Efficient Day-and-Night NO2 Abatement by Polyaniline/TiO2 Nanocomposites

Finding innovative and highly performing approaches for NOx degradation represents a key challenge to enhance the air quality of our environment. In this study, the high efficiency of PANI/TiO2 nanostructures in the NO2 abatement both in the dark and under light irradiation is demonstrated for the first time. Heterostructures were synthesized by a “green” method and their composition, structure, morphology and oxidation state were investigated by a combination of characterization techniques. The results show that the unique PANI structure promotes two mechanisms for the NO2 abatement in the dark (adsorption on the polymeric chains and chemical reduction to NO), whereas the photocatalytic behavior prevails under light irradiation, leading to the complete NOx degradation. The best-performing materials were subjected to recycling tests, thereby showing high stability without any significant activity loss. Overall, the presented material can represent an innovative and efficient night-and-day solution for NOx abatement.


Introduction
Air pollution today continues to pose a threat to health worldwide. Recently, the World Health Organization (WHO) has estimated that about 4.2 and 3.8 million deaths per year are due to the exposure to outdoor air pollution and to the household exposure to smoke from dirty cookstoves and fuels, respectively. In addition, about 91% of the world population lives in places where the air quality exceeds the WHO guideline limits [1]. All these updates suggest that a proper control of air pollutant emissions is required.
Among the main classes of air pollutants, nitrogen oxides (NOx) identify a family of binary compounds of nitrogen and oxygen, including nitric oxide (NO) and nitrogen dioxide (NO 2 ) [2,3]. Both these species have been considered extremely harmful pollutants since 1952, when their role in the formation of photochemical smog was confirmed [4]. However, more in detail, NO, a colorless gas, is about four times less toxic than NO 2 , a reddish-brown gas [5,6].
In this drastic scenario, over the years numerous strategies have been developed and properly optimized with the final aim of decomposing NOx species. Among all the possibilities, their selective catalytic reduction by ammonia as a reducing agent (NH 3 -SCR) is one of the most studied and applied technologies [7][8][9], as well as absorption and adsorption routes [6,10]. However, from a practical viewpoint all these techniques suffer from several drawbacks, limiting their real application [3,8,11]. In this context, the ever more stringent legislative constraints related to NOx emissions are spurring the optimization of highly efficient techniques for NOx abatement [12]. Following these perspectives, innovative strategies based on the photocatalysis principles have emerged Materials 2023, 16, 1304 3 of 17 in the final composite was accurately dosed to the final enhancement of the night-time activity without affecting photocatalytic properties. The role of PANI in the NO 2 abatement was also investigated based on its physicochemical characteristics. The best materials were subjected to recycling tests to demonstrate their stability in the working conditions.
In brief, Evonik P25 TiO 2 (anatase: rutile 80: 20, 50 m 2 /g) was added under vigorous stirring to 290 mL of an aqueous solution prepared by dissolving 1 g of N-(4aminophenyl)aniline in HCl 0.1 M. The use of HCl is crucial not only to solubilize the reagent (N-(4-aminophenyl)aniline), but also for the progress of the reaction that requires acidic conditions and to obtain the final PANI in its conducting form (half-oxidized and half-protonated). Then, the suspension was irradiated with a UV halogen lamp (500 W, 30 mW/cm 2 ) for 2 h and subsequently stirred in the dark for further 4 h at room temperature. TiO 2 particles guarantee the formation of PANI oligomers and act as catalysts during the subsequent step that consists in the addition of a suitable amount of 30% H 2 O 2 aqueous solution to carry out oxidative polymerization reaction. The reaction mixture was then stirred in the dark and at room temperature for 1 day. Finally, the precipitate was collected by filtration, washed with water and acetone and dried at room temperature. Samples with different TiO 2 contents were synthesized by varying the (N-(4-aminophenyl)aniline: TiO 2 : H 2 O 2 molar ratios, namely 1.0:0.4:3.0, 1.0:0.4:1.0 and 1.0:0.8:1.0, which will be named PANI-TiO 2 _15%, PANI-TiO 2 _31% and PANI-TiO 2 _67%, respectively, where the percentage is related to the amount of the polymer in the final composites.
For the sake of comparison, pure PANI samples, prepared according to the literature syntheses, were also investigated as the references. The first, labeled as PANI_ref1, was prepared according to a conventional synthetic approach [51,72] starting from 0.04 M aniline aqueous solution acidified by HCl (aniline/HCl molar ratio = 0.1) and using K 2 S 2 O 8 as oxidant. The reaction was carried out at 0 • C, adopting an aniline/K 2 S 2 O 8 molar ratio of 0.67. The precipitate was collected by filtration after 8 h, then washed with water and acetone and dried at room temperature.
A second reference (named PANI_ref2) sample was synthesized from N-(4-aminophenyl) aniline using H 2 O 2 as the sole oxidant in the presence of FeCl 3 (N-(4-aminophenyl)aniline/Fe 3+ = 4000 wt/wt), able to catalyze the oxidative polymerization reaction [73,74]. In brief, 2.35 mL of 30% H 2 O 2 aqueous solution was added to 250 mL of an aqueous solution prepared by dissolving 1 g of N-(4-aminophenyl aniline) in HCl 0.1 M. Then, 0.15 mL of a 5 g/L FeCl 3 aqueous solution was added. The precipitate was collected by filtration after 24 h, washed with water and acetone and dried at room temperature.

Sample Characterization
Powder X-ray diffraction (PXRP) was performed on a Philips PW 3710 Bragg-Brentano goniometer equipped with a graphite-monochromatic Cu Kα radiation source, scintillation counter, 1 • divergence slit, 0.2 mm receiving slit and a 0.04 • Soller slit system. Diffractograms were collected at a nominal X-ray power of 40 kV and 40 mA, in a 2θ range between 10 • and 80 • .
Thermogravimetric analyses (TGA) were collected on a TGA/DSC 3+ Mettler Toledo instrument equipped with a 70 µL alumina crucible. Measurements were carried out between 30 and 900 • C with a heating rate of 5 • C/min and in air.
Specific surface area values were determined from adsorption isotherms of N 2 in subcritical conditions, measured using a Coulter SA3100 instrument (Beckman Life Sciences, Los Angeles, CA, USA) by the Brunauer-Emmett-Teller (BET) model. UV-vis absorption spectra of the samples, dissolved in DMF with and without the HCl addition, were recorded in 200-1000 nm range on a Shimadzu UV-2600 UV-vis spectrophotometer (Kyoto, Japan).
The morphologic characterization was carried out without any sample pretreatment using a scanning electron microscope, operating with a Field Emission source (model TESCAN S9000G, (Overcoached, Germany); Source: Schottky type FEG; Resolution: 0.7 nm at 15 keV (in In-Beam SE mode) and equipped with EDS Oxford Ultim Max (operated with Aztec software 6.0).

NO 2 Abatement
A total of 50 mg of each material was suspended in 5 mL isopropanol and sonicated by an ultrasonic bath to obtain a uniform dispersion. After 10 min, the mixture was deposited on a glass plate (3.3 cm × 11.5 cm) and dried in air overnight to obtain a thin film. Then, the glass plate was placed inside a 20 L Pyrex cylindrical batch reactor and NO 2 gas abatement tests were performed. They were carried out at room temperature for 3 h in the dark and under both LED (350 mA, 9-48 V, 16.8 W, 2900 lx) and UVA (λ max = 365 nm, 100 W/m 2 ) irradiation. The initial concentration of NO 2 was 500 ± 50 ppb. An ENVEA AC32e directly connected to the reactor measured the NOx (NO and NO 2 ) concentration at 30, 60 and 180 min by chemiluminescence principle. Photolysis tests indicate 10% pollutant degradation after 3 h of UVA light irradiation.

Reusability Tests
PANI_ref1, bare TiO 2 and PANI-TiO 2 _31% were selected to be tested for four consecutive tests to evaluate their stability. The samples were prepared as described in Section 2.1.
The tests were carried out at room temperature, exposing the material to 500 ppb of NO 2 . PANI_ref1 was maintained in the dark for 3 h, whereas bare TiO 2 and PANI-TiO 2 _31% were maintained in the dark for 1 h and then irradiated by UVA light for other 2 h. The chemiluminescence detector monitored NO, NO 2 and NOx concentration at 30, 60 and 180 min. After each test, the material was removed from the reactor, left in air for 12 h and then used for the subsequent cycle.

Results and Discussion
All the synthesized materials were characterized by several techniques (TG, BET, PXRD, SEM-EDS, FTIR and UV-vis) and used for the NO 2 abatement, as described below.

Materials Characterization
PANI-TiO 2 composites with different TiO 2 contents (from 15 to ca. 67%) were prepared varying the reactant molar ratios.
The actual composition, determined by TG analyses in air and expressed as weight percentage, is reported in Table 1, whereas the relative TG curves of the pristine PANIs and PANI-TiO 2 composites are reported in Figures S1 and 1. Physisorbed water content was determined from the weight loss at around 100 °C, whereas the HCl amount was related to the weight loss at 150-200 °C and the polymer content was estimated from the weight loss starting around 350 °C, assigned to the polymer backbone degradation. The remaining fraction at 700 °C was attributed to TiO2. Interestingly, the thermal stability of the polymer backbone is affected by the TiO2 content, as higher oxide amounts weaken the polyaniline inter-chain interactions [75].  The different PANI content is also appreciable in the PXRD patterns ( Figure 2) from the relative intensity of the peaks attributed to the TiO2 polymorphs (anatase and rutile) and to the ES-I phase of the polymer [76] (Figure 2, inset). Moreover, decreasing the PANI content leads to a more ordered arrangement of the polymer chains, i.e., a higher crystallinity of the PANI phases, as appreciable from the better resolved peaks. Physisorbed water content was determined from the weight loss at around 100 • C, whereas the HCl amount was related to the weight loss at 150-200 • C and the polymer content was estimated from the weight loss starting around 350 • C, assigned to the polymer backbone degradation. The remaining fraction at 700 • C was attributed to TiO 2 . Interestingly, the thermal stability of the polymer backbone is affected by the TiO 2 content, as higher oxide amounts weaken the polyaniline inter-chain interactions [75].
The different PANI content is also appreciable in the PXRD patterns ( Figure 2) from the relative intensity of the peaks attributed to the TiO 2 polymorphs (anatase and rutile) and to the ES-I phase of the polymer [76] (Figure 2, inset). Moreover, decreasing the PANI content leads to a more ordered arrangement of the polymer chains, i.e., a higher crystallinity of the PANI phases, as appreciable from the better resolved peaks.
The enhanced crystallinity of the PANI-TiO 2 heterostructures with higher oxide content pairs with the larger specific surface area values that are greatly promoted compared to the PANI reference materials (Table 1 and Figures S2 and S3). These differences between the nanocomposites and pristine PANIs can be traced back to the morphologies observed by SEM analyses. While the PANI_ref2 displays a compact morphology ( Figure S4), PANI-TiO 2 composites are composed of a matrix of PANI nanorods and aggregates of rounded oxide nanoparticles ( Figure 3). Among them, the PANI-TiO 2 _15% sample presents a less porous structure (Figure 3a) with PANI rods that look fused, explaining the lower sur- The enhanced crystallinity of the PANI-TiO2 heterostructures with higher oxide content pairs with the larger specific surface area values that are greatly promoted compared to the PANI reference materials (Table 1 and Figures S2 and S3). These differences between the nanocomposites and pristine PANIs can be traced back to the morphologies observed by SEM analyses. While the PANI_ref2 displays a compact morphology ( Figure S4), PANI-TiO2 composites are composed of a matrix of PANI nanorods and aggregates of rounded oxide nanoparticles ( Figure 3). Among them, the PANI-TiO2_15% sample presents a less porous structure (Figure 3a) with PANI rods that look fused, explaining the lower surface area. PANI_ref1 displays an intermediate surface area, which could be related to its coral-like structure, displaying micrometric rods covered by nanometric spikes (Figure 3d).
The different structural features of the synthesized materials can be associated with the diverse synthetic mechanisms involved. In fact, the traditional synthesis of PANI_ref1 consists in a two-steps reaction, involving the production of active radical species (aniline and its dimers) during the step 1, followed by a chain elongation process (step 2) [55]. The possibility to control the first step permits obtaining well organized structures characterized by high surface areas and porous morphologies. Similarly, the synthesis of PANI/TiO2 composites follows the same approach. However, as shown in Table 1, in this case the control of the radical species' production is correlated to the TiO2 quantity. In fact, for low amount of oxide particles, materials with poor characteristics are obtained (low value of surface area and more compact morphology). On the contrary, when the first step of the reaction occurs in the presence of larger amount of TiO2, the properties of the final composites improve. It is supposed that the first step of the reaction consists in the UVmediated PANI oligomer formation on the TiO2 surface and that, during the second step, the oligomers' polymerization occurs [71].
On the contrary, the reaction that leads to the production of PANI_ref2 does not permit a fine control of the radicals formation, because the oxidative polymerization of N-(4-aminophenyl)aniline is too fast to be controlled, leading to a decrease in the structural characteristics of the synthesized product [55].  EDX mapping of the PANI-TiO2 composites confirms the non-uniform distribution of the oxide nanoparticles in the polymer matrix. In particular, the TiO2 particles seem to be localized along the polymer fibers ( Figure 4). The different structural features of the synthesized materials can be associated with the diverse synthetic mechanisms involved. In fact, the traditional synthesis of PANI_ref1 consists in a two-steps reaction, involving the production of active radical species (aniline and its dimers) during the step 1, followed by a chain elongation process (step 2) [55]. The possibility to control the first step permits obtaining well organized structures characterized by high surface areas and porous morphologies. Similarly, the synthesis of PANI/TiO 2 composites follows the same approach. However, as shown in Table 1, in this case the control of the radical species' production is correlated to the TiO 2 quantity. In fact, for low amount of oxide particles, materials with poor characteristics are obtained (low value of surface area and more compact morphology). On the contrary, when the first step of the reaction occurs in the presence of larger amount of TiO 2 , the properties of the final composites improve. It is supposed that the first step of the reaction consists in the UVmediated PANI oligomer formation on the TiO 2 surface and that, during the second step, the oligomers' polymerization occurs [71].
On the contrary, the reaction that leads to the production of PANI_ref2 does not permit a fine control of the radicals formation, because the oxidative polymerization of N-(4-aminophenyl)aniline is too fast to be controlled, leading to a decrease in the structural characteristics of the synthesized product [55].
EDX mapping of the PANI-TiO 2 composites confirms the non-uniform distribution of the oxide nanoparticles in the polymer matrix. In particular, the TiO 2 particles seem to be localized along the polymer fibers ( Figure 4). EDX mapping of the PANI-TiO2 composites confirms the non-uniform distribution of the oxide nanoparticles in the polymer matrix. In particular, the TiO2 particles seem to be localized along the polymer fibers ( Figure 4).  FTIR spectra ( Figure 5) confirmed the formation of PANI in its emeraldine form, as appreciable from the similar relative intensity of the band of the quinoid ring C=C stretching (1570 cm −1 ) and of the stretching mode of the benzenoid ring (1490 cm −1 ) [51]. The band at about 1300 cm −1 is attributed to the C-N bending vibration in aromatic amines, whereas those at 1070 cm −1 and 870 cm −1 are related to the in-plane and out-of-plane bending modes of C-N. The stretching mode of N=Q=N (Q = quinoid ring) is responsible for the band at 1000 cm −1 . Similar results were obtained for the two pristine PANI samples ( Figure S5). The absence of signals in the 3600-3700 cm −1 region, characteristic of the free hydroxyl groups of TiO 2 , suggests a complete coating of TiO 2 particles with the PANI matrix.
whereas those at 1070 cm −1 and 870 cm −1 are related to the in-plane and out-of-plane bending modes of C-N. The stretching mode of N=Q=N (Q = quinoid ring) is responsible for the band at 1000 cm −1 . Similar results were obtained for the two pristine PANI samples ( Figure S5). The absence of signals in the 3600-3700 cm −1 region, characteristic of the free hydroxyl groups of TiO2, suggests a complete coating of TiO2 particles with the PANI matrix. This observation is also supported by the UV-vis spectra of the samples dissolved in DMF (Figure 6), where the signals of both the oxidized and reduced part of the polymer are appreciable. As an example, the UV-vis spectrum of PANI-TiO2_31% is shown in Figure 6. In particular, the π − π* transition of benzene is observed at ca. 310 nm, whereas the band at 580 nm can be attributed to the azaquinoid moieties π − π* transition band. Upon the HCl addition, the characteristic bands of emeraldine salt are observed, testifying the acid-base properties of the polymer: the signal around 310 nm is due to benzene rings, whereas the band around 430 nm is attributed to π − polaron and polaron − π* transitions, and the band at 900 nm has been related to polymer conformation changes, conjugation extension or to the de-aggregation of polymer chains in solution [77]. This observation is also supported by the UV-vis spectra of the samples dissolved in DMF (Figure 6), where the signals of both the oxidized and reduced part of the polymer are appreciable. As an example, the UV-vis spectrum of PANI-TiO 2 _31% is shown in Figure 6. In particular, the π − π* transition of benzene is observed at ca. 310 nm, whereas the band at 580 nm can be attributed to the azaquinoid moieties π − π* transition band. Upon the HCl addition, the characteristic bands of emeraldine salt are observed, testifying the acid-base properties of the polymer: the signal around 310 nm is due to benzene rings, whereas the band around 430 nm is attributed to π − polaron and polaron − π* transitions, and the band at 900 nm has been related to polymer conformation changes, conjugation extension or to the de-aggregation of polymer chains in solution [77]. In contrast, the two UV-vis spectra of the two pristine PANIs show an important difference ( Figure S6), strongly affecting the activity of the materials toward the NO2 abatement. In fact, comparing the intensity of the two main peaks, at ca. 310 nm and 600 nm, the second one, associated to quinone rings, is more intense for PANI_ref1, if compared to PANI_ref2, suggesting a greater number of imino-quinone rings in the chains In contrast, the two UV-vis spectra of the two pristine PANIs show an important difference ( Figure S6), strongly affecting the activity of the materials toward the NO 2 abatement. In fact, comparing the intensity of the two main peaks, at ca. 310 nm and 600 nm, the second one, associated to quinone rings, is more intense for PANI_ref1, if compared to PANI_ref2, suggesting a greater number of imino-quinone rings in the chains on the polymer.

NO 2 Abatement
All the synthesized materials were tested for the NOx abatement, exploiting the different characteristics of the components: redox properties and adsorption features of the PANI material and photocatalytic activity of TiO 2 . Figure 7 shows the results obtained when pristine PANI samples are used as active materials in dark conditions. In conditions, it was also evaluated that pristine TiO 2 leads to a 5% of NOx adsorption. Both PANI_ref1 and PANI_ref2 exhibit comparable activity in the NO2 abatement, reaching about 100% removal in 3 h (Figure 7a). If, at first, this extraordinary result could Both PANI_ref1 and PANI_ref2 exhibit comparable activity in the NO 2 abatement, reaching about 100% removal in 3 h (Figure 7a). If, at first, this extraordinary result could be attributed to the adsorption capability of the materials, already demonstrated for the VOC reduction in air matrix [55], the results of NOx abatement (Figure 7b) suggest otherwise. In fact, PANI_ref1 leads to the removal of 47% of the initial NOx, whereas PANI_ref2 reaches only 14% abatement. On the other hand, the NO analyses (Figure 7c) display a gradual increase in the NO amount during the tests. Based on these outcomes, it is possible to affirm that for the NOx removal, PANI acts by two simultaneous different paths: NO 2 adsorption and chemical reduction in NO 2 to NO, as described in Scheme 1. be attributed to the adsorption capability of the materials, already demonstrated for the VOC reduction in air matrix [55], the results of NOx abatement (Figure 7b) suggest otherwise. In fact, PANI_ref1 leads to the removal of 47% of the initial NOx, whereas PANI_ref2 reaches only 14% abatement. On the other hand, the NO analyses (Figure 7c) display a gradual increase in the NO amount during the tests. Based on these outcomes, it is possible to affirm that for the NOx removal, PANI acts by two simultaneous different paths: NO2 adsorption and chemical reduction in NO2 to NO, as described in Scheme 1. The chemical reduction in NO2 to NO can be reasonably associated to the redox properties of the PANI chains. In fact, as described in Scheme 1, the amino-benzenic units are able to reduce NO2 to NO-producing imino-quinones units. In air, these latter are quickly regenerated maintaining the materials' activity. As reported in the literature [78], NO2 is a strong electron acceptor gas able to carry out oxidation reactions. So far, the adsorption ability of PANI has been only highlighted, affirming that, at room temperature, NO2 is strongly adsorbed on the polymer, needing a proper thermal The chemical reduction in NO 2 to NO can be reasonably associated to the redox properties of the PANI chains. In fact, as described in Scheme 1, the amino-benzenic units are able to reduce NO 2 to NO-producing imino-quinones units. In air, these latter are quickly regenerated maintaining the materials' activity. As reported in the literature [78], NO 2 is a strong electron acceptor gas able to carry out oxidation reactions. So far, the adsorption ability of PANI has been only highlighted, affirming that, at room temperature, NO 2 is strongly adsorbed on the polymer, needing a proper thermal treatment for the material regeneration. Here, for the first time, we demonstrate that PANI acts through both NO 2 adsorption and chemical reduction to NO that simultaneously permit the NO 2 abatement.
It is worth noting that, although both the materials exhibit very similar performances in the NO 2 removal, they differ in the preferential mechanism employed.
In fact, if, on the one hand, PANI_ref1 acts through both approaches, converting 54% NO 2 to NO and adsorbing 46% of the initial NO 2 , on the other hand, PANI_ref2 proceeds almost exclusively by NO 2 reduction, reducing 86% NO 2 into NO and adsorbing only 14% of the initial gas. It is important to highlight that, unlike NO 2 , NO is not an irritant gas and much less toxic than NO 2 [79]. No major role is attributed to an oxidation of inhaled NO into NO 2 in the lungs since, after inhalation, NO is eliminated faster than it is oxidized NO [80].
The different behavior of the two polymers can be attributed to their difference in specific surface area values and oxidation level of the polymeric chains, as described above in Table 1 and Figure S6.
The attempts carried out to characterize the used materials by UV-vis spectroscopy did not show any evident differences compared to the fresh ones. In fact, the low NO 2 initial concentration (50 ppb) and the fast internal redox chemical rearrangement between imino-quinone and amino-benzenic units do not permit to evidence a change in the polymer structure.
The test carried out under light irradiation (both UVA and LED) led to similar results ( Figures S7 and S8), confirming that, in photocatalytic processes, pristine PANI does not act as a photocatalyst, but as a photosensitizer [81].
PANI/TiO 2 nanocomposites showed a behavior intermediate between that of PANI and that of bare TiO 2 . In fact, when exposed to NO 2 gas in the dark, they maintained the same activity of pristine PANI_ref2, removing almost completely the gaseous pollutant ( Figure S9a). In contrast, concerning the NOx abatement, the performance of the materials is strictly dependent on the polymer content, increasing with the amount of PANI in the composites. Passing from PANI-TiO 2 _15% to PANI-TiO 2 _67%, the NOx abatement ranges from 35% to 50% ( Figure S9b). On the contrary, the percentage of NO 2 converted into NO follows the opposite trend: PANI-TiO 2 _15% > PANI-TiO 2 _31% > PANI-TiO 2 _67% ( Figure S9c), attributing to the increase in the polymer content this property of the heterostructures. As expected, in the dark, TiO 2 does not show any activity in the NO 2 abatement.
When exposed to UVA light irradiation, PANI/TiO 2 nanocomposites exhibit superior activity compared to the two bare components (PANI and TiO 2 ), as reported in Figure 8.
PANI/TiO 2 nanocomposites maintain a high capability toward the NO 2 abatement that was similar to PANIs and greater than the performance of bare TiO 2 . In fact, all the heterostructures can remove more than 90% NO 2 in 1 h, whereas the result with pristine TiO 2 settles at 70% in the same conditions (Figure 8a). Moreover, concerning the NOx abatement, the heterostructures characterized by higher TiO 2 content (PANI-TiO 2 _15% and PANI-TiO 2 _31%) lead to full NOx removal in 3 h, whereas when PANI content exceeds a certain amount (PANI-TiO 2 _67%), the NOx degradation drops to 85% (Figure 8b). Finally, all nanocomposites in the first 30 min of light irradiation, acting by the chemical reduction path typical of PANI, convert 55-60% of initial NO 2 to NO (Figure 8c). However, continuing with the irradiation, all the heterostructures are also able to degrade NO. The degradation percentage depends on the materials' composition, leading to full NO removal in 3 h for the composites with higher TiO 2 content (PANI-TiO 2 _15% and PANI-TiO 2 _31%) and 90% NO abatement for the photocatalyst more reachable in the polymer (PANI-TiO 2 _67%).
As reported in the literature [82], the photocatalytic NOx degradation consists in a gradual oxidation of both NO and NO 2 to HNO 3 or NO 3 − depending on the reaction conditions. However, in the presence of a certain amount of water droplets in the air (e.g., rain or artificial sprays of water), NO 2 can undergo disproportionation, leading to HNO 3 and NO that in air is quickly re-oxidized to NO 2 . and NO that in air is quickly re-oxidized to NO2.
Although PANI is known to be a very promising photosensitizer able to extend the photoactivity of TiO2 in the visible part of the electromagnetic spectrum, the photocatalytic tests carried out under visible light irradiation showed that, under these conditions, all the nanocomposites maintain the activity of the polymeric component without any improvement in the photocatalytic properties ( Figure S10). This is associated with the low TiO2 content of the heterostructures, if compared with other similar PANI/TiO2 materials employed in the photodegradation of pollutants [83]. Although PANI is known to be a very promising photosensitizer able to extend the photoactivity of TiO 2 in the visible part of the electromagnetic spectrum, the photocatalytic tests carried out under visible light irradiation showed that, under these conditions, all the nanocomposites maintain the activity of the polymeric component without any improvement in the photocatalytic properties ( Figure S10). This is associated with the low TiO 2 content of the heterostructures, if compared with other similar PANI/TiO 2 materials employed in the photodegradation of pollutants [83].

Reusability Tests
PANI_ref1 and PANI-TiO 2 _31% were selected to test their stability on the basis of the reusability tests. Each material was tested for four consecutive cycles, as reported above.
The reusability results obtained in the NO 2 and NOx abatement and NO formation in the dark using PANI_ref1 as active material are reported in Figure S11, confirming the high stability of the polymer that exhibits the same behavior for all the performed tests.
As expected, bare TiO 2 particles are completely inactive in the first 60 min, when the UVA light is off, whereas they quickly lead to NOx and NO 2 photodegradation under light irradiation ( Figure S12). The NO degradation is not reported, but the concentration was zero after light irradiation. Figure 9 shows the results of the reusability tests performed by PANI-TiO 2 _31%. PANI-TiO2_31% exhibits very high stability for all the recycle tests. It is able to degrade about 60% NO2 in the dark, converting only 20% of the initial gas to NO ( Figure  9a). Under light irradiation, its photocatalytic properties prevail, leading to quite complete NOx degradation (Figure 9b,c). PANI-TiO 2 _31% exhibits very high stability for all the recycle tests. It is able to degrade about 60% NO 2 in the dark, converting only 20% of the initial gas to NO (Figure 9a). Under light irradiation, its photocatalytic properties prevail, leading to quite complete NOx degradation (Figure 9b,c).

Conclusions
In this work, PANI/TiO 2 nanocomposites were properly synthesized by a green UV light-driven oxidative polymerization and used for NO 2 abatement. For the first time, the double role of the PANI matrix in the NO 2 removal by both adsorption and chemical reduction to NO was demonstrated. The first mechanism (adsorption) can be correlated to the van der Walls interaction as well as to the formation of hydrogen bonds between the polluting gas and the polymer's protonated amino and imino groups. Moreover, these organic groups (in particular -NH 2 ) are able to strongly interact with polar species such as NO 2 thanks to electrostatic dipole-dipole forces resulting in enhanced adsorption. In contrast, the second one (chemical reduction) affects the amino-benzenic units of PANI that are easily oxidized to the corresponding imino-quinone groups. The double mechanism ensures the total removal of NO 2 in dark conditions but at the same time the NO accumulation. However, exploiting the photocatalytic properties of the heterostructures, when UVA light is turned on, all the nanocomposites are able to complete the NOx photodegradation in 3 h.
The high stability of the materials, tested through four consecutive cycles, makes them attractive for further investigation for the realization of innovative and efficient night-and-day surfaces to be applied for indoor and outdoor environments.

Data Availability Statement:
The data that support the plots within this paper are available from the corresponding author on reasonable request.

Conflicts of Interest:
The authors declare no conflict of interest.