Quinoline (C₉H₇N, MW = 129) was acquired from Sinopharm, China, and used without further purification. TiO₂ (P25; ca. 80% anatase and 20% rutile) was purchased from Deguessa, Germany, and had a BET surface area of 55 m²/g. Anhydrous alcohol was obtained from Sigma. The natural pH of the synthetic wastewater containing quinoline was 6.7. Prior to photodegradation, the pH of the wastewater was adjusted using a small amount of either H₂SO₄ or NaOH, and was monitored by a digital pH meter (WTW Multi 340i, Germany). All chemicals used were reagent grade. The distilled water used to prepare all sample solutions was prepared in our laboratory.

Degussa P25 TiO₂ was immobilized on porous Ni (denoted as TiO₂/Ni) via a facile dip-coating method without the use of coating equipment, as reported in our previous study [12]. P25 TiO₂ was added to anhydrous alcohol [13], and then dispersed for 10 min under ultrasonic agitation to confect the P25 sol to 10% concentration.

Foam Ni with a specific surface area of 0.25 m²/g was used as a support because of its homogeneous vesicular structure and facile machining properties, which are beneficial to increasing the contact of photocatalysts with UV rays [14,15]. Pre-dried foam Ni was cut into tubes of 40 cm × Ø 7 cm, and then immersed into the P25 sol for 20 s. The tubes were taken out and air dried at room temperature. After ethanol volatilization overnight, the TiO₂/Ni film was calcined at 90 °C for 60 min, and then at 450 °C for 30 min at a heating rate of 3 °C/min in a muffle furnace. The TiO₂/Ni film obtained was washed with boiling water to remove uncoated TiO₂ particles. The surface morphology of the prepared TiO₂/Ni film and bare porous Ni substrate were observed using a field emission scanning electron microscopy (FESEM) system (XL-30 ESEM FEG, Philips). The crystal phase structures were measured by X-ray diffraction (XRD) on a Rigaku D/max-3c diffractometer (CuKα radiation, λ = 0.15405 nm). The specific surface area of the prepared TiO₂/Ni film was 8.5 m²/g, as determined by nitrogen adsorption-desorption measurements (TriStar 3000).

An internal airlift loop photocatalytic reactor made of Pyrex glass was used for the photocatalytic experiments, as shown in Figure 1. The effective volume of the reactor was 3.5 L, and the dimensions were 55 cm × Ø 10 cm. A UV lamp (15 W, 254 nm) was longitudinally fixed in the center of the reactor. The prepared tubular TiO₂/Ni film was placed on a fixed support outside the UV lamp, resulting in a 2.5 cm gap. A microporous aerator was fixed below the reactor to maintain aerobic conditions inside the TiO₂/Ni tube. Gas bubbled up from the bottom and accelerated around the liquid. The liquid inside the TiO₂/Ni tube flowed in two directions: (i) some liquid gradually flowed out through the tube pores, and (ii) the others flowed up to the top of the tube. Consequently, the liquid outside the tube flowed down from the top; (iii) due to the different internal and external flow rates, resulting in water circulation in the photocatalytic reactor. This process benefited the enhancement of the contacting probability between the water contaminant and the P25 TiO₂ on the surface of porous Ni. The experimental temperature of the entire photoreactor was controlled by a surrounding external water-cooling jacket.

In the photocatalytic process, 3.5 L of aqueous quinoline solution was poured into the reactor. The UV lamp was switched on to start the photo-oxidation process. Compressed gas was bubbled upward through the gas distributor at the bottom of the reactor. At pre-defined irradiation times, 50 μL of solution was collected and immediately analyzed to avoid further reaction. The dependences of the quinoline degradation rate on initial concentration, light intensity, reaction temperature, aeration, and pH were investigated. Each experiment was conducted three times under identical conditions. The data presented in the text and figures are the mean values. The light absorption characteristics of the solutions were recorded using Persee T6 UV/visible spectrophotometer at a distance of 100 cm from the light source per the manufacturer’s instructions [16]. The concentration of quinoline was analyzed using a high-performance liquid chromatography system (LC-20A). The chromatographic column was a Shim-pack VP-ODS-C18 (4.6 mm × 150 mm, 5 μm). The mobile phase was a mixture of methanol and water (60:40, v/v) with a flow rate of 1 mL/min. The testing wavelength was 275 nm.

To obtain the intermediate products of quinoline degradation, the experiment was performed in a photocatalytic reactor with 100 mg/L quinoline under ambient conditions. After 30 min of irradiation time, 200 mL of the effluent was collected in a separating funnel, and then extracted for qualitative analysis by gas chromatography (GC)/mass spectroscopy (MS). The effluent for the GC/MS sample was further assayed by the method of Li et al. [17], except that 30 mL of anhydrous ether was used as the extractant. The extracts were decanted into a rotary evaporator, and then dried at 35 °C to evaporate the ether. The residue was dissolved in 10 mL of anhydrous ether for the GC/MS analysis.

GC/MS analysis was performed on an Agilent GC system (6890N/MSD 5973) equipped with an HP-5MS fused silica capillary column (30 m × 0.25 mm i.d., 0.25 μm film thickness). The column was first heated to 80 °C, kept constant for 3 min, linearly programmed to increase to a final temperature of 300 °C at a rate of 10 °C/min, and then isothermally held for 2 min. The temperature for the MS ion source was 200 °C, and the electron energy was 70 eV.

The FESEM images of porous Ni and the TiO₂/Ni film are shown in Figure 2. The surface of bare porous Ni without immobilized P25 was smooth (Figure 2a). After P25 TiO₂ coating, the originally smooth surface of the coated TiO₂/Ni film (Figure 2b) became relatively rough and had a number of microcracks. Then, the specific surface area increased remarkably from 0.25 m²/g to 8.5 m²/g. Consequently, a high photocatalytic efficiency was obtained because there were more interfaces between the catalyst and reactants.

Figure 3 shows the XRD patterns of P25 TiO₂ powder, TiO₂/Ni film, and porous Ni. Three strong diffraction peaks found at 2θ = 44.5°, 51.8°, and 76.1° were ascribed to the (111), (200), and (220) planes of Ni (ICDD PDF Card 87-0712) (JCPDS. 2001), respectively, in accordance with earlier studies [18,19].

Two diffraction peaks at 2θ = 37.2° and 68.3° were attributed to rutile phase (ICDD PDF Card 21-1276) [20]. The presence of other crystallographic plains labeled with “A” belongs to the anatase phase (ICDD PDF Card 21-1272) [20]. P25 TiO₂ coated on porous Ni caused no apparent change in the half width of the anatase (101) XRD peak, regardless of the heating treatment, and the crystallite size was calculated to be 21 nm from the anatase (101) peak width using the Scherrer equation. This result indicated that nano-sized TiO₂ particles in P25 sol have been reproduced onto the surface of porous Ni. Therefore, the TiO₂/Ni film possessed good photocatalytic activity.

To analyze the degradation process of quinoline, a UV absorption experiment was performed on the reactants, as described in the Materials and Methods Section. With increased illumination time, a series of UV absorption spectra was recorded within the 180–380 nm range, as shown in Figure 4a. The characteristic absorption of quinoline significantly decreased with increased irradiation time, and almost disappeared at 60 min. The UV₂₅₄ of the reaction solution initially increased and then gradually decreased. The peak appeared at 15 min (Figure 4b). This finding indicated that the aromatic or pyridine ring of quinoline opened in the initial stage, and formed a series of intermediate products containing conjugated double bond. As a result, UV₂₅₄ absorbance was enhanced. UV absorbance within the range 180–240 nm also gradually decreased with prolonged irradiation time. This result indicated the further mineralization of aliphatic saturated as well as cyclic hydrocarbons and their derivatives in solution [21]. Interestingly, UV absorbance above 320 nm initially increased, reached the peak at 45 min, and then apparently decreased. These findings were in accordance with the color of the reaction solution changing from transparent to pale yellow, to yellow-green, and then back to transparent.

The synthetic wastewater containing quinoline varied in color, as described in Section 3.2, indicating that there existed some intermediate products in the photocatalytic process. Based on the GC/MS analysis performed, the effluent of the synthetic wastewater after irradiation for 30 min contained 2-pyridinecarboxaldehyde, 3-pyridinecarboxaldehyde, quinoline, and 2(1H)-quinolinone. Figure 12 shows the possible mechanisms for the formation of the three intermediate products. The formation of the first two intermediates may have depended on the electrophilic substitution between quinoline and hydroxyl radicals. The initial oxidation mainly occurred at positions 5 and 8 of the aromatic ring, resulting in the formation of 5,8-quinolinedione. Further oxidation by active radicals resulted in the formation of 2-pyridinecarboxaldehyde or 3-pyridinecarboxaldehyde based on the phenyl ring opening process via quinoline-5,8-dione. This mechanism was in accordance with that reported by Thomsen [33] and Zhu et al. [34]. Considering that the electron cloud density of the pyridine ring is lower than that of an aromatic ring, the pyridine ring is relatively difficult to be oxidized. Hence, the formation of 2(1H)-quinolinone is related to the superoxide ion (O₂⁻) generated by dissolved oxygen on the TiO₂ surface trapped electron under UV irradiation. Ishibashi et al. [35] have reported that the superoxide ion (O₂⁻) is generated at the water-TiO₂ interface with a longer lifetime of about 70 s. Its absolute number pre-illuminated is calculated to be 2 × 10¹⁴/cm², and it has a low mobility [36]. During the adsorption of quinoline onto the TiO₂ surface, the initial addition reaction between quinoline and the superoxide ion (O₂⁻) may take place at the C-2 position of the pyridine ring, similar with an ordinary nucleophile [34,37]. This phenomenon benefited the further reaction with active radicals to generate 2(¹H)-quinolinone.