visible light was studied using the same configuration with 48 membranes, and the results are shown in Figure 8b. Under visible light conditions, the photocatalytic membrane reactor made of 48 Al 2 O 3 -based hollow fiber membranes functionalized by N-TiO 2 photocatalysts performed well. The photocatalytic membrane reactor with Al 2 O 3 -based hollow fiber membranes functionalized by undoped TiO 2 photocatalyst had a maximum NH 3 heterogeneous photocatalytic degradation capacity of 30% after 30 min.
In this work, different configurations were also investigated in order to reduce the number of Al 2 O 3 -based hollow fiber membranes functionalized by N-TiO 2 photocatalysts. Three new prototype lab-scale photocatalytic membrane reactors with 30, 36, and 42 membranes were developed in our laboratory (see Supplementary Material, Figure S2). Furthermore, due to the high intensity of the Xenon lamp used in the above setup test case, as well as the increased temperature when used for an extended period of time, a small LED light source (white, blue, and ultraviolet) was installed in the middle of the photocatalytic membrane reactor to be more suitable and compact. By combining three LED lamps in a triangular shape, an internally mounted LED system with a total of nine lamps was developed (see Supplementary Material, Figure S7). It should be noted that the ideal compact and portable photocatalytic membrane reactor configuration would be to have the light source integrated into the photocatalytic system. This new compact configuration, which places the light source in the middle of the photocatalytic membrane reactor rather than on the outside, was tested for heterogeneous photocatalytic degradation of NH 3 gas at room temperature (see Supplementary Material, Figure S8).
The capacity of an NH 3 heterogeneous photocatalytic membrane reactor consisting of 36 Al 2 O 3 -based hollow fiber membranes functionalized by undoped TiO 2 and N-TiO 2 photocatalysts under an LED light source (white, blue, and ultraviolet) was investigated with the goals of (1) discovering relationships between NH 3 heterogeneous photocatalytic performance and photocatalysts and then (2) determining the optimum operational conditions concerning photocatalysts. The results of this experiment are shown in Figure S9 (see Supplementary Material). As could be expected, under LED light sources, the photocatalytic membrane reactor made of 36 Al 2 O 3 -based hollow fiber membranes functionalized by undoped TiO 2 photocatalyst always showed lower performance than the N-TiO 2 photocatalyst (see Figure S9). Based on the experimental results obtained with the photocatalytic membrane reactor made with Al 2 O 3 -based hollow fiber membranes functionalized by undoped TiO 2 and N-TiO 2 photocatalysts under ultraviolet-visible light (see Figure 8) and LED light sources (see Figure S9), the N-TiO 2 photocatalysts were used in the following section for the fabrication of different NH 3 heterogeneous photocatalytic membrane reactors.
The NH 3 heterogeneous photocatalytic degradation capacity of (a) 30, (b) 36, and (c) 42 Al 2 O 3 -based hollow fiber membranes functionalized by N-TiO 2 photocatalysts under LED light sources (white, blue, and ultraviolet) is shown in Figure 9. As expected, the number of Al 2 O 3 -based hollow fiber membranes functionalized by N-TiO 2 photocatalysts increases the NH 3 heterogeneous photocatalytic degradation capacity. In all cases, the NH 3 heterogeneous photocatalytic degradation capacity under a white LED light source was only slightly lower than that under a blue or ultraviolet light source.
The prototype lab-scale photocatalytic membrane reactor made up of 42 Al 2 O 3 -based hollow fiber membranes functionalized by N-TiO 2 photocatalysts can successfully reduce the gaseous NH 3 in the indoor environment to zero in a short amount of time (fifteen minutes) while maintaining the same performance level.
To the best of our knowledge, this is the first report on using Al 2 O 3 -based hollow fiber membranes functionalized by N-TiO 2 photocatalysts to develop a novel compact photocatalytic membrane reactor to completely reduce the NH 3 pollutant concentration, and its excellent performance is expected to expand our understanding of indoor air pollution.
In conclusion, the heterogeneous photocatalytic degradation of ammonia gas pollutant in a compact and portable photocatalytic membrane reactor with Al 2 O 3 -based hollow fiber membranes functionalized by N-TiO 2 photocatalysts provides an inexpensive alternative not only to conventional technologies, such as thermal and catalytic oxidation processes, condensation, adsorption on solids, scrubbing and biofiltration [26], but also to alternative methods [1,2]. The prototype lab-scale photocatalytic membrane reactor made up of 42 Al2O3-based hollow fiber membranes functionalized by N-TiO2 photocatalysts can successfully reduce the gaseous NH3 in the indoor environment to zero in a short amount of time (fifteen minutes) while maintaining the same performance level.
To the best of our knowledge, this is the first report on using Al2O3-based hollow fiber membranes functionalized by N-TiO2 photocatalysts to develop a novel compact photocatalytic membrane reactor to completely reduce the NH3 pollutant concentration, and its excellent performance is expected to expand our understanding of indoor air pollution.
In conclusion, the heterogeneous photocatalytic degradation of ammonia gas pollutant in a compact and portable photocatalytic membrane reactor with Al2O3-based hollow fiber membranes functionalized by N-TiO2 photocatalysts provides an inexpensive alternative not only to conventional technologies, such as thermal and catalytic oxidation processes, condensation, adsorption on solids, scrubbing and biofiltration [26], but also to alternative methods [1,2].

Conclusions
Previous research on new processes to eliminate the gaseous NH 3 pollutant from indoor environments has highlighted the importance of discovering new potentials in the heterogeneous photocatalytic degradation of NH 3 gas. Our findings demonstrate that Al 2 O 3 -based hollow fiber membranes functionalized with nitrogen-doped titanium dioxide (N-TiO 2 ) can be used successfully for heterogeneous photocatalytic NH 3 gas degradation.
In summary, Al 2 O 3 -based hollow fiber membranes functionalized by N-TiO 2 were investigated as an advanced photocatalyst for heterogeneous photocatalytic degradation of the NH 3 gas pollutant in a novel prototype lab-scale photocatalytic membrane reactor, with a modular design, based on 30, 36, 42, and 48 membranes exposed to UV and visible light irradiation. The NH 3 gas degradation increases with increasing Al 2 O 3 -based hollow fiber membranes functionalized by N-TiO 2 , according to the heterogeneous photocatalytic test results. Under optimal conditions, the NH 3 gas degradation efficiency reached 100% after only fifteen minutes.
Overall, the results point to a promising future for NH 3 heterogeneous photocatalytic degradation in compact and portable photocatalytic membrane reactors with Al 2 O 3 -based hollow fiber membranes functionalized by N-TiO 2 photocatalysts.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/membranes12070693/s1, Table S1: Chemicals used in the phase inversion process to prepare the Al2O3 hollow fiber membrane and the dip-coating deposition of TiO 2 films. Figure S1: Digital photographs of the prepared (a) undoped TiO 2 and (b) N-TiO 2 photocatalysts; Figure S2: (a) Photocatalytic membrane reactor (below) with the brushless DC fan (above), (b) NH 3 flux through the photocatalytic membrane reactor, (c) disassembled photocatalytic membrane reactor, assembled photocatalytic membrane reactor, (e) without light photocatalytic membrane reactor, (f) with light photocatalytic membrane reactor, and photocatalytic membrane reactor based on