Photocatalytic Activity in the In-Flow Degradation of NO on Porous TiO 2 –Coated Glasses from Hybrid Inorganic–Organic Thin Films Prepared by a Combined ALD/MLD Deposition Strategy

: A combined ALD/MLD (where ALD and MLD stand for atomic and molecular layer deposition, respectively) deposition strategy using TiCl 4 , H 2 O and HQ (hydroquinone) as precursors has been applied for the preparation of inorganic–organic thin ﬁlms on soda-lime glasses. The alternate deposition of TiO 2 layers, by pulsing TiCl 4 /H 2 O (ALD), and hybrid layers, using TiCl 4 /HQ (MLD), results in the formation of thin ﬁlms that are precursors for porous TiO 2 -coatings after removal of the HQ template by annealing. The coated-glassed show good photocatalytic activity in the degradation of NO with up to 15% reduction of NO concentration in three successive photocatalytic cycles of 5 h each. Surface Scanning Electron Microscopy (SEM) images show that the TiO 2 -coating is composed of large grains that are made up of ﬁner subgrains resulting in a porous structure with an average pore size of 3–4 nm. Transmission Electron Microscopy (TEM) images show two regions, a porous columnar structure on top and a denser region over the glass substrate. Energy Dispersive X-Ray (EDX) analysis, nanocrystal electron diffraction and Raman spectroscopy conﬁrm the presence of the anatase phase, which, together with the porosity of the material, accounts for the observed photocatalytic activity. 180 nm and evidence a columnar structure with the grains elongated along the thickness of the ﬁlm. TEM images at higher magniﬁcation reveal two distinct regions in the ﬁlm: a dense zone ( ≈ 45 nm) was observed at the bottom (near the substrate) and a highly porous zone at the top ( ≈ 125 nm). A thin layer of protective SiO 2 ( ≈ 10nm) is observed below the deposited ﬁlm (Figure 5b). EDX analysis and the electron diffraction pattern conﬁrms that the composition of the coating is TiO 2 and corresponds to the anatase phase.


Introduction
The combustion of fossil fuels in mobile and stationary power sources results in the generation of nitrogen oxides (NO X ), which are key contributors to the photochemical smog and air pollution in urban areas. Photochemical smog consists mainly of tropospheric ozone (O 3 ), peroxyacyl nitrates (PANs), particulate matter (PM), and nitric acid. The latter is formed by the reaction of NO 2 with water, whilst the generation of radicals by the interaction of NO X and sunlight is, to a great extent, the origin of O 3 , PANs and PM. The environmental and health related problems associated with the emissions of NO X 's are diverse, and include (i) acidic depositions, which are harmful to aquatic life, crops and forests, (ii) and adverse effects on human health, which primarily affect the respiratory and cardiovascular systems. In this context, the development of systems that reduce the concentration of NO X 's in the atmosphere is of paramount importance to improve the quality of life in major cities [1][2][3].
Titanium dioxide (TiO 2 ) is a prolific semiconductor photocatalyst [4][5][6][7][8] that has shown activity in the degradation of NO X 's via UV irradiation, since the band-gap of these materi-We report herein on the preparation of glasses coated with a thin film of porous anatase deposited by ALD/MLD and the evaluation of their photocatalytic activity in the inflow degradation of NO(g). In this regard, we have followed two different approaches: (i) MLD deposition of titanicone thin films, and (ii) ALD/MLD deposition of hybrid inorganic-organic superlattice thin films. The presence of porous anatase in the photocatalytic coatings has been confirmed by Raman spectroscopy, X-ray diffraction (XRD) and electron microscopy techniques.

Materials and Methods
The studies were carried out on flat soda-lime glasses of 100 mm× 50 mm × 6 mm (height, width and thickness) manufactured by Guardian coated with a 10 nm thin film of SiO2 deposited by PVD (Physical Vapor Deposition-Magnetron Sputtering) at the Department of Applied Physics of the University of Zaragoza [51]. Previously, the glasses were washed with water and soap and dried on an absorbent paper. The SiO2 coating plays a multifaceted role, as it inhibits the diffusion of alkaline ions from the glass substrate to the photocatalytic coating during the annealing process [45,52], provides thermal stability to the supported TiO2 via the formation of covalent Ti-O-Si bonds [22], while it prevents agglomeration, phase transformation and assist the adsorption process, increasing optical path length [53].
The ALD/MLD depositions were carried out in a PLAY Series Tool (CTECHnano, Coating Technologies S.L, San Sebastián, Spain) using nitrogen (N2) as a carrier gas for the precursors and as a purging gas. Flat soda-lime glasses with 10 nm of SiO2 deposited by sputtering and silicon wafers were used as substrates. Titanicone films [38] (labeled as Titanicone EG) were deposited using titanium tetrachloride (TiCl4; Honeywell Fluka ≥ 99.0%, Madrid, Spain) and ethylene glycol (EG; Sigma-Aldrich 99.8%, St. Louis, MO, USA) as precursors according to the following pulse sequence TiCl4 (0. Hybrid [(TiO2)m(Ti-O-C6H4-O-)k]120 superlattice was deposited using TiCl4 and HQ as a precursors and deionized water (H2O) as a co-reactant, by mixing the deposition process for Titanicone HQ films with the TiCl4/H2O ALD process for TiO2 [54]. To deposit an ALD cycle of TiO2, the precursors were pulsed in the following manner: TiCl4 (0.2 s)/N2 (60 s)/H2O (0.2 s)/N2 (65 s). To reach the precursor vapor pressures required for efficient precursor transport to the substrate, HQ was heated up to 120 °C, whereas TiCl4 and H2O were kept at room temperature. The deposition temperature was set at 210 °C for all of the experiments. Two different structures were prepared. The first one consisted in 20 cycles of TiCl4/H2O following by 500 cycles of TiCl4/HQ. The second one was prepared by mixing m cycles of TiCl4/H2O, with then k cycles of TiCl4/HQ, and repeated the process 120 times in order to fabricate thin films with superlattice structures. Hence, different k:m We report herein on the preparation of glasses coated with a thin film of porous anatase deposited by ALD/MLD and the evaluation of their photocatalytic activity in the in-flow degradation of NO(g). In this regard, we have followed two different approaches: (i) MLD deposition of titanicone thin films, and (ii) ALD/MLD deposition of hybrid inorganicorganic superlattice thin films. The presence of porous anatase in the photocatalytic coatings has been confirmed by Raman spectroscopy, X-ray diffraction (XRD) and electron microscopy techniques.

Materials and Methods
The studies were carried out on flat soda-lime glasses of 100 mm× 50 mm × 6 mm (height, width and thickness) manufactured by Guardian coated with a 10 nm thin film of SiO 2 deposited by PVD (Physical Vapor Deposition-Magnetron Sputtering) at the Department of Applied Physics of the University of Zaragoza [51]. Previously, the glasses were washed with water and soap and dried on an absorbent paper. The SiO 2 coating plays a multifaceted role, as it inhibits the diffusion of alkaline ions from the glass substrate to the photocatalytic coating during the annealing process [45,52], provides thermal stability to the supported TiO 2 via the formation of covalent Ti-O-Si bonds [22], while it prevents agglomeration, phase transformation and assist the adsorption process, increasing optical path length [53].
Hybrid [(TiO 2 ) m (Ti-O-C 6 H 4 -O-) k ] 120 superlattice was deposited using TiCl 4 and HQ as a precursors and deionized water (H 2 O) as a co-reactant, by mixing the deposition process for Titanicone HQ films with the TiCl 4 /H 2 O ALD process for TiO 2 [54]. To deposit an ALD cycle of TiO 2 , the precursors were pulsed in the following manner: TiCl 4 (0.2 s)/N 2 (60 s)/H 2 O (0.2 s)/N 2 (65 s). To reach the precursor vapor pressures required for efficient precursor transport to the substrate, HQ was heated up to 120 • C, whereas TiCl 4 and H 2 O were kept at room temperature. The deposition temperature was set at 210 • C for all of the experiments. Two different structures were prepared. The first one consisted in 20 cycles of TiCl 4 /H 2 O following by 500 cycles of TiCl 4 /HQ. The second one was prepared by mixing m cycles of TiCl 4 /H 2 O, with then k cycles of TiCl 4 /HQ, and repeated the process 120 times in order to fabricate thin films with superlattice structures. Hence, different k:m ratio was effectively used to control the superlattice period, the total number of deposition cycles was obtained as (m + k) × 120. After the deposition process, samples were subjected to thermal annealing in a UNITEMP RSO-200 series furnace (UniTemp, Pfaffenhofen; Germany). The optimized treatment was an instantaneous temperature rise to 250 • C, a temperature ramp to 450 • C in 90 min (2.2 • C/min), and finally 3 h at that temperature.
Film thicknesses were determined by means of X-ray reflectivity (XRR) in a X'Pert PRO PANalytical X-ray diffractometer (Malvern Panalytical, Malvern, UK). Scanning Electron Microscopy (SEM) and High-Resolution Transmission Electron Microscopy (HRTEM) images of the coated-glasses were obtained using eSEM-FEI QuantaTM 250 (and FEI Titan 80-300 TEM microscopes (Fei Company, Hillsboro, USA), respectively, at the Structural-Morphology Characterization Service of CIC NanoGUNE. A thin sample for TEM was prepared using dualbeam instrument FEI Helios 450 with standard lift out and thinning protocols [55]. Raman and UV-Vis spectra were performed at the Chromatography and Spectroscopy Service of the ISQCH. Raman spectra were recorded on a JASCO NRS 3100 spectrophotometer (JASCO, Tokyo, Japan), equipped with a high performance 532 nm laser, and are the result of subtracting the glass signal to the Raman spectra obtained from the coated-glasses. UV-Vis spectra were recorded directly on coated-glass samples on a JASCO V-670 dual beam spectrophotometer (JASCO, Tokyo, Japan), using clean uncovered glass as reference sample. Bandgaps were determined by Tauc plot method [56] using the bandgap calculation subroutine embedded in Jasco Spectra Analysis software (version 1.0). X-ray diffraction measurements were carried out at the X-ray Diffraction and Fluorescence Analysis Service of the Servicio General de Apoyo a la Investigación-SAI, Universidad de Zaragoza. XRD patterns were recorded on a RIGAKU diffractometer (Rigaku Corporation, Tokyo, Japan), model Ru2500, equipped with a rotating Cu anode. The diffractometer operates at 40Kv and 80 mA using a graphite monochromator to select the CuKα radiation. Data were collected in a 2θ range of 20-60 • , step = 0.03 • , t = 3 s/step.
The photocatalytic activity of the coated-glasses was measured in an inert flat-bed photoreactor designed and manufactured by the company Ariño Duglass according to ISO 22197-1:2007 [48]. The experimental setup including the gas supply, photoreactor, and analytical system is shown in Figure 1. The photoreactor consists of a stainless-steel reaction chamber designed to hold 50 cm 2 glass samples (50 mm × 100 mm) of variable thicknesses on Teflon holders equipped with a borosilicate glass window. The irradiation source was a 36 W black light (Philips Actinic BL TL-DK; the wavelengths emitted range from 340 to 400 nm with a maximum at 370 nm) placed over the reactor at 6.5 cm from the sample with an irradiance (incident power/area) of 10 W/m 2 . The reactor is fed with a NO/N 2 certified mixture as polluting gas (120 ppm of NO(g), Nippon Gases) and high purity synthetic air (Nippon Gases). A stream of air was bubbled through a gas wash bottle containing water to control the humidity. The dry air, moist air and N 2 /NO streams were regulated with mass flow controllers (Bronkhorst, F-201CV) and mixed to obtain the desired NO(g) concentration. The system is equipped with a stainless-steel static gas mixer (Koflo Pipe Mixer) before reactor inlet, and temperature and humidity sensors at the reactor inlet and outlet. The gas flows were: 1.50 and 1.47 L/min for dry and wet synthetic air, respectively, and 30 mL/min for the N 2 /NO polluting gas. After a stabilization period, the concentration of NO(g) through the reactor was 1 ± 0.05 ppm with a flow rate of 3 L/min. Note that the stabilization period is the time required to reach a stable concentration of 1 ± 0.05 ppm of NO and a humidity of 50%, since the flows of NO, dry synthetic air and wet synthetic air need to mix correctly before the start of the experiment. NO(g) and NO 2 (g) concentrations in the outlet gas stream from the reactor were continuously monitored by using electrochemical gas sensors (Alphasense Ltd., Braintree, UK, NO-A4 and NO 2 -A4F3). The concentration of both gases, temperature and humidity data were recorded in a computer by using software designed by the Electronic Instrumentation Service (SAI) from the University of Zaragoza for this purpose. The outlet gas stream is bubbled trough a gas wash bottle containing a concentrated NaOH solution. The photocatalytic tests were performed at room temperature with a relative humidity of 50% at 25 °C. The measurements were carried out continuously for 4-5 h with a data collection every 7 s. At the end of the measurement, the coated-glasses were washed with distilled water in order to determinate the NO2 − and NO3 − concentration using a colorimetric indicator from Hasch. A test strip was dipped into the solution obtained by washing the surface with distilled water (2 mL) after the photocatalytic experiments. Positive tests result in a color change which is compared with the chart provided by the manufacturer for the semi-quantitative determination.
Coated-glassed with significant photocatalytic activity were washed with Milli-Q water, dried on an absorbent paper and subjected to successive photocatalytic cycles every 24 h under the same conditions.

MLD of Titanicone EG Thin Films on Glass as Precursors for Photocatalytic Coatings
Titanicone EG thin films on glass were prepared according to the literature procedure [57]. The hybrid inorganic-organic coatings based on titanium-alkoxide were deposited using TiCl4 and ethylene glycol (EG), a relatively low boiling point diol, as precursors using a standard pulse sequence (see above). In parallel, the deposition was performed on silicon wafers (1 cm × 1 cm) to determine the thickness of the coating by X-ray reflectometry (XRR). Under these conditions, the measured film thickness after 200 cycles was about 90 nm, which corresponds to a growth per cycle value (GPC) of 0.45 nm/cycle. Finally, annealing at 400 °C for 2 h in air to remove the organic template and transform the amorphous TiO2 into the anatase crystalline phase resulted in a uniform and transparent coated-glass. The coating thickness was reduced to 30 nm after the thermal treatment, which agrees with the elimination of the organic groups connecting the different titanium centers in the hybrid material.
The photocatalytic activity of this coated-glass in the degradation of NO(g) is shown in Figure 2a. As can be seen, the NO(g) concentration is slightly reduced, by 0.04 ppm, once the illumination is turned on and it is maintained over the 4 h of the experiment until the illumination is turned off (NO2(g) formation was not observed). SEM analysis of surface morphology showed no evidence of porosity suggesting a microporous structure with a pore size of less than 2 nm ( Figure S22). The coating thickness determined by the cross-sectional image was approximately 20 nm, which is of the same order of magnitude The photocatalytic tests were performed at room temperature with a relative humidity of 50% at 25 • C. The measurements were carried out continuously for 4-5 h with a data collection every 7 s. At the end of the measurement, the coated-glasses were washed with distilled water in order to determinate the NO 2 − and NO 3 − concentration using a colorimetric indicator from Hasch. A test strip was dipped into the solution obtained by washing the surface with distilled water (≈2 mL) after the photocatalytic experiments. Positive tests result in a color change which is compared with the chart provided by the manufacturer for the semi-quantitative determination.
Coated-glassed with significant photocatalytic activity were washed with Milli-Q water, dried on an absorbent paper and subjected to successive photocatalytic cycles every 24 h under the same conditions.

MLD of Titanicone EG Thin Films on Glass as Precursors for Photocatalytic Coatings
Titanicone EG thin films on glass were prepared according to the literature procedure [57]. The hybrid inorganic-organic coatings based on titanium-alkoxide were deposited using TiCl 4 and ethylene glycol (EG), a relatively low boiling point diol, as precursors using a standard pulse sequence (see above). In parallel, the deposition was performed on silicon wafers (1 cm × 1 cm) to determine the thickness of the coating by X-ray reflectometry (XRR). Under these conditions, the measured film thickness after 200 cycles was about 90 nm, which corresponds to a growth per cycle value (GPC) of 0.45 nm/cycle. Finally, annealing at 400 • C for 2 h in air to remove the organic template and transform the amorphous TiO 2 into the anatase crystalline phase resulted in a uniform and transparent coated-glass. The coating thickness was reduced to 30 nm after the thermal treatment, which agrees with the elimination of the organic groups connecting the different titanium centers in the hybrid material.
The photocatalytic activity of this coated-glass in the degradation of NO(g) is shown in Figure 2a. As can be seen, the NO(g) concentration is slightly reduced, by 0.04 ppm, once the illumination is turned on and it is maintained over the 4 h of the experiment until the illumination is turned off (NO 2 (g) formation was not observed). SEM analysis of surface morphology showed no evidence of porosity suggesting a microporous structure with a pore size of less than 2 nm ( Figure S22). The coating thickness determined by the cross-sectional image was approximately 20 nm, which is of the same order of magnitude as that determined by XRR (30 nm). The Raman spectrum showed three strong characteristic bands at 639, 513, and 399 cm −1 corresponding to the anatase phase, which is consistent with the observed photocatalytic activity. Interestingly, no Raman bands for the rutile phase were observed. In contrast, a coated-glass with TiO 2 using TiCl 4  ( Figures S16 and S17).

and H 2 O as precursors
The influence of the thermal treatment on the photocatalytic activity of the coatedglasses has been investigated as well. Thus, an inorganic-organic film deposited on glass by MLD under the same conditions as described above was subjected to a different annealing program. First, an instantaneous temperature rise to 250 °C was applied, a temperature ramp to 450 °C in 90 min (2.2 °C /min), and finally 3 h at that temperature. The resulting coated-glass showed a slightly higher photocatalytic activity than the glass treated at 400 °C , with a reduction of the NO(g) concentration of 0.05 ppm over 4 h ( Figure  S2). The slight improvement in the photocatalytic activity prompted us to apply these conditions in the following experiments (optimized annealing conditions). The limited photocatalytic activity of the samples evaluated in Figure 2 is probably a consequence of the lack of porosity. To find the best conditions to prepare porous TiO2 anatase thin films on glass by MLD, we have investigated the influence of increasing: (i) the EG temperature from 70 to 80 °C , (ii) the EG pulse time from 1 to 1.5 s, (iii) the maximum temperature of the heating ramp from 450 to 550 °C during annealing, and (iv) the time at the maximum temperature during annealing (450 °C ) for 5 h (Tables S2-S6). Unfortunately, no significant increase in photocatalytic activity was observed in any case. Nevertheless, increasing the EG pulse time to 1.5 s produces a slight improvement of the photocatalytic activity at the beginning of the experiment although after 2 h the average reduction of the NO(g) concentration is 0.04 ppm, identical to that observed with a time of 1.0 s. On the other hand, increasing the maximum temperature to 550 °C or prolonged heating showed a negative effect. This activity drop is likely due to the collapse of the porous structure, which is pivotal to achieve catalytic activity in the photodegradation of NO, although anatase-to-rutile phase transformation cannot be discarded.
Finally, two coated-glasses have been prepared in an attempt to maximize the EG transfer to the substrate. In the first one, the EG pulse time was increased to 4.0 s, and in the second, two consecutive EG pulses of 4.0 s each were programmed, keeping in both cases identical deposition conditions (EG temperature of 80 °C ) and the optimized annealing treatment. The first of the coated-glass showed an initial decrease in NO(g) concentration to 0.89 ppm and NO2(g) formation of 0.02 ppm, although the photocatalytic activity decays with time and stabilizes at 0.97 ppm of NO(g) after 2 h (Figure 2b). The second coated-glass showed no photocatalytic activity, which may be due to the fact that the addition of two consecutive EG pulses results in a fragile structure that collapses during the annealing process. The influence of the thermal treatment on the photocatalytic activity of the coatedglasses has been investigated as well. Thus, an inorganic-organic film deposited on glass by MLD under the same conditions as described above was subjected to a different annealing program. First, an instantaneous temperature rise to 250 • C was applied, a temperature ramp to 450 • C in 90 min (2.2 • C/min), and finally 3 h at that temperature. The resulting coated-glass showed a slightly higher photocatalytic activity than the glass treated at 400 • C, with a reduction of the NO(g) concentration of 0.05 ppm over 4 h ( Figure S2). The slight improvement in the photocatalytic activity prompted us to apply these conditions in the following experiments (optimized annealing conditions).
The limited photocatalytic activity of the samples evaluated in Figure 2 is probably a consequence of the lack of porosity. To find the best conditions to prepare porous TiO 2 anatase thin films on glass by MLD, we have investigated the influence of increasing: (i) the EG temperature from 70 to 80 • C, (ii) the EG pulse time from 1 to 1.5 s, (iii) the maximum temperature of the heating ramp from 450 to 550 • C during annealing, and (iv) the time at the maximum temperature during annealing (450 • C) for 5 h (Tables S2-S6). Unfortunately, no significant increase in photocatalytic activity was observed in any case. Nevertheless, increasing the EG pulse time to 1.5 s produces a slight improvement of the photocatalytic activity at the beginning of the experiment although after 2 h the average reduction of the NO(g) concentration is 0.04 ppm, identical to that observed with a time of 1.0 s. On the other hand, increasing the maximum temperature to 550 • C or prolonged heating showed a negative effect. This activity drop is likely due to the collapse of the porous structure, which is pivotal to achieve catalytic activity in the photodegradation of NO, although anatase-to-rutile phase transformation cannot be discarded.
Finally, two coated-glasses have been prepared in an attempt to maximize the EG transfer to the substrate. In the first one, the EG pulse time was increased to 4.0 s, and in the second, two consecutive EG pulses of 4.0 s each were programmed, keeping in both cases identical deposition conditions (EG temperature of 80 • C) and the optimized annealing treatment. The first of the coated-glass showed an initial decrease in NO(g) concentration to 0.89 ppm and NO 2 (g) formation of 0.02 ppm, although the photocatalytic activity decays with time and stabilizes at 0.97 ppm of NO(g) after 2 h (Figure 2b). The second coated-glass showed no photocatalytic activity, which may be due to the fact that the addition of two consecutive EG pulses results in a fragile structure that collapses during the annealing process.

Photocatalytic Coatings on Glass from Hybrid Inorganic-Organic Superlattice Thin Fims
Klepper and co-workers have described the preparation of alucones, inorganic-organic hybrid films based on trimethylaluminum (TMA). A range of aromatic and aliphatic dicarboxylic acids have been used for the preparation of alucone coatings on soda-lime glass and silicon single crystals [57,58] However, hybrid inorganic-organic thin films prepared by MLD using TiCl 4 and terephthalic acid (benzene-1,4-dicarboxylic acid) under different deposition conditions resulted, after the optimized annealing treatment, in coated-glasses that showed no photocatalytic activity in the degradation of NO(g).
Recently, Knez and co-workers have described the preparation of alucones using hydroquinone (HQ) as an organic template [59]. Therefore, the preparation of inorganicorganic hybrid coatings using TiCl 4 and hydroquinone (HQ) could be possible. However, the small size and rigidity of HQ likely limits its ability to induce porosity. Therefore, we decided to follow the methodology reported by Karppien and co-workers [54], who developed a combined ALD/MLD deposition strategy to prepare inorganic-organic thin films using TiCl 4

Photocatalytic Coatings on Glass from Hybrid Inorganic-Organic Superlattice Thin Fims
Klepper and co-workers have described the preparation of alucones, inorganic-organic hybrid films based on trimethylaluminum (TMA). A range of aromatic and aliphatic dicarboxylic acids have been used for the preparation of alucone coatings on soda-lime glass and silicon single crystals [57,58] However, hybrid inorganic-organic thin films prepared by MLD using TiCl4 and terephthalic acid (benzene-1,4-dicarboxylic acid) under different deposition conditions resulted, after the optimized annealing treatment, in coated-glasses that showed no photocatalytic activity in the degradation of NO(g).
Recently, Knez and co-workers have described the preparation of alucones using hydroquinone (HQ) as an organic template [59]. Therefore, the preparation of inorganic-organic hybrid coatings using TiCl4 and hydroquinone (HQ) could be possible. However, the small size and rigidity of HQ likely limits its ability to induce porosity. Therefore, we decided to follow the methodology reported by Karppien and co-workers [54], who developed a combined ALD/MLD deposition strategy to prepare inorganic-organic thin films using TiCl4, H2O and HQ as precursors. Thus, the combination of hybrid  The first coating was prepared in two steps. Initially, TiCl4 and H2O were pulsed under the specified conditions (20 cycles) to deposit a TiO2 layer. Subsequently, a monolayer of the hybrid film was deposited using TiCl4 and HQ (500 cycles) (Figure 3a). The resulting coating has a thickness of 84 nm, which was reduced to about 30 nm (XRR) after the optimized annealing treatment. The obtained coated-glass is transparent and shows a uniform coating. However, the glass showed virtually no photocatalytic activity, i.e., no significant decrease of the NO concentration has been observed under irradiation ( Figure S10).
In view of this result, it was decided to explore a more homogeneous alternating The first coating was prepared in two steps. Initially, TiCl 4 and H 2 O were pulsed under the specified conditions (20 cycles) to deposit a TiO 2 layer. Subsequently, a monolayer of the hybrid film was deposited using TiCl 4 and HQ (500 cycles) (Figure 3a). The resulting coating has a thickness of 84 nm, which was reduced to about 30 nm (XRR) after the optimized annealing treatment. The obtained coated-glass is transparent and shows a uniform coating. However, the glass showed virtually no photocatalytic activity, i.e., no significant decrease of the NO concentration has been observed under irradiation ( Figure S10).
In view of this result, it was decided to explore a more homogeneous alternating deposition sequence. First, TiCl 4 /H 2 O were used as precursors (2 cycles) and subsequently TiCl 4 /HQ (4 cycles), repeating this sequence 120 times following the deposition conditions described in Section 2 ( Figure 3b). The coated-glass obtained after the optimized annealing treatment is also transparent and shows a uniform coating over the entire surface. However, in contrast with the previous sample, the film thickness after the thermal treatment increases from 25 to 46 nm (XRR), which could be indicative of increased porosity induction. The transparent coated-glass fabricated under these conditions showed photocatalytic activity. As can be seen in Figure 4a, initially there is a decrease in NO(g) concentration to 0.75 ppm and NO 2 (g) formation to 0.07 ppm and then, the photocatalytic activity steadily stabilizes at 0.90 ppm of NO(g) and 0.04 ppm of NO 2 (g) over the remaining 4 h of experiment. The coated-glass was subjected to successive photocatalytic cycles by washing several times with distilled water and dried on absorbent paper after each cycle. Although the photocatalytic activity is maintained in the second photocatalytic cycle, a slight decrease was observed in the third cycle ( Figure S11). Interestingly, analysis of the washings with test strips showed the presence of NO 2 − and NO 3 − which is consistent with the observed photocatalytic activity.
Coatings 2022, 12, x FOR PEER REVIEW 8 of 12 described in Section 2 ( Figure 3b). The coated-glass obtained after the optimized annealing treatment is also transparent and shows a uniform coating over the entire surface. However, in contrast with the previous sample, the film thickness after the thermal treatment increases from 25 to 46 nm (XRR), which could be indicative of increased porosity induction. The transparent coated-glass fabricated under these conditions showed photocatalytic activity. As can be seen in Figure 4a, initially there is a decrease in NO(g) concentration to 0.75 ppm and NO2(g) formation to 0.07 ppm and then, the photocatalytic activity steadily stabilizes at 0.90 ppm of NO(g) and 0.04 ppm of NO2(g) over the remaining 4 h of experiment. The coated-glass was subjected to successive photocatalytic cycles by washing several times with distilled water and dried on absorbent paper after each cycle. Although the photocatalytic activity is maintained in the second photocatalytic cycle, a slight decrease was observed in the third cycle ( Figure S11). Interestingly, analysis of the washings with test strips showed the presence of NO2 − and NO3 − which is consistent with the observed photocatalytic activity. Figure 4. Photocatalytic activity in the degradation of NO(g) of coated-glasses formed by alternating inorganic (TiO2) and inorganic-organic (TiCl4/HQ) layers after optimized thermal treatment: (a) layers deposited at 2/4 ratio (120 cycles), (b) layers deposited at 2/8 ratio (120 cycles). Photocatalytic activity of the last coated-glass in the second (c) and third photocatalytic cycle (d).
After this successful experiment we decided to explore a similar alternating deposition sequence increasing the frequency of the TiCl4/HQ pulses. Thus, alternate layers of inorganic (TiO2) and hybrid (Ti/HQ) films have been deposited using TiCl4/H2O (2 cycles) and TiCl4/HQ (8 cycles) as precursors, repeating this sequence 120 times, under the conditions specified above. The coated-glass fabricated after annealing under optimized conditions showed higher photocatalytic activity. Initially, the amount of NO(g) decreases to 0.64 ppm and stabilizes at 0.85 ppm NO(g) and 0.08 ppm NO2(g) along the experiment (Figure 4b). The pronounced decrease in NO(g) concentration at the beginning of the experiment compared to the previous glass is a diagnostic of increased coating porosity which is consistent with the improved photocatalytic activity. It should be noted that this After this successful experiment we decided to explore a similar alternating deposition sequence increasing the frequency of the TiCl 4 /HQ pulses. Thus, alternate layers of inorganic (TiO 2 ) and hybrid (Ti/HQ) films have been deposited using TiCl 4 /H 2 O (2 cycles) and TiCl 4 /HQ (8 cycles) as precursors, repeating this sequence 120 times, under the conditions specified above. The coated-glass fabricated after annealing under optimized conditions showed higher photocatalytic activity. Initially, the amount of NO(g) decreases to 0.64 ppm and stabilizes at 0.85 ppm NO(g) and 0.08 ppm NO 2 (g) along the experiment (Figure 4b). The pronounced decrease in NO(g) concentration at the beginning of the experiment compared to the previous glass is a diagnostic of increased coating porosity which is consistent with the improved photocatalytic activity. It should be noted that this coated-glass showed stable photocatalytic behavior, since the activity is preserved in successive photocatalytic cycles. Thus, the NO(g) concentration stabilizes at 0.87 ppm with NO 2 (g) formation of 0.07 ppm along the second and third photocatalytic cycles (Figure 4c,d). This coated-glass is iridescent with two zones with different light behavior, one violet and the other bluish. The Raman spectra of both zones are very similar showing the presence of anatase, bands at 639, 513, and 399 cm −1 . However, the intensity of the Raman bands is different, suggesting that this effect could be a consequence of the difference in thickness of the coating. In addition, the presence of the anatase TiO 2 phase was also confirmed by the XRD diffraction pattern ( Figure S20).
Finally, we explored the effect of increasing the frequency of the TiCl 4 /H 2 O pulses. A deposition sequence of TiCl 4 /H 2 O (4 cycles) and TiCl 4 /HQ (2 cycles) was applied and repeated 120 times. The coated-glass prepared under these conditions showed very little photocatalytic activity with a NO(g) concentration stabilization value of 0.98 ppm throughout the experiment. Therefore, decreasing the relative amount of HQ has a negative impact on the photocatalytic activity, which confirms the key role played by hydroquinone in the generation of porous structures.
SEM and TEM images of the coated-glass exhibiting best photocatalytic performance ( Figure 4b) are shown in Figure 5 and in Figures S23-S25 and S27. Surface SEM images show that the film is composed of grains (≈30-150 nm) (Figure 5a) which are built of finer subgrains resulting in a porous structure with an average pore size of 3-4 nm (min. 0.8 nm, max. 13 nm). Low magnification cross-sectional TEM images show a film thickness of about ≈ 180 nm and evidence a columnar structure with the grains elongated along the thickness of the film. TEM images at higher magnification reveal two distinct regions in the film: a dense zone (≈45 nm) was observed at the bottom (near the substrate) and a highly porous zone at the top (≈125 nm). A thin layer of protective SiO 2 (≈10nm) is observed below the deposited film (Figure 5b). EDX analysis and the electron diffraction pattern confirms that the composition of the coating is TiO 2 and corresponds to the anatase phase. , x FOR PEER REVIEW 9 of 12 coated-glass showed stable photocatalytic behavior, since the activity is preserved in successive photocatalytic cycles. Thus, the NO(g) concentration stabilizes at 0.87 ppm with NO2(g) formation of 0.07 ppm along the second and third photocatalytic cycles ( Figure  4c,d). This coated-glass is iridescent with two zones with different light behavior, one violet and the other bluish. The Raman spectra of both zones are very similar showing the presence of anatase, bands at 639, 513, and 399 cm −1 . However, the intensity of the Raman bands is different, suggesting that this effect could be a consequence of the difference in thickness of the coating. In addition, the presence of the anatase TiO2 phase was also confirmed by the XRD diffraction pattern ( Figure S20). Finally, we explored the effect of increasing the frequency of the TiCl4/H2O pulses. A deposition sequence of TiCl4/H2O (4 cycles) and TiCl4/HQ (2 cycles) was applied and repeated 120 times. The coated-glass prepared under these conditions showed very little photocatalytic activity with a NO(g) concentration stabilization value of 0.98 ppm throughout the experiment. Therefore, decreasing the relative amount of HQ has a negative impact on the photocatalytic activity, which confirms the key role played by hydroquinone in the generation of porous structures.
SEM and TEM images of the coated-glass exhibiting best photocatalytic performance ( Figure 4b) are shown in Figure 5 and in Figures S23-S25 and S27. Surface SEM images show that the film is composed of grains (30-150 nm) (Figure 5a) which are built of finer subgrains resulting in a porous structure with an average pore size of 3-4 nm (min. 0.8 nm, max. 13 nm). Low magnification cross-sectional TEM images show a film thickness of about  180 nm and evidence a columnar structure with the grains elongated along the thickness of the film. TEM images at higher magnification reveal two distinct regions in the film: a dense zone (45 nm) was observed at the bottom (near the substrate) and a highly porous zone at the top (125 nm). A thin layer of protective SiO2 (10nm) is observed below the deposited film (Figure 5b). EDX analysis and the electron diffraction pattern confirms that the composition of the coating is TiO2 and corresponds to the anatase phase. The highly porous TiO2 anatase layer on top of the coating accounts for the observed photocatalytic activity. However, the appearance of a dense non-porous TiO2 layer at the bottom of the coating was completely unexpected. We have fabricated a coated-glass under the same deposition conditions increasing the time at maximum temperature (400 °C ) in the thermal treatment from 3 to 5 h. The UV-Vis spectra of both coated-glasses showed a strong absorption at 312 nm in the UV region which is attributed to the presence of TiO2. The estimated bandgaps values using the Tauc method were 3.35 and 3.39 eV, respectively, which are in the expected range for photocatalytic coatings ( Figure S21). The spectra also showed a second absorption band around 368 nm. As the intensity of this absorption is lower in the second sample, it could be tentatively attributed to incomplete removal of the The highly porous TiO 2 anatase layer on top of the coating accounts for the observed photocatalytic activity. However, the appearance of a dense non-porous TiO 2 layer at the bottom of the coating was completely unexpected. We have fabricated a coated-glass under the same deposition conditions increasing the time at maximum temperature (400 • C) in the thermal treatment from 3 to 5 h. The UV-Vis spectra of both coated-glasses showed a strong absorption at 312 nm in the UV region which is attributed to the presence of TiO 2 . The estimated bandgaps values using the Tauc method were 3.35 and 3.39 eV, respectively, which are in the expected range for photocatalytic coatings ( Figure S21). The spectra also showed a second absorption band around 368 nm. As the intensity of this absorption is lower in the second sample, it could be tentatively attributed to incomplete removal of the hydroquinone template. However, this coated glass showed significantly lower photocatalytic activity ( Figure S14) which might be related to the collapse of the porous structure. Since the deposition conditions did not change throughout the different cycles, we hypothesized that this dense layer was formed during the annealing process.

Conclusions
We have developed a convenient ALD/MLD methodology for the deposition of porous TiO 2 -coatings on soda-lime glass that show activity in the photocatalytic oxidation of nitrogen oxide. Initially, we studied the deposition of hybrid inorganic-organic coatings using TiCl 4 and EG by MLD, which leads to a maximum NO removal of 4% over 4 h. A study of the annealing process on these samples reveals that long thermal treatments result in the loss of catalytic activity, likely due to the collapse of the porous structure. Subsequently, we explored the preparation of TiO 2 -coatings by the deposition of inorganicorganic superlattice thin films. The best performing coating was prepared by deposition of a TiO 2 layer by pulsing TiCl 4 /H 2 O (2 cycles), followed by hybrid film using TiCl 4 and HQ (8 cycles), repeating this sequence 120 times. The resulting material was able to reduce the NO concentration ca. 15% in three successive photocatalytic cycles of 5 h each. The TEM and SEM images of this TiO 2 -coating show two regions, a porous columnar structure on top, and a denser region immediately over the thin layer of SiO 2 deposited by PVD. EDX analysis, nanocrystal electron diffraction and Raman spectroscopy confirm the presence of the anatase phase, which, together with the porosity of the material, accounts for the observed photocatalytic activity.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/coatings12040488/s1. Tables S1-S14: MLD and ALD/MLD thin film deposition and annealing conditions. Figures S1-S14: In-flow monitoring of the photocatalytic degradation of NO(g) by coated-glasses fabricated by MLD and ALD/MLD strategies. Figure S15: Photograph of the experimental setup. Figures S16-S19: Raman spectra of selected coated-glasses. Figure