Photocatalytic Decomposition of N₂O by Using Nanostructured Graphitic Carbon Nitride/Zinc Oxide Photocatalysts Immobilized on Foam

Abstract

The aim of this work was to deposit cost-effective g-C₃N₄/ZnO nanocomposite photocatalysts (weight ratios of g-C₃N₄:ZnO from 0.05:1 to 3:1) as well as pure ZnO and g-C₃N₄ on Al₂O₃ foam and to study their photocatalytic efficiency for the photocatalytic decomposition of N₂O, which was studied in a home-made batch photoreactor under ultraviolet A irradiation (λ = 365 nm). Based on the photocatalysis measurements, it was found that photocatalytic decomposition of N₂O in the presence of all the prepared samples was significantly higher in comparison with photolysis. The photoactivity of the investigated nanocomposite photocatalysts increased in the following order: g-C₃N₄/ZnO (3:1) ≈ g-C₃N₄/ZnO (0.45:1) ≤ g-C₃N₄/ZnO (2:1) ZnO < g-C₃N₄ < g-C₃N₄/ZnO (0.05:1). The g-C₃N₄/ZnO (0.05:1) nanocomposite showed the best photocatalytic behavior and the most effective separation of photoinduced electron–hole pairs from all nanocomposites. The key roles played in photocatalytic activity were the electron–hole separation and the position and potential of the valence and conduction band. On the other hand, the specific surface area and band gap energy were not the significant factors in N₂O photocatalytic decomposition. Immobilization of the photocatalyst on the foam permits facile manipulation after photocatalytic reaction and their repeated application.

1. Introduction

Nitrous oxide is one of the most important greenhouse gases and represents 6.4% of the total global radiative forcing. Its contribution to global warming is important, because it is nearly 300 times higher than that of carbon dioxide [1]. N₂O is also a significant contributor to the destruction of the ozone layer in the stratosphere. The increase of N₂O emissions in the atmosphere (approximately 0.3% per year) is caused particularly by anthropogenic activities [1]. Synthetic nitrogen fertilizers in agriculture, combustion (fossil fuel, biomass), and the production of adipic acid and nitric acid are among the main contributors to N₂O emissions. The decomposition of nitrous oxide into N₂ and O₂ offers a simple method for its transformation to natural air elements. Decomposition of N₂O under UV irradiation on semiconductor materials (Equation (1)) is one possible solution for its removal:

$$N_{2}O\stackrel{h\nu , catalyst}{\to }{N}_2+\frac{1}{2}{O}_2.$$

(1)

Photocatalytic decomposition of nitrous oxide was studied on Cu ions embedded on various oxides, e.g., (SiO₂·Al₂O₃, Al₂O₃, SiO₂) [2], ZnS [3], TiO₂ [4], TiO₂ doped by Ag [5,6,7], Ce [8,9].

Graphitic carbon nitride (g-C₃N₄) is an n-type semiconductor with a unique structure and useful optical, electric, and physicochemical properties. Although g-C₃N₄ has attracted increasing interest worldwide, its photocatalytic performance is still limited due to fast recombination of charge carriers [10,11,12,13]. Enhancing the photoactivity of g-C₃N₄-based photocatalysts can be achieved by the construction of semiconductors’ heterojunctions [13,14]. The combination of different semiconductor photocatalysts may create new materials with optimal performances. The formation of unique semiconductor–semiconductor junctions can enhance the electron–hole separation for improved photocatalytic efficiency and extend the spectral range for light absorption. Looking into the mechanisms, one can identify that four dissimilar models of semiconductor–semiconductor junctions have been utilized to enhance the charge kinetics, including semiconductor sensitization, Type II, phase junction, and Z-scheme [15,16]. In our previous work, heterojunction photocatalysts (g-C₃N₄/TiO₂ and g-C₃N₄/WO₃ photocatalysts with various weight ratios) were prepared and studied for N₂O photocatalytic decomposition [17,18,19]. The formation of heterojunctions contributed to the increase of the N₂O conversion.

Recent investigations on semiconductors’ photocatalysis are oriented to photocatalyst immobilization [16]. The immobilization of photocatalysts removes any imperfections with the powder form of the photocatalysts: (i) The difficulties of separating photocatalysts, (ii) the inclination of the nanoparticles of photocatalysts to aggregate, and (iii) the difficulty of applying them to continuous flow reactors [20].

For those reasons, we decided to prepare g-C₃N₄/ZnO nanocomposites with different weight ratios to immobilize them on Al₂O₃ foam, and to investigate them for N₂O photocatalytic decomposition under UVA (365 nm) irradiation. The g-C₃N₄/ZnO photocatalysts have not yet been studied for N₂O photocatalytic decomposition.

2. Results and Discussion

2.1. Photocatalysts Characterization

The real content of ZnO in prepared samples was determined by atomic absorption measurements (

Table 1. Structural and microstructural properties and content of ZnO ZnO, g-C₃N₄, and g-C₃N₄/ZnO.

Photocatalysts	SBET (m2/g)	V net (mm3lig/g)	Content of Zn2+ (wt.%)
ZnO	7	61	80
g-C3N4/ZnO (0.05:1)	8	62	77
g-C3N4/ZnO (0.45:1)	37	212	55
g-C3N4/ZnO (2:1)	58	298	26
g-C3N4/ZnO (3:1)	68	339	21
g-C3N4	85	413	0

). The nitrogen physisorption characterization technique was used for the specific surface area determination (Figure 1). Each photocatalyst displayed a mesoporous-macroporous character. From the comparison of textural properties of the prepared samples, it is evident that the g-C₃N₄ addition positively affected the properties of the photocatalyst’s porous structure, significantly enhancing its specific surface area as well as the net pore volume, e.g., pure ZnO showed a specific surface area of 7 m²/g and a net pore volume of 61 mm³_liq/g, while g-C₃N₄ /ZnO (3:1) showed a specific surface area of 68 m²/g and a net pore volume of 339 mm³_liq/g (Table 1).

The X-ray powder diffraction (XRD) patterns of all investigated photocatalysts are shown in Figure 2. Two different phases based on zinc, zinc oxide, and zinc cyanamide were observed in all mixtures with higher contents of g-C₃N₄. The zinc oxide phase was markedly more significant than the zinc cyanide phase. Graphitic carbon nitride was confirmed by the presence of the reflection at position 14.9° 2Theta (d = 0.695 nm) indexed as the (100) plane and attributed to the in-plane tri-s-triazine units forming one-dimension melon strands. The most typical reflection was found around 32.2° 2Theta (d = 0.322 nm) and could be interpreted as long range interplanar stacking of the aromatic systems with the hkl plane (002). Unfortunately, this reflection was overlapped with the zinc cyanamide reflection (211) in all measured samples.

For the obtained g-C₃N₄/ZnO photocatalysts, nearly all of the characteristic diffraction peaks of zinc oxide were observed in the XRD pattern and confirmed the presence of the ZnO phase. In the case of the g-C₃N₄ phase, the (002) peak of g-C₃N₄ was almost not observed. The relatively poor crystallinity of the g-C₃N₄ phase indicates that simultaneous crystallization of the ZnO phase interferes with the condensation process and interlayer stacking patterns of the g-C₃N₄ phase, as shown by its weak diffraction intensity [21]. The XRD patterns in g-C₃N₄/ZnO photocatalysts indicate that these two phases occur separately in the composite and they do not have strong crystal lattice incorporation. This result was also confirmed by transmission electron microscopy (TEM).

All phases in the XRD patterns are marked by different symbols (Figure 2). All phases were identified using the PDF-2 database issued by the international center diffraction data. Zinc oxide corresponds to PDF-2 card no. 01-080-0075 and zinc cyanamide to PDF-2 card no. 01-070-4898. All XRD patterns were recorded using co-radiation with wavelength lambda = 1.78 Å. For this reason, all peaks were shifted to the higher 2Theta angles (based on the Braggs equation) than Cu radiation, which is more common.

The UV–vis diffuse reflectance spectra of g-C₃N₄, ZnO, and g-C₃N₄/ZnO photocatalysts are compared in Figure 3a. The band gaps’ energy values of the prepared nanocomposites were estimated by plotting the Kubelka–Munk function, as shown in Figure 3b. The largest band gap was 3.09 eV and it belongs to pure ZnO. On the other hand, pure g-C₃N₄ had a band gap energy of 2.64 eV. The nanocomposites band gap energies were 3.00, 2.75, 2.75, and 2.75 eV for g-C₃N₄/ZnO (0.05:1), g-C₃N₄/ZnO (0.45:1), g-C₃N₄/ZnO (2:1), and g-C₃N₄/ZnO (3:1), respectively. The presence of g-C₃N₄ in the nanocomposites extended the light absorption of the composites into the visible region compared to pure ZnO photocatalyst.

The infrared spectra are seen in Figure 4. The spectra of samples g-C₃N₄, g-C₃N₄/ZnO (3:1), g-C₃N₄/ZnO (2:1), and g-C₃N₄/ZnO (0.45:1) primarily contain bands that are characteristic for graphitic carbon nitride [22,23,24]. The regions are labelled as N1, N2, and N3. The region N1 contains bands at 3280, 3255, 3160, and 3090 cm⁻¹. These bands can be assigned to the stretching vibration of N-H bonds. The region N2 contains intensive bands at 1680 (shoulder), 1635, 1575, 1545, 1460, 1430 (shoulder), 1410, 1335 (shoulder), 1320, 1280, (shoulder), 1240, and 1210 cm⁻¹. Characteristic bands of the skeletal “breathing” vibration of aromatic rings of nitrogen heterocyclic compounds and stretching vibration bonds between carbon and nitrogen were found in the N2 region. The N3 region contains only a medium intensive band at 815 cm⁻¹, which is due to the breathing mode of the triazine unit. Moreover, the spectra of samples g-C₃N₄/ZnO (3:1), g-C₃N₄/ZnO (2:1), and g-C₃N₄/ZnO (0.45:1) also contain bands of complex cyanide compounds at 2085 and 2045 cm⁻¹ [25]. The bands have a medium or weak intensity (see region CN). The relative red shifts of the complex cyanide compounds compared with the normal wavenumber of 2250 cm⁻¹ are due to the conjugation effect of the extended C-N aromatic system [21]. The spectra of samples g-C₃N₄/ZnO (0.05:1) and ZnO only contain medium or weak bands of carbonates (see region C1 and C2) and a broad band of zinc oxide (see region Z). The carbonate bands can be assigned to asymmetric stretching vibration of carbonate (1450 cm⁻¹) and symmetric deformation vibration of carbonate (885 cm⁻¹), respectively [25]. The band of vibration of zinc oxide [26] is seen at 480 cm⁻¹. This band is seen also in the spectrum of sample g-C₃N₄/ZnO (0.45:1).

The Raman spectra are seen in Figure 5. The spectra are in the “full-scale” mode, thus the presented spectra have different scales of intensity axes. The spectra of samples g-C₃N₄, g-C₃N₄/ZnO (3:1), g-C₃N₄/ZnO (2:1), and g-C₃N₄/ZnO (0.45:1) only contain the bands characteristic for graphitic carbon nitride (i.e., [22,24,27]) in the regions labelled N1, N2, N3, N4, and N5. The N1 region contains strong bands at 1230 and 1215 cm⁻¹ and weak bands at 1150 and 1115 cm⁻¹. The N2 region contains a medium band at 770 cm⁻¹ and a strong band at 705 cm⁻¹. The N3 region contains two medium bands at 485 and 470 cm⁻¹ and one weak band at 455 cm⁻¹. Medium bands at 355 and 210 cm⁻¹ are in the N4 and N5 regions, respectively. The assignment of the above-mentioned Raman bands of graphitic carbon nitride has not been published yet. We can suppose the bands are related to vibration of the tri-s-triazine ring (i.e., [27]). The spectra of samples g-C₃N₄/ZnO (0.05:1) and ZnO contain two bands of carbonate (see region C1 and C2) and three bands of zinc oxide (see region Z1, Z2, and Z3). The carbonate bands are presented at 1080 cm⁻¹ (symmetric stretching vibration) and at 700 cm⁻¹ (symmetric deformation vibration), respectively [25]. The bands characteristic for zinc oxide are presented at 435, 330, and 185 cm⁻¹.

Figure 6a–d demonstrate the TEM images of pure ZnO, g-C₃N₄, and g-C₃N₄/ZnO (3:1) photocatalysts, respectively. The ZnO nanoparticles (Figure 6a) have hexagonal shapes and the particle size is about 50 nm while pure g-C₃N₄ (Figure 6b) shows a 2D lamellar structure. The TEM micrographs of g-C₃N₄/ZnO (3:1) (Figure 6c) testify that the ZnO nanoparticles were not supported well on the surface of g-C₃N₄. It is possible to infer that the salt (from preparation) can be observed in the g-C₃N₄/ZnO samples (Figure 6d). It is also evident from the results of the EDS analysis (Figure 6e), which confirmed not only the presence of Zn, C, O, and N elements but also Na element. The presence of the salt was also confirmed by Raman and infrared spectroscopy.

Photoelectrochemical measurements of all photocatalysts are shown in Figure 7. It is evident that the magnitude of the generated current decreased with an increasing amount of g-C₃N₄ presented at the photocatalysts until the g-C₃N₄ was in the majority. Higher photocurrents measured in the presence of the ZnO sample show that there are significantly more charge carriers produced in ZnO compared to g-C₃N₄. The photocurrent decreased with an increasing amount of g-C₃N₄, which formed a significantly smaller photocurrent.

2.2. Photocatalytic Decomposition of N₂O

The influence of prepared nanocomposites on the photocatalytic decomposition of nitrous oxide was studied from 0 to 22 h (Figure 8). The photolysis was also conducted in order to confirm the photocatalytic activity of the prepared samples. The time dependences of the nitrous oxide conversion of all g-C₃N₄/ZnO nanocomposites, including pure ZnO and g-C₃N₄ photocatalysts, are depicted in Figure 8. It is evident that each prepared photocatalyst proved to be much more effective in comparison with photolysis. The g-C₃N₄/ZnO (0.05:1) and pure g-C₃N₄ and ZnO showed a significantly higher efficiency than other photocatalysts.

The highest N₂O conversion under UVA irradiation was reached over the g-C₃N₄/ZnO (0.05:1) photocatalyst. The N₂O conversion in the presence of this photocatalyst was 11% after 22 h of UVA irradiation. For comparison, the N₂O conversion without photocatalyst was only 6% after 22 h of UVA irradiation. The same foam with photocatalyst was used for repeated tests three times. This confirms the stability and ability to reuse these materials without losing photocatalytic activity.

It is hard to compare the observed photocatalytic results with the literature, as it is usually measured under different reaction conditions (geometry of photoreactor, irradiation source, concentration of N₂O, amount of photocatalyst, whether it is immobilized, and also the irradiation time). Each of these conditions influences the photocatalytic decomposition of N₂O in a different way and therefore accurate comparison is impossible. Also, not many research groups have investigated the photocatalytic decomposition of N₂O. Therefore, the obtained results were compared with our previous works. N₂O conversion for the g-C₃N₄/BiVO₄ and g-C₃N₄/WO₃ composites after 20 h of irradiation was 12.5% and 14.5%, respectively. However, both previous sets of photocatalysts were measured in a different photoreactor (different geometry) and therefore, from that point of view, the observed N₂O conversion of approximately 10% after 20 h of irradiation seems to be comparable to the previous results.

2.3. Proposed Mechanism for Photocatalytic Activity Enhancement

When g-C₃N₄ is combined with ZnO, the generated electrons (e⁻) and holes (h⁺) can migrate by two possible mechanisms: Type II heterojunction and the Z-scheme system. In the case of the type II heterojunction of g-C₃N₄/ZnO, the generated e^- in the conduction band (CB) of g-C₃N₄ can transfer to the CB of ZnO and simultaneously, the h⁺ in the valence band (VB) of ZnO can migrate to the VB of g-C₃N₄. In the case of the Z-scheme system of g-C₃N₄/ZnO, e⁻ in the CB of ZnO are directly recombined with the h⁺ in the VB of g-C₃N₄ at the heterojunction interface, and so the recombination of e⁻ and h⁺ in ZnO and g-C₃N₄ is inhibited. In this case, the e^- are accumulated in the CB of g-C₃N₄ and h⁺ in the VB of ZnO. Due to these two mechanisms, a prolonged lifetime of electrons and holes is gained [14].

The band edge positions of materials were calculated via Equations (2) and (3) [28]:

$$E_{VB}=\chi -E_0+0.5E_g,$$

(2)

$$E_{CB}=E_{VB}-E_g,$$

(3)

where χ is the electronegativity of the material, E_VB is the VB potential, E_CB is the CB potential, E_g is the band gap energy of a material, and E₀ is the energy of free electrons (4.5 V vs. NHE). The χ values for g-C₃N₄ and ZnO are 4.72 and 5.95 eV, respectively [28]. The E_VB of g-C₃N₄ and ZnO were 1.75 and 3.05 eV, respectively. The E_CB of g-C₃N₄ and ZnO were −1.31 and −0.15 V, respectively.

The formation of the Z-scheme system of g-C₃N₄/ZnO can be confirmed by the hydroxyl radical (•OH) trapping test using modified photoluminescence (PL) measurements [29]. After the photoexcitation, the h⁺ with sufficient oxidation potential can be either directly involved in the photocatalytic degradation reactions or generate active species. The reaction between terephthalic acid (TA) and hydroxyl radicals (•OH) can generate hydroxyterephthalic acid (HTA), which is detectable by the PL spectrometer (signal with maximal intensity centered at 425 nm) [30]. The PL peak intensity of the HTA is in proportion to the amount of produced •OH.

Since the photoluminescence intensity increased in the presence of each photocatalyst over time, with the exception of pure g-C₃N₄, the presence of hydroxyl radicals was confirmed (Figure 9a). According to the band structures of ZnO and g-C₃N₄, the photoinduced h⁺ on g-C₃N₄ cannot oxidize the adsorbed H₂O molecules to •OH because the VB potential of g-C₃N₄ (+1.75 V vs. NHE) is less positive than the standard redox potential of •OH/OH⁻ (2.4 V vs. NHE). In the case of Z-scheme formation, the photogenerated h⁺ remain on the ZnO. The h⁺ accumulation on ZnO is in accordance with the migration mechanism of the direct Z-scheme. The creation of the Z-scheme system was observed by Wang et al. [14], Liu et al. [31], and Yu et. al. [21]. In our case, it was not confirmed that g-C₃N₄/ZnO nanocomposite had contact with the Z-scheme system because the nanocomposites were not proven to have a higher PL intensity than pure ZnO. The PL intensity decreased with an increasing content of ZnO, which can produce hydroxyl radical.

According to the literature, the standard reduction potential of N₂O/N₂ is +1.355 V [32]. The conduction bands’ potentials of both pure and prepared photocatalysts (g-C₃N₄ and ZnO) are more negative (−1.19 and −0.17 V, respectively) than the standard reduction potential of N₂O/N₂. Therefore, both pure photocatalysts showed similar photocatalytic activity.

The photocatalytic activity of g-C₃N₄/ZnO, with the exception of g-C₃N₄/ZnO (0.05:1), was lower with regard to pure photocatalysts. These three nanocomposites (g-C₃N₄/ZnO (0.45:1), g-C₃N₄/ZnO (2:1), and g-C₃N₄/ZnO (3:1)) had not only similar photocatalytic activity but also band gap energy. The photocatalytic activity of nanocomposites (g-C₃N₄/ZnO (0.45:1), g-C₃N₄/ZnO (2:1), and g-C₃N₄/ZnO (3:1)) decreased both by the presence of cyanide compounds formed during preparation (Figure 4) and by the imperfect contact of g-C₃N₄ and ZnO, which is visible in Figure 6c. The detection of cyanide compounds in g-C₃N₄/ZnO indicates that simultaneous ZnO crystallization could interfere with the thermal polymerization of melamine to generate a well-defined g-C₃N₄ phase. Based on the previous discussion about XRD patterns and infrared spectra, such interference in ZnO crystallization can result in relatively poor crystallinity and give rise to defect states within the g-C₃N₄ phase [21]. These defect states can catch generated charge carriers; however, they are not accessible for the adsorbed molecule of nitrous oxide, which consequently leads to a decrease in photoactivity.

A slightly different situation is the case of g-C₃N₄/ZnO (0.05:1), which proved to have the highest photocatalytic activity. This nanocomposite contained only 4% of g-C₃N₄, which means most of the holes in VB of g-C₃N₄ probably recombined with some of the electrons from the CB of ZnO. In the meantime, the photogenerated electrons in the CB of g-C₃N₄ and a big part of the electrons in the CB of ZnO, which was not used for recombination with holes of g-C₃N₄, reduced N₂O to N₂ and •O⁻. The photoexcited holes in the VB of ZnO oxidized •O⁻ to O₂ [19]. This result shows that only a small loading of g-C₃N₄ in g-C₃N₄/ZnO can increase the effectiveness of photogenerated electron–hole separation.

In heterogeneous photocatalysis, on the contrary to catalysis, the specific surface (S_BET) is not the most crucial parameter. It is obvious in this work, where g-C₃N₄/ZnO (0.05:1) and g-C₃N₄ had a similarly high photoactivity though the S_BET of g-C₃N₄ was more than 10 times higher than in the case of g-C₃N₄/ZnO (0.05:1). Many authors have quoted the band gap energy as one of the most important parameters. The photoelectrochemical measurements validated the capability of all nanocomposites to absorb UVA. Moreover, both pure photocatalysts, which had similar photoactivity, have different band gap energy. Consequently, the band gap energy is not the most decisive parameter. The most significant factor, which is valid in photocatalytic reactions generally, is the position and potential of the valence and conduction band.

3. Materials and Methods

3.1. Preparation of ZnO, g-C₃N₄, and g-C₃N₄/ZnO Nanocomposites

ZnO was prepared by precipitation [33]. G-C₃N₄ was formed by heating of melamine [18,34]. In a synthesis procedure of g-C₃N₄/ZnO, the determined amounts of ZnO and g-C₃N₄ were dispersed in deionized water and evenly mixed (4 h). Subsequently, the mixture was dried (at 60 °C in air) and calcined [18]. The weight ratios of prepared g-C₃N₄/ZnO photocatalysts were set at 0.05:1, 0.45:1, 2:1, and 3:1, respectively. The prepared nanocomposites (0.1 g) were anchored on a commercial support VUKOPOR (Lanik Ltd., Boskovice, Czech Republic) [35,36]. For detailed preparation information, see the Supplementary Materials.

3.2. Characterization of g-C₃N₄/ZnO Nanocomposite and Photocatalytic Decomposition of N₂O

The products were characterized in detail by N₂ physisorption (Micromeritics), X-ray powder diffraction (XRD, Rigaku, Tokyo, Japan), UV–Vis diffuse reflectance spectra (DRS, IRS-2600Plus), transmission electron microscopy (TEM, JEOL 2100 equipped with an EDS detector, Tokyo, Japan), Fourier transform infrared spectroscopy (FT-IR, Nexus 470 ThermoScientific, Waltham, MA, USA), Raman spectroscopy (DXR SmartRaman (ThermoScientific), photoluminescence (PL) measurements (FLSP920 Series, Edinburgh Instruments, Livingston, UK), and photoelectrochemical measurements (Instytut Fotonowy, Krakow, Poland), and their photocatalytic activity was studied for the photocatalytic decomposition of N₂O according to the reported method [18,29]. The photocatalytic inactivity of the bare foam was confirmed via blank tests and compared to the blank test without foam.

4. Conclusions

g-C₃N₄/ZnO nanocomposites with different ratios of components as well as pure ZnO and g-C₃N₄ were prepared with success. Each photocatalyst was coated on commercial Al₂O₃ foam. The prepared nanocomposites were investigated for N₂O photocatalytic decomposition for the first time. g-C₃N₄ addition positively affects the properties of the photocatalysts’ porous structure, significantly enhancing its specific surface area as well as the net pore volume. In all mixtures with higher contents of g-C₃N₄, two phases based on zinc were observed, -zinc oxide (markedly more significant) and zinc cyanamide. ZnO nanoparticles have hexagonal shapes and pure g-C₃N₄ shows a 2D lamellar structure. Based on the experimental and characterization results, N₂O conversion in the presence of photocatalysts was higher in comparison with photolysis; in the case of the best photocatalysts, g-C₃N₄/ZnO (0.05:1) was almost twice higher. FT-IR characterization showed that samples of the g-C₃N₄/ZnO (3:1), g-C₃N₄/ZnO (2:1), and g-C₃N₄/ZnO (0.45:1) photocatalysts also contained cyanide compounds, which together with an imperfect contact of g-C₃N₄ and ZnO caused a decrease of photocatalytic activity. The key role played in photocatalytic activity was the electron–hole separation and the position and potential of the valence and conduction band. On the other hand, the specific surface area and band gap energy were not significant factors in N₂O photocatalytic decomposition. The immobilization on commercial Al₂O₃ foam proved to be promising. A short-term test confirmed the photocatalysts’ stability and reproducible activity. Immobilization might be an efficient way to bring photocatalysis closer to an application, since the powder form is not usable for several reasons, such as filtration from the liquid phase or blowing off by the gas stream. No such drawbacks exist with the immobilized photocatalyst on the foam.