Low-Temperature O3 Decomposition over Pd-TiO2 Hybrid Catalysts

: In aircraft and spacecraft, outside air is not directly fed to the passenger because it contains ozone at elevated altitudes. The decomposition of low concentration ozone in the air was carried out at 25 ◦ C by catalytic oxidation on Pd-based catalysts supported on a high surface area hybrid TiO 2 . The use of these hybrid catalysts has shown a beneﬁcial effect, both on the catalytic activity and on the catalyst stability. Kinetic studies showed that the most promising catalytic phase (Pd/TiO 2 _100) was the one obtained from the TiO 2 support containing the lowest content of citrate ligands and leading to small Pd particles (around 4 nm). The effect of catalyst synthesis on the decomposition of O 3 gas (15 ppm) in a dry and humid (HR = 10%) stream in a closed environment such as aircraft or spacecraft was also investigated in this study and further elucidated by detailed characterizations. It was shown that the system could be used as an effective treatment for air coming from outside.


Introduction
Human activity in various industrial sectors can be considered the main cause of air pollution that leads to the presence of gaseous air pollutants such as CO, O 3 , NO 2 , and SO 2 [1,2]. Ozone is considered to be a secondary air pollutant with environmental and human toxicity [3][4][5][6]. It results from complex photochemical reactions between different pollutants such as nitrogen oxides (NO x ), carbon monoxide (CO), volatile organic compounds (VOCs) (fuels, paints, manufacturing facilities, etc.), and sunlight [7]. However, at low concentrations, ozone is corrosive, strongly oxidizing, and toxic to human health [8][9][10]. Thermodynamically, ozone is an unstable gas (at 298 K ∆rG • = −163 kJ mol −1 and ∆rH • = −138 kJ mol −1 ).
It forms naturally in the stratosphere, forming the protective layer of ozone around 20 km altitude, by a photochemical phenomenon involving oxygen and UV radiation [8,11].
The U.S. Environmental Protection Agency (EPA) fixed the outdoor ozone levels at 0.08 ppm for 8 h. The Food and Drug Administration (FDA) requires ozone output from indoor medical devices to be less than 0.05 ppm, but short-term exposures to ozone at levels below international standards also result in an increased risk of death [12].
For example, in airplanes and spacecraft, outside air is not directly supplied to passengers because it contains a high concentration of ozone at high altitudes [23,24]. Catalytic converters are hence needed to lower ozone concentrations to the authorized values. Such onboard equipment already exists but must evolve to be adapted to future aircraft and spacecraft technologies and will require working at much lower temperatures and in the presence of lower relative humidity. In the decomposition of ozone, the use of catalysts based on transition or noble metals alone, or supported on oxides, is the most widely used process because it allows high efficiency and low energy consumption [8][9][10][25][26][27]. Moreover, in the literature [12], catalysts based on manganese and palladium are known to be effective in the reaction of the decomposition of this trioxygen (O 3 ) molecule [28][29][30][31]. Likewise, previous studies have been performed in a temperature range between 20 and 100 • C and O 3 concentrations between 5 and 1600 ppm. In addition, various parameters can influence the performance of catalysts used, such as sulfur, NO x , humidity, ozone concentration, and temperature [8,11]. An important specific surface area and small size of metal nanoparticles have been shown to achieve very high efficiency in the decomposition of ozone [17,32]. In addition, among the catalysts, palladium supported on titanium dioxide presents an important challenge for the O 3 decomposition because of its large specific surface area and the synthesis method [17,18]. In this study, we developed new Pd-based heterogeneous catalysts for the decomposition of ozone. For this purpose, titania-supported palladium nanoparticles by one-pot deposition method using citrate-functionalized TiO 2 hybrid support nanomaterials were investigated [33,34].
Functionalization of the titanium oxide surface by citrate (Cit) groups could provide active centers for (i) the selective Pd deposition, (ii) "in situ" reduction, and (iii) control of the growth step of Pd nanoparticles (NPs). At room temperature, different Cit/Ti molar ratios were evaluated in the decomposition of ozone under very specific conditions (low concentration of the order of 15 ppm, relative humidity of the order of 10%) and close to the real conditions for the elimination of this molecule in the cabins of planes and spacecraft.

Catalysts Synthesis
The catalysts were synthesized in two major steps, the first one consisting in hydrolyzing, via the Sol-Gel process, a mixture of Ti(O i Pr) 4 , and a heteroleptic alkoxide precursor of general formula [Ti(OiPr) x (Cit) y ] m to obtain TiO 2 nanoparticles surface functionalized by citrate ligands. Depending on the initial citrate/Ti molar ratio (x), a series of hybrid nano-objects (Cit) 1 (TiO 2 ) x (x = 20; 50 and 100) with controlled surface citrate density are obtained. These latter are used as a reduction site in a second step to deposit Pd nanoparticles (PdNPs) on TiO 2 . A mild thermal treatment (70 • C under air) is finally applied to clean the surface from any residual solvent.

Physicochemical and Morphological Properties of Catalysts
The properties of catalysts were first assessed by in situ ATR-IR. Results of the Pd/TiO 2 _x catalysts have shown similar profiles to those of the supports alone ((Cit) 1 (TiO 2 ) x ) ( Figure 1) with bands at 1620 cm −1 and 1380 cm −1 that were attributed to the presence of carboxylate groups COO − [35]. The bands due to stretching vibrations ν (C=O) groups of citric acid (at 1735 cm −1 and 1710 cm −1 ) were not detected, indicating the presence of higher TiO 2 _citrate interactions.
showed the disappearance of the band relative to νCOOlocated at 1385 cm −1 . For all catalysts (Pd/TiO2_x), two new vibration bands situated at 1535 and 1398 cm −1 were observed, suggesting that the deposition of Pd allows modifying the citrate coordination on the TiO2 surface since a higher proportion of bridged moieties after Pd deposition was observed. This result would be in agreement with the presence of a strong interaction of the carboxylic groups of citrates with Pd.  Table 1. Irrespective of the catalyst, the amount of palladium deposited was close to the desired value (0.40 wt%). Moreover, the percentage of Pd fixed on the various hybrid supports (Cit)1(TiO2)x seemed to be related to the quantity of citrate present. Indeed, in the presence of a high citrate content in the hybrid support (Pd/TiO2_20), the quantity of Pd measured (0.40 wt%) is the one excepted. A decrease in citrate in the support led to a decrease in the amount of Pd deposited (0.32% of Pd in Pd/TiO2_100), suggesting that the amount of citrate on the TiO2 surface is not sufficient to reduce the amount of palladium initially introduced (0.40%). The isotherms of adsorption-desorption of N2 of all Pd-based catalysts were of type IV, according to the IUPAC classification ( Figure 2). The presence of hysteresis suggested a mesoporosity of the surface of the studied materials in which a capillary condensation occurred. After functionalization with citrate precursors, the hysteresis loop extended over a relatively wide range of relative pressures; at P/P° values between 0.40 and 0.47, this decrease could be related to the increase in the Ti/Cit ratio. This value was higher (close to 0.54) for Pd/TiO2_free citrate. The results for the measurements of specific surface The characteristic vibrations of the inorganic TiO 2 : Ti-O-Ti (1120 cm −1 ) and Ti-O (962 and 720 cm −1 ) [36] did not change, showing that the inorganic structure of TiO 2 was not impacted by the presence of citric acid. After palladium deposition, the ATR-IR spectra showed the disappearance of the band relative to νCOO − located at 1385 cm −1 . For all catalysts (Pd/TiO 2 _x), two new vibration bands situated at 1535 and 1398 cm −1 were observed, suggesting that the deposition of Pd allows modifying the citrate coordination on the TiO 2 surface since a higher proportion of bridged moieties after Pd deposition was observed. This result would be in agreement with the presence of a strong interaction of the carboxylic groups of citrates with Pd. ICP-AES analyses for different catalysts are summarized in Table 1. Irrespective of the catalyst, the amount of palladium deposited was close to the desired value (0.40 wt%). Moreover, the percentage of Pd fixed on the various hybrid supports (Cit) 1 (TiO 2 ) x seemed to be related to the quantity of citrate present. Indeed, in the presence of a high citrate content in the hybrid support (Pd/TiO 2 _20), the quantity of Pd measured (0.40 wt%) is the one excepted. A decrease in citrate in the support led to a decrease in the amount of Pd deposited (0.32% of Pd in Pd/TiO 2 _100), suggesting that the amount of citrate on the TiO 2 surface is not sufficient to reduce the amount of palladium initially introduced (0.40%). The isotherms of adsorption-desorption of N 2 of all Pd-based catalysts were of type IV, according to the IUPAC classification ( Figure 2). The presence of hysteresis suggested a mesoporosity of the surface of the studied materials in which a capillary condensation occurred. After functionalization with citrate precursors, the hysteresis loop extended over a relatively wide range of relative pressures; at P/P • values between 0.40 and 0.47, this decrease could be related to the increase in the Ti/Cit ratio. This value was higher (close to 0.54) for Pd/TiO 2 _free citrate. The results for the measurements of specific surface area (S BET ), pore volume (Vp), and average pore diameter (Dp) of the catalysts prepared in the presence and absence of citrate precursors are given in Table 2. nium oxide particles. BET surface areas increased with increasing citrate content; hence a higher value was obtained for the highly functionalized material TiO2_20 exhibiting a surface area of 335 m 2 g −1 . This was consistent with the presence of an important amount of surface organic functionalization, which could induce self-assembly of the nanocrystallites and create extra mesoporosity, as evidenced by the N2 adsorption-desorption isotherm ( Figure 2). The XRD patterns of the different catalysts are shown in Figure 3. All functionalization materials exhibited the anatase phase. The material containing the largest amount of citrate also displayed the poorest crystallinity, showing that a significant amount of citrate on the surface resulted in a disruption of the crystal lattice. After reductive deposition of nanoparticles of Pd and compared with the supports, the crystallinity of TiO2 anatase did not change (JCPDS card no.21-1272). Moreover, the metallic Pd phase was never detected in the XRD diffractograms, which may come from the small amount of deposited Pd but also indicates a good dispersion of Pd nanoparticles on the surface of TiO2.
Analysis of the XPS spectra of Pd 3d5/2 showed the presence of two contributions corresponding, respectively, to metallic Pd° and PdOx species ( Figure 4). Binding energies for 3d5/2 and 3d3/2 appeared at 334.3 eV and 339.6 eV for Pd° and 336.1 eV and 341.4 eV for PdOx, respectively [37,38]. In addition, the presence of citrate species favored the reducibility of Pd. The amount of reduced palladium (Pd°) increased as the Cit/Ti ratio  Interestingly, all the modified titanium oxides exhibited higher S BET specific surfaces, lower pore volumes (Vp), and smaller pore sizes than Pd/TiO 2 _citrate free. The low values of the C BET constant confirmed that the -COOH groups were located on the surface of titanium oxide particles. BET surface areas increased with increasing citrate content; hence a higher value was obtained for the highly functionalized material TiO 2 _20 exhibiting a surface area of 335 m 2 g −1 . This was consistent with the presence of an important amount of surface organic functionalization, which could induce self-assembly of the nanocrystallites and create extra mesoporosity, as evidenced by the N 2 adsorption-desorption isotherm ( Figure 2). The XRD patterns of the different catalysts are shown in Figure 3. increased: 75% of Pd° using (Cit)1(TiO2)20 versus 70% using (Cit)1(TiO2)100 as support (Table  1). In contrast, the Pd/Ti atomic ratio tended to decrease when increasing the citrate content from x = 100 to 20. This suggests that too high citrate content can lead to a depleted dispersion of Pd onto the TiO2 surface and the entrapment of a non-negligible fraction of Pd inside the TiO2 support (Table 1).   All functionalization materials exhibited the anatase phase. The material containing the largest amount of citrate also displayed the poorest crystallinity, showing that a significant amount of citrate on the surface resulted in a disruption of the crystal lattice. After reductive deposition of nanoparticles of Pd and compared with the supports, the crystallinity of TiO 2 anatase did not change (JCPDS card no. . Moreover, the metallic Pd phase was never detected in the XRD diffractograms, which may come from the small amount of deposited Pd but also indicates a good dispersion of Pd nanoparticles on the surface of TiO 2 . Analysis of the XPS spectra of Pd 3d 5/2 showed the presence of two contributions corresponding, respectively, to metallic Pd • and PdO x species (Figure 4). Binding energies for 3d 5/2 and 3d 3/2 appeared at 334.3 eV and 339.6 eV for Pd • and 336.1 eV and 341.4 eV for PdO x , respectively [37,38]. In addition, the presence of citrate species favored the reducibility of Pd. The amount of reduced palladium (Pd • ) increased as the Cit/Ti ratio increased: 75% of Pd • using (Cit) 1 (TiO 2 ) 20 versus 70% using (Cit) 1 (TiO 2 ) 100 as support ( Table 1). In contrast, the Pd/Ti atomic ratio tended to decrease when increasing the citrate content from x = 100 to 20. This suggests that too high citrate content can lead to a depleted dispersion of Pd onto the TiO 2 surface and the entrapment of a non-negligible fraction of Pd inside the TiO 2 support (Table 1).
increased: 75% of Pd° using (Cit)1(TiO2)20 versus 70% using (Cit)1(TiO2)100 as support (Table  1). In contrast, the Pd/Ti atomic ratio tended to decrease when increasing the citrate content from x = 100 to 20. This suggests that too high citrate content can lead to a depleted dispersion of Pd onto the TiO2 surface and the entrapment of a non-negligible fraction of Pd inside the TiO2 support (Table 1).  The size distributions of Pd nanoparticles (PdNPs) prepared using different (Cit)1(TiO2)x supports, obtained by TEM analysis, are shown in Figure 5. TEM micrographs showed that the produced PdNPs were of spherical shape with an average size between 3 and 7 nm (Table 2). The size distributions of Pd nanoparticles (PdNPs) prepared using different (Cit) 1 (TiO 2 ) x supports, obtained by TEM analysis, are shown in Figure 5. TEM micrographs showed that the produced PdNPs were of spherical shape with an average size between 3 and 7 nm ( Table 2). This result confirmed the reducing effect of the carboxylate groups of the citrates in agreement with ATR-IR results. Citrate amounts seemed to have an impact on the diameter distribution of palladium nanoparticles. It is interesting to note that PdNPs prepared from (Cit)1(TiO2)50 have the smallest mean diameter, which is 3 ± 0.8 nm. By increasing the citrate amount (x = 20), the PdNPs had a mean diameter of 6 nm and showed a broader size distribution than (Cit)1(TiO2)50 and (Cit)1(TiO2)100 supports ( Figure 5). This result could be explained by the agglomeration of very small particles of Pd, in agreement with the results reported by Y. Sun et al. [39]. Indeed, a higher concentration of citrate could generate the appearance of more Pd nuclei at the same place and thus favor the growth of Pd nanoparticles.

Effect of Pd Free Support on O3 Decomposition
The results of the decomposition of ozone on the different synthesized and commercial TiO2 supports (TiO2_citrate free, used as reference) under dry and humid conditions at 25 °C have been presented in Figure 6. For all supports, a transient decrease in ozone conversion efficiency occurred systematically initially (for about 1 h) until stabilization of the ozone conversion (at around 20% on TiO2_citrate free). All hybrid TiO2 materials This result confirmed the reducing effect of the carboxylate groups of the citrates in agreement with ATR-IR results. Citrate amounts seemed to have an impact on the diameter distribution of palladium nanoparticles. It is interesting to note that PdNPs prepared from (Cit)1(TiO 2 ) 50 have the smallest mean diameter, which is 3 ± 0.8 nm. By increasing the citrate amount (x = 20), the PdNPs had a mean diameter of 6 nm and showed a broader size distribution than (Cit) 1 (TiO 2 ) 50 and (Cit) 1 (TiO 2 ) 100 supports ( Figure 5). This result could be explained by the agglomeration of very small particles of Pd, in agreement with the results reported by Y. Sun et al. [39]. Indeed, a higher concentration of citrate could generate the appearance of more Pd nuclei at the same place and thus favor the growth of Pd nanoparticles.

Effect of Pd Free Support on O 3 Decomposition
The results of the decomposition of ozone on the different synthesized and commercial TiO 2 supports (TiO 2 _citrate free, used as reference) under dry and humid conditions at 25 • C have been presented in Figure 6. For all supports, a transient decrease in ozone conversion efficiency occurred systematically initially (for about 1 h) until stabilization of the ozone conversion (at around 20% on TiO 2 _citrate free). All hybrid TiO 2 materials exhibited higher activity in ozone decomposition than TiO 2 _citrate free. This efficiency could be partly explained by a modification of the textural properties of these TiO 2 hybrid materials since their specific surface area (between 263-335 m 2 g −1 ) was much larger than the one of TiO 2 _citrate free (232 m 2 g −1 ). Regardless of the solid, the presence of water (in presence of humid air) led to a decrease in the ozone conversion compared with experiments carried out in the presence of dry air. As an example, the ozone conversion was close to 30% under dry air on (Cit) 1 (TiO 2 ) 100 and equal to 23% when humid conditions were used.
Catalysts 2022, 12, x 7 of 14 exhibited higher activity in ozone decomposition than TiO2_citrate free. This efficiency could be partly explained by a modification of the textural properties of these TiO2 hybrid materials since their specific surface area (between 263-335 m 2 g −1 ) was much larger than the one of TiO2_citrate free (232 m 2 g −1 ). Regardless of the solid, the presence of water (in presence of humid air) led to a decrease in the ozone conversion compared with experiments carried out in the presence of dry air. As an example, the ozone conversion was close to 30% under dry air on (Cit)1(TiO2)100 and equal to 23% when humid conditions were used.
However, the addition of palladium remarkably improved the efficiency of these nanomaterials in the decomposition of O3 under both dry and humid conditions ( Figure  7a,b), indicating the beneficial effect of palladium on the limitation of decreasing O3 decomposition (Conversion (%) = 40-100) [8,24].
However, the addition of palladium remarkably improved the efficiency of these nanomaterials in the decomposition of O 3 under both dry and humid conditions (Figure 7a,b), indicating the beneficial effect of palladium on the limitation of decreasing O 3 decomposition (Conversion (%) = 40-100) [8,24].
These results show that impregnating Pd can promote the conversion of O 3 very effectively for humid and dry ozone gas mixtures. Independently of the hybrid-supported Pd-based catalyst used, the ozone conversion was between 76% and 100% after 3 h of reaction under dry and humid conditions. On the other hand, a significant influence of the nature of the support was noticed in both conditions. The catalysts based on hybrid TiO 2 showed higher efficiency for ozone decomposition compared with Pd/TiO 2 _citrate free. Indeed, Pd/TiO 2 _100 as a catalyst allowed obtaining a total conversion of ozone with or without dry air, whereas the ozone conversion determined over Pd/TiO 2 _citrate free after 180 min of reaction time was close to 40% under dry and 21% under humid air. The beneficial effect could be attributed to a better dispersion of Pd on the surface of TiO 2 because of the presence of a citrate group at the surface of the different supports. Indeed, according to XPS and TEM results, the catalyst containing the lowest citrate content (Pd/TiO 2 _100) exhibited Pd nanoparticles with an average diameter of about 4 nm. By increasing the amount of citrate, the particle size was also increased to 6 nm, causing a decrease in O 3 conversion (Pd/TiO 2 _20). Therefore, the small size of Pd nanoparticles appeared to increase the conversion of ozone to oxygen at the surface of the catalyst. on ozone conversion for support only ((Cit)1(TiO2)x (x = 20, 50 and 100) and TiO2_citrate free) (Conversion (%) = 20-24) ( Figure 6).
However, the addition of palladium remarkably improved the efficiency of these nanomaterials in the decomposition of O3 under both dry and humid conditions ( Figure  7a,b), indicating the beneficial effect of palladium on the limitation of decreasing O3 decomposition (Conversion (%) = 40-100) [8,24]. On all Pd supported on hybrid supports, the TOF values were always between 1.7 and 2.5 times higher than over Pd/TiO 2 _citrate free (Table 3). However, considering their respective Pd loadings and particle sizes as determined by BET and TEM, TOF calculations showed that the intrinsic activities of the Pd supported on hybrid supports (Pd/TiO 2 _20, Pd/TiO 2 _50, and Pd/TiO 2 _100) and Pd/TiO 2 _citrate free catalysts were really different reaching values around 135.1 min −1 for Pd/TiO 2 _citrate free and 235.4-348.7 min −1 for Pd/TiO 2 _x in dry conditions. Under humid conditions (RH = 10%), a decrease in ozone conversion efficiency was observed for the Pd/TiO 2 _citrate free catalyst indicating a strong inhibiting effect of water on the decomposition of ozone. This phenomenon was highlighted by a drop in TOF (>10%). On the other hand, the TOF values were not really very different (<0.5%) for Pd supported on hybrid supports, indicating that the presence of citrate decreased the water impact and thus highlighted the beneficial effect of the presence of citrate moieties on the catalytic properties of Pd supported on hybrid materials. The positive effect of palladium in limiting the decrease in ozone decomposition in the presence of water has already been observed [13,[40][41][42][43], although several authors did not observe the inhibitory effect of water vapor [9,10,27,44,45]. This difference could be explained by the presence of NO x inside the reactor when the air was used instead of oxygen to generate ozone [46].
Indeed, the effect of water has been shown to be limited in the presence of NO x , as demonstrated by Subrahmanyam et al. on activated carbon [46] and Mehandjiev et al. on catalysts based on manganese oxide [47][48][49]. However, the addition of a small amount of palladium was found to considerably limit this undesirable effect. It has been proposed that the inhibiting effect of water was not due to a surface modification of the catalyst but rather to the blocking of active sites responsible for ozone decomposition by water adsorption [41,50]. If NO x species were produced in the gas stream, adsorbed water could react with them to form nitric or nitrous acid and hence limit the negative effect of water [11]. This would result in the removal of water and gas phase NO x species and the release of sites for ozone reaction. The role of palladium could be favored by electron transfer to titanium, which contributed to the desorption of water from titanium cations [40,41,51]. The stability of such catalysts could be explained by the presence of Pd, which is well-known for its high efficiency in oxidation reactions; on the other hand, the citrate-modified TiO 2 supports, rich in OH groups, as is known by a large specific surface area, has an important effect also in this capacity of decomposition of O 3 .

Effect of the Catalyst Weight in the O 3 Decomposition
The amount of the catalyst is an important parameter because it determines the ozone decomposition capacity under operating conditions. As expected, the increase in catalyst quantity led to an increase in ozone decomposition both in dry and humid air (Figure 8).
Catalysts 2022, 12, x 9 of 14 rather to the blocking of active sites responsible for ozone decomposition by water adsorption [41,50]. If NOx species were produced in the gas stream, adsorbed water could react with them to form nitric or nitrous acid and hence limit the negative effect of water [11]. This would result in the removal of water and gas phase NOx species and the release of sites for ozone reaction. The role of palladium could be favored by electron transfer to titanium, which contributed to the desorption of water from titanium cations [40,41,51].
The stability of such catalysts could be explained by the presence of Pd, which is wellknown for its high efficiency in oxidation reactions; on the other hand, the citrate-modified TiO2 supports, rich in OH groups, as is known by a large specific surface area, has an important effect also in this capacity of decomposition of O3.

Effect of the Catalyst Weight in the O3 Decomposition
The amount of the catalyst is an important parameter because it determines the ozone decomposition capacity under operating conditions. As expected, the increase in catalyst quantity led to an increase in ozone decomposition both in dry and humid air (Figure 8).
Above 20 mg of catalyst, full ozone conversion was noticed whatever the Pd supported on hybrid supports. For Pd/ TiO2_citrate free, the conversion of ozone seemed to be limited to 52% conversion in dry air and 43% in the humid air.

Preparation of Hybrid (Cit)1(TiO2)x (x = 20; 50 and 100)
The hybrid titanium oxide synthesis method used was described previously [33]. Under argon, 1.06 mL of titanium isopropoxide Ti(O i Pr)4 and 248 mg of citric acid (Ti: Cit molar ratio = 3) were added to 10 mL of THF. The medium was kept under vigorous stirring for 16 h. The 13 C NMR analysis of the mixture shows the formation of the Ti3O(O i Pr)8(OOCCH2-C(OH)(COO)-CH2COOH) complex. Then, 10 mL of isopropanol (HO i Pr) and (x-3) equivalent mol of Ti(O i Pr)4, regarding to citrates, were added to the complex allowing to obtain three hybrid supports with various Ti/Cit ratios (20, 50, and 100). This solution was added to 0.1 equivalents of NBu4Br in 100 mL of distilled water in a two-necked flask maintained under reflux at 100 °C. The suspensions were heated under reflux for 3 h. The precipitate was then washed twice with water and once with ethanol and finally dried at 70 °C for 20 h. Above 20 mg of catalyst, full ozone conversion was noticed whatever the Pd supported on hybrid supports. For Pd/TiO 2 _citrate free, the conversion of ozone seemed to be limited to 52% conversion in dry air and 43% in the humid air.

Materials
Titanium isopropoxide (Ti(O i Pr) 4  The hybrid titanium oxide synthesis method used was described previously [33]. Under argon, 1.06 mL of titanium isopropoxide Ti(O i Pr) 4 4 , regarding to citrates, were added to the complex allowing to obtain three hybrid supports with various Ti/Cit ratios (20, 50, and 100). This solution was added to 0.1 equivalents of NBu 4 Br in 100 mL of distilled water in a two-necked flask maintained under reflux at 100 • C. The suspensions were heated under reflux for 3 h. The precipitate was then washed twice with water and once with ethanol and finally dried at 70 • C for 20 h.

Preparation of TiO 2 _Citrate Free
The same method was used as described in Section 3.2.1 without using citric acid. The support is used as a reference.
(b) Preparation of 0.4%Pd/TiO 2 _citrate free catalysts (Pd/TiO 2 _citrate free). Pd was deposited on TiO 2 -citrate free from an aqueous solution of Na 2 PdCl 4 by simple impregnation: 0.01 mmol (5 mL) of Pd was mixed with 250 mg of support. Then, this catalyst was reduced by an aqueous solution of NaBH 4 (0.1 mmol), washed with distilled water, and dried at 70 • C.

Characterization Methods
X-ray powder diffractograms were obtained using an Analytical XPERTMPD Pro diffractometer using Cu Kα radiation (λ = 1.542 Å). The Pd loading amount in the catalysts was determined by inductively coupled plasma-atomic emission spectroscopy (ICP-AES) using the Horiba Jobin Yvon instrument. The morphology of Pd nanoparticles was characterized by transmission electron microscopy higher resolution (HR-TEM) with a JEOL 2010 instrument equipped with EDX capabilities. The size and size distribution of Pd were determined using Sigma Scan Pro.5 software. N 2 adsorption-desorption isotherms at 77 K were performed using a Micromeritics ASAP 2010 instrument. The Brunauer-Emmett-Teller (BET) equation was used to calculate the specific surface area (S BET ). The infrared absorption spectra of the synthesized products were recorded using a PerkinElmer Paragon 500 FT-IR spectrometer using a diamond/ZnSe ATR crystal in the region of 4000 to 400 cm −1 . Chemical states of the palladium and carbon in the catalyst surface were investigated by X-ray photoelectron spectroscopy (XPS) on an Axis Ultra DLD (Kratos Analytical) equipped with a dual Al/Mg anode. The spectra were obtained using the Al Kα source (1486.6 eV). All spectra were measured between 0 and 1200 eV pass energy. The constant charging of the different catalysts was corrected by referencing all the energies to the C 1s peak at 284.6 eV arising from adventitious carbon. Elemental carbon analysis is determined using a brand CHNS elemental analyzer Horiba Jobin Yvon model EMIA 220 V.

Catalytic Test
The catalytic activity tests were carried out in a conventional flow reactor under atmospheric pressure at 25 • C. The composition of the reacting gas flow was 15 ppm of O 3 in 1000 mL min −1 of air. Various amounts of powder catalyst (2.5, 5, 10, 20, or 40 mg) were used.
Ozone was generated with a homemade apparatus by flowing pure oxygen through a nonthermal plasma reactor, the existing ozone-oxygen mixture was diluted by air. Water was introduced into the gas flow by bubbling a part of the airflow in a saturator-type vessel containing water prior to being mixed with ozone-air mixture to obtain the humid conditions. The amount of water was fixed by controlling the temperature of the saturator by a cryostat. The relative humidity was fixed at 10%, which corresponds to 3130 ppm of water introduced in the gas flow at 25 • C. Ozone was analyzed by an ozone analyzer (Environment S.A., type "O 3 42 M") based on the UV photometric method at 254 nm [8].
Our analytical system is presented in Figure 9, and ozone concentration was chosen according to the reaction of ozone activation on a catalyst: Catalysts 2022, 12, x 11 of 14 analytical system is presented in Figure 9, and ozone concentration was chosen according to the reaction of ozone activation on a catalyst: where: -mPd is the mass of Pd in the catalytic (g); - Vgas is the total molar flow rate (=1000/(60 × 22,400)) = 7.5 × 10 −4 mol·s −1 ; -XO3 is conversion of O3; -YO3 is the mole fraction of O3 in the in-let gas mixture (80% Ar/20% O2); -MPd is the molecular weight of Pd (106.42 g·mol −1 ); -d is the dispersion of the Pd particles determined by the dynamic pulsed hydrogen chemisorption technique.

Conclusions
Ozone decomposition was studied at 25 °C on Pd catalysts with various textural properties, as well as supported on citrate-modified TiO2 under dry and humid where: -mPd is the mass of Pd in the catalytic (g); - Vgas is the total molar flow rate (=1000/(60 × 22,400)) = 7.5 × 10 −4 mol·s −1 ; -X O3 is conversion of O 3 ; -Y O3 is the mole fraction of O 3 in the in-let gas mixture (80% Ar/20% O 2 ); -M Pd is the molecular weight of Pd (106.42 g·mol −1 ); -d is the dispersion of the Pd particles determined by the dynamic pulsed hydrogen chemisorption technique.

Conclusions
Ozone decomposition was studied at 25 • C on Pd catalysts with various textural properties, as well as supported on citrate-modified TiO 2 under dry and humid conditions. This study clearly showed the importance of surface citrate in the reduction of palladium nanoparticles on the surface of TiO 2 . Depending on the Ti/Cit molar ratio, the particle size of the Pd nanoparticles could be controlled, greatly modifying the activity properties by modifying the electronic properties of the final Pd/TiO 2 _x catalysts. Compared with TiO 2 _citrate free, highly intrinsically active Pd/TiO 2 _x catalysts could be obtained for the decomposition of ozone. These citrate groups on the support surface have shown great persistence to water vapor.
First of all, the activities of the hybrid supports (Cit)(TiO 2 ) x (x = 20, 50, 100) and TiO 2 _citrate free were slightly decreased at the beginning of the reaction and then became stable under dry conditions. However, in the presence of humidity, the activity was clearly sensitive to the presence of water vapor for all supports (around an 8% decrease). The TiO 2 _100 hybrid support, with the lowest amount of citrate, showed a significantly higher efficiency compared with the reference TiO 2 _citrate free support, as well as hybrid supports richer in citrate.
The Pd-doped TiO 2 _x supports were very active in the O 3 decomposition reaction, whatever the operating conditions (humid or dry air catalytic tests). This showed that the presence of Pd had a very beneficial effect in the presence of water vapor. In contrast, the low citrate catalyst (Pd/TiO 2 _100) was very efficient in both the wet and dry air reactions. We have also shown that this catalyst was more effective regardless of the amount of material used for 3 h of reaction. This efficiency was explained by the particle size of Pd, which was smaller for Pd/TiO 2 _100 compared with Pd/TiO 2 _20. Thus, the total elimination of 15 ppm of ozone by this catalytic system under specific experimental conditions (residence time lower than 1.38 ms) allows us to envisage this method for treating indoor and outside air in different settings that could be beneficial to human health and environmental factors.