Photocatalytic degradation of hydrocarbons and methylene blue using floatable titanium dioxide catalysts in contaminated water

Mineral oil-based pollution of water bodies is a problem worldwide today. The sources of pollution are diverse and include roadways, accidents, pollution from shipping, etc. Polluted water can be found in rainwater retention areas, docks, and, ultimately, in the natural water resources. Additionally, anthropogenic trace substances, such as pharmaceuticals, additives from personal care products, industrial chemicals, and pesticides, have been the focus of urban water management research for some time (Bergmann et al. 2011). Particularly in the context of further wastewater treatment, these substances and the process-technical possibilities for their elimination from water are considered.

Separation and/or degradation of contaminants from water is possible through physical, chemical, and biological means. While biological approaches rely on natural self-purification capabilities of the water body, this may take a considerable amount of time and results may be unsatisfactory (Weichgrebe et al. 1991). Physical and chemical methods tend to be more cost-intensive but offer faster results than biological methods (Pinnekamp et al. 2010). Precipitation, flocculation, disinfection, radiation, reverse osmosis, electro-coagulation, and absorption by activated charcoal are conventional methods (Pinnekamp et al. 2010).

Flexible and decentralized solutions are especially required in the case of accidents involving the release of large quantities of mineral oil hydrocarbons. Furthermore, increasing demand and parallel scarcity of freshwater and increasing pollution of water bodies, especially in developing countries and areas with insufficient infrastructure, create the need for a sustainable low-cost alternative for water clean-up. Particularly in warm regions, which receive high solar radiation throughout the year, a possible answer could be the use of solar-powered photocatalytic purification.

When irradiated by sunlight, titanium dioxide forms highly reactive oxygen species that can disintegrate organic molecules to their mineral components (Pedanekar et al. 2020) and are potentially lethal to microbes and can even destroy viruses (Cabiscol 2000). Several important features for the heterogeneous photocatalysis have extended its feasible applications in water treatment, such as ambient operating temperature and pressure, complete mineralization of organic substances, their derivate and decomposition products without secondary pollution, as well as low operating costs (Chong et al. 2010). It is well-documented (Mehrjouei et al. 2015) that highly reactive oxygen species cause cell lysis, generated as a result of the photo-induced charge separation on titanium dioxide (TiO₂) surfaces, which leads to microbial inactivation and organic mineralization without creating secondary pollution. The most important reactive oxygen species are hydroxyl radicals, which are formed by the reaction of hydroxide-ions and water molecules with the electron defects in the semiconductor (Equations (3) and (4)) (Irawaty et al. 2014). Superoxide anion radicals can be formed as further oxidizing agents in the electron band (Equation (5)). The electron defects themselves are also strong oxidizing agents, which can react directly with the organic compounds to be broken down (Equation (2)) (Vargas & Núñez 2010).

The degradation takes place through the reaction between C–H bonds with the reactive oxygen species under the formation of radical residuals of the hydrocarbons, which are so unstable that they disintegrate further on their own (Fox & Dulay 1993). The principal reaction between the hydroxyl radicals and the organic pollutant is shown in Equation (2). The well-known kinetics of the heterogeneous catalysis defines the step of the pollutant adsorption as the step of the kinetics which decides the velocity of the reaction. Therefore, a rough surface structure of the titanium dioxide coating will help to accelerate the disintegration of the pollutant (Didier et al. 2013).

Nevertheless, there are still a series of technical challenges for the application of TiO₂ catalysts to water treatment. The post-separation of the semiconductor TiO₂ catalyst after water treatment remains the major obstacle to its practicality as an industrial process (Barka et al. 2014). The fine particle size of the TiO₂ along with their large surface area-to-volume ratio and surface energy creates a strong tendency for catalyst agglomeration during the operation. Such agglomeration of the particles is highly detrimental given the particles' size preservation, surface-area reduction, and its reusable lifespan (Chong et al. 2010).

Using the catalyst designs on a fixed surface has the advantage that only a small amount of TiO₂ powder is necessary and only a small fraction ends up in the water column, thus not requiring elaborate separation. The reduction of free nanoparticles is a notable ecological benefit over the use of loose TiO₂ powder nanoparticles, which harm aquatic life forms (Adams et al. 2006; Haynes et al. 2017; Li et al. 2017). Photocatalytic disintegration with carrier-bound titanium dioxide catalysts provides a practical, low-cost approach for dealing with organic pollution originating from petrol-derived hydrocarbons (Cazoir et al. 2012), pharmaceuticals and their derivates (Schnabel et al. 2020), and even pathogens while possessing self-cleaning capabilities (Cabiscol 2000). The catalysts can be left on the water, utilizing a floating or fixed frame to hold them in place. Constant and predictable solar radiation is required, and no further energy use is needed. This simple low-technology method does not need trained personnel and is advantageous in acquisition and operation. These floating catalysts could be used to break down various types of pollution on the water surface with the help of solar radiation. Mineral oil-based pollution in particular that floats on the surface is a worthwhile cleaning goal for this approach.

Therefore, the goal of this study is to design and present floatable titanium dioxide-catalyst materials, which have a high surface contact with non-polar hydrocarbons that also float on the surface of the waterbody. The idea of a floatable sun driven catalyst (Qiu et al. 2019) was already described with bismuth-based materials (Qiu et al. 2019; Nasir et al. 2020). In this study, the goal was to investigate the suitability of three different types of floatable titanium dioxide catalysts (foam glass, glass fiber, and steel grid), provided by the Laboratories for Material Testing and Research (MFPA) Weimar, to disintegrate polar and non-polar substances in water. Ideally, these catalysts can be placed on polluted water bodies (such as water reservoirs and rainwater basins) or used as an additional cleaning stage in sewage treatment plants (Schnabel 2016; Mayer 2017), purifying the water by removing organic pollutants like drugs and their derivates and pathogenic germs (Cabiscol 2000). In the experiments, diesel fuel served as a substitute for non-polar pollutants such as petrol-derived hydrocarbons, while the pigment methylene blue was used to investigate the degradability of polar organic substances.

A series of experiments was conducted for each type of catalyst. To simulate solar radiation, an actively cooled array of UV-LEDs manufactured by Synantik (Ohrdruf, Thuringia, Germany) was used. The array has a power rating of 8 W with an efficiency of approximately 8% at 385 nm from electrical to radiometric power. Conventional 200 mL aluminum trays from the food industry with a resulting water surface area of 70 cm² were used as containers for each experiment. Therefore, the specific irradiation power was 91.4 W/m². Each aluminum tray was filled with 100 mL demineralized water, into which 100 μL (0.082 g) of diesel fuel was injected beneath the water surface. Subsequently, the fuel rose to the surface and spread evenly. The coated glass foam was put in the tray until the water surface was covered. The other photocatalytic materials (coated steel grid and glass fiber tissue) were attached to wood pieces, which helped the material float on the surface. The UV-array was placed over one of the trays at 1.5 cm on a UV light-transmissive glass panel. During the experiments, the reactors were not magnetically stirred. Various experiments with different radiation times were executed, and a new preparation for every single experiment was applied. Within the closed reactor system, there was a gas phase present, so that evaporation of volatile components occurred. To exclude the effect of evaporation, two setups were used. One was exposed to UV light, while the other was shielded from it. This allowed us to estimate the influence of evaporation and distinguish it from the effect of photocatalytic disintegration. Figures 1–3 show the floating materials.

The extraction of the remaining hydrocarbons was performed using cyclohexane. A total of 20 mL of eluent was poured into the tray, and a sample was drawn after 15 min of stirring to be analyzed via GC-FID (Agilent ChemStation 6890N and Agilent 7683B Injector). This resulted in a complete sampling of all hydrocarbons within the reactor system including adsorbed and undissolved components. The experiments using coated glass fiber and steel grit were conducted in the same manner, with the exception of the glass fibers, which were provided with a floating frame. The grits were folded such that they would remain approximately 1 mm under the water surface.

Field tests were conducted in Weimar, Germany (50°N, 11°E) at 28 °C and under sunny weather on the 9 August 2017 to show the performance of the three catalysts for polar organic compounds and to include real sunlight. The experiments were performed to compare the activity of the different materials in a homogenous solution. Methylene blue (20 μmol/L) was used as a representative of water-soluble polar substances. Discoloration and thus disintegration were quantified by a ‘Merck SQ 118’ photometer. The extinction was measured with a 1 cm quartz glass cuvette at a wavelength of 620 nm. The extinction has a linear relation to the methylene blue concentration, so the concentration was calculated using a 10-point calibration curve between 1 and 20 μmol/L methylene blue.

The degradation of mixtures of mineral oil hydrocarbons, which are present as a separate phase, is further significantly influenced by the specific solubility and vapor pressures of their individual compounds. Mineral oil hydrocarbons are not miscible with water but are soluble in water to a small extent. Diesel, as a representative of the mineral oil hydrocarbons, belongs to the middle distillates, which are characterized by a boiling range of 160–390 °C and a chain length of 9–24 C atoms (Cho et al. 2006). The total water solubility of diesel as a mixture of mineral oil hydrocarbons is determined by the respective water solubility of the individual components. Furthermore, in the case of mineral oil mixtures, there is a mutual influence of the individual compounds, which reduces their individual solubility. With increasing chain length, a decrease of the solubility can be recognized. With the same number of C atoms, a decrease in solubility is evident in the following series: aromatics > alkenes > cycloalkanes > i-alkanes > n-alkanes. The total solubility of diesel can be classified as low and amounts to 5–20 mg/L (Speight 2020). The volatility of the individual compounds behaves analogously, whereby a decrease in volatility occurs to a greater extent with increasing molecule size compared with solubility. According to Speight (2020), the total vapor pressure of diesel mixtures is 0.05 kPa, whereby mixtures of substances above 0.01 kPa are classified as volatile.

In this equation, is the amount of mineral oil which could not be disintegrated further during the time of the experiment. is the maximum concentration which could be disintegrated, is the exponential disintegration factor, and t is the time in hours. The kinetics show that the percentile disintegration does not depend on the concentration, but the reaction velocity depends on the concentration (Equation (9)). Within this formula, the total decrease in diesel is described by an exponential disintegration factor. As a result, both photocatalytic degradation and evaporation are represented by the disintegration factor. The influence of the evaporation was covered by an analogous data processing of the test series without the use of a catalyst. Furthermore, the photocatalytic degradation within the experiments is influenced by the mass transfer within the reactor system as well as between mineral oil and water phase, since no devices for the reactor stirring were included. For methylene blue as a water-soluble substance, the term equals zero.

Three different photocatalytic materials were used for the experiments. All three materials were coated with a layer of titanium dioxide nanoparticles, which were sinter-fused on the hot material surface out of a waterish aerosol suspension. The load of titan dioxide was calculated so that it would be 30 g/m² in all experiments. The three materials used were glass foam with a diameter between 1 and 2 mm and glass fiber tissue and stainless-steel tissue with 200 μm mesh. The steel grit is made of V4A high-grade stainless steel (316 L/1.4404). An aqueous nanoparticle suspension made from anatase-modified titanium dioxide nanoparticles of 14 nm in size was used as the coating. The anatase content of the titanium dioxide was 90%, and the BET surface area of the TiO₂ particles was 50 m²/g. The aqueous suspension was sprayed onto the heated carrier material (150–200 °C surface temperature). The fast evaporation of the solvent creates a fractal coating structure with a high inner surface (Figures 1–3). The concentration of the aqueous suspension was 10 m% titanium dioxide. The scanning electron microscope (SEM) images (Figures 1–3) show the different materials before and after the coating with the nanoparticle suspension. As can be seen in the transmission electron microscope (TEM) images, the coating of the glass tissue and the steel grit is very homogeneous while the coating of the glass foam is not as even and needs to be improved in the future to reach better photocatalytic properties. The following images show the floatability of the glass foam pearls (Figure 4(a) and 4(b)) and the other materials with the wooden floaters (Figure 4(c)).

In this study, the disintegration induced by the photocatalysis of organic substances with titanium dioxide in the presence of UV light was investigated. The disintegration of UV light-exposed and non-exposed diesel fuel with different forms of catalyst was measured via gas chromatography as well as photometer measurements for methylene blue decoloring. For the evaluation of the measurements, an exponential curve fit was using the described kinetics for the disintegration of the mineral oil and methylene blue.

With the floating foam glass design, the concentration of diesel fuel could be lowered in 68 h from 688 to 271 mg/L; this equals the disintegration of 417 mg/L or 60.6%. In the experiment without UV light, the concentration decreased through evaporation of the volatile mineral oil compounds from 688 to 418 mg/L over the same time; this equals evaporation of 270 mg/L or 39%. Figure 5 shows the results as a concentration of mineral oil hydrocarbons over time. The contrast between radiated and non-radiated experiments is clearly visible throughout the tests: it is always higher in the non-radiated experiments. The calculated difference of non-radiated and radiated values in the arrangements indicates the effect of photocatalytic disintegration.

The experiments using diesel fuel were conducted with artificial UV light from LEDs because the effect of evaporation proved to be influential in an open-air setup with natural solar radiation and heating, thus corrupting the outcome. The analysis of the FID chromatograms in Figure 6 shows that the smaller compounds disintegrate faster by hydroxyl radicals than the long-chained aliphatic compounds of the mineral oil. This is due to lower volatility, solubility, and possibly lower photocatalytic reaction rates of longer-chain components. These effects overlap within the selected experimental setup, which is why this question cannot be solved completely by the existing results.

This was also shown in the fluorescent behavior of the mineral oil in the experiments after 2 h of irradiation that the mineral oil film under the photocatalytic materials was not fluorescent anymore. This provides strong evidence that especially the short aromatic compounds in the mineral oil (benzene and toluene) disintegrate faster than others. Figure 7 shows the irradiated and unirradiated samples after 2 h.

The glass fiber material has, in comparison, the highest surface area. The problem remains that this material does not float by itself and needs floaters of wood to stay close to the surface of the water. Otherwise, the contact of the catalyst and the mineral oil is not given, which may decrease reaction rate for low turbulence applications. Regardless of those problems, a higher and faster disintegration of mineral oil was achieved in comparison to the experiment without UV light (Figure 8). In 16 h, the evaporation was 189 mg/L (23% of the initial concentration). In the experiment with UV light, a disintegration of 591 mg/L (71% of the initial concentration) was achieved. The same effect regarding fluorescing of the mineral oil like with the glass foam could be observed. Additionally, the chromatograms were equal.

The experiment with the steel grid materials showed the smallest difference between evaporation and photocatalytic disintegration (Figure 9). The experiments showed higher variation between values, which is reflected by R². Presumably, insufficient adsorption of diesel fuel on the catalyst surface led to inconsistent measurements. One hypothesis is that this effect occurs due to the greater distance of the catalyst surface from the mineral oil phase. However, glass wool, which is also a non-floatable material, showed a very high degradation rate for diesel fuel. It seems more likely that the adsorption of petroleum hydrocarbons is significantly reduced due to the electrochemical properties of steel-based catalyst carriers.

To compare all experimental setups, Table 1 provides the functions and a combination of evaporation and disintegration after 16 h. The results of the comparison showed that the difference between evaporation and photocatalytic experiment is 329 mg/L for the glass foam, 401 mg/L for the glass fiber materials, and 55 mg/L for the steel grid. This means that the glass fiber material achieved the highest photocatalytic disintegration rate, and the steel grid achieved almost no photocatalytic activity with the disintegration of mineral oil.

Thus, the decomposition of diesel fuel with coated foam glass and glass fiber proved to be promising, while the design of the coated steel grid was less efficient. Using immobilized TiO₂, a solar reactor with collectors and groundwater samples with similar hydrocarbon concentrations but a mixture of different hydrocarbons (Cho et al. 2006) achieved only a degradation of 10% after 4 h. When switching to 0.2% TiO₂ slurry 80% of the hydrocarbons could be degraded after 4 h. In these experiments, the measured solar irradiation power per surface was 16 W/m², which is lower than the value chosen in this paper. However, due to the use the reactor designs, a higher specific irradiation surface of 0.55 m²/L was created. The volume-specific irradiation power of 8.8 W/L is of a similar order of magnitude to the selected experimental design (6.37 W/L). The higher degradation rate of the hydrocarbons (Cho et al. 2006) can be explained by the use of groundwater samples which only contain the soluble contents of mineral oil hydrocarbons. In conclusion, it is shown that the floating, carrier-bound catalysts can achieve nearly the same reaction velocity for the degradation of mineral oil as particulate material.

The experiments with methylene blue depletion were conducted with real sunlight in sunny weather conditions, as direct solar radiation had – in contrast to the experiments with diesel fuel – no observable effect on dye concentration via evaporation. As shown in Figure 10, glass foam and steel grit catalysts could disintegrate the pigment more efficiently and in a shorter time than the glass foam. The control of uncoated glass foam had no observable effect on the dye. The attainable degradation was much higher than in the experiments with diesel fuel.

Decoloring of methylene blue solution for steel grit yielded up to 99.1% decoloring after 4 h, for glass fiber 96.6% after 6 h, and for glass foam 56.4% after 6 h. The trend line fits well with the values measured with exponential decay of the first order (except foam glass without TiO₂), which diverges from it only to a small extent as shown by its correlation (R²). The catalysts investigated showed a higher effect on the polar substance methylene blue than on the diesel fuel. Foam glass could disintegrate over 50% of methylene blue after 6 h, while only foam glass was able to mineralize 47% of diesel fuel concentration after 16 h. Glass fiber and steel grit could diminish methylene blue concentration by 96.6% after 4 h 99.1% after 6 h, respectively. Given that the experiments with diesel and glass fiber/steel grit produced inconsistent results, the outcome of the methylene blue examinations, in contrast, showed that the capability of the catalysts relies mostly on adequate adsorption and contact of the organic compounds to make them available for photocatalysis. The non-polar nature of hydrocarbons and the polarity of dissolved methylene blue may also play an important role during adsorption. Table 2 provides the exponential functions of the methylene blue experiment as well as the half-life time for the different materials.

Within the scope of this work, it could be demonstrated that carrier-bound TiO₂ catalysts can achieve similar degradation rates in their application as floating catalysts to those achieved by particulate catalyst material. However, the investigated catalysts show a slightly reduced reaction rate, which in practical applications should be compensated by a much easier handling. The treatment of mineral oil phases in larger water bodies represents a promising application of this technology. In the carried out experiments, the floating design of foam glass coated with titanium dioxide could reduce the concentration of diesel fuel by 329 mg/L in 16 h; the submerged designs for coated glass fiber and coated steel grit could reduce methylene blue concentration by 96.6% after 4 h and 99.1% after 6 h, respectively. In relation to an application for mineral oil degradation, the question arises as to how the behavior of dissolved mineral oil hydrocarbons in large water bodies with high dilution potentials and higher turbulence is influenced by the positioning of the catalysts. It is assumed that the distance between catalyst and mineral oil phase has a significant influence on the degradation rate. Given this context, the use of self-floatable catalysts, such as glass foam as a carrier material, seems to be promising. The conducted experiments have shown that this material is generally suitable but shows a lower activity. An optimization of the coating process is, therefore, recommended.

For the degradation of diesel fuel, the experiments further showed a limitation to the maximum amount accessible to photocatalytic degradation. Photocatalytic degradation requires a dissolved state for the compounds to be oxidized. For longer-chain mineral oil components, this is only given to a very limited extent. It should be noted that highly insoluble and correspondingly less mobile substances usually have an equally low ecotoxicological impact potential. It would also be feasible to counteract this effect by using solubilizers (e.g. surfactants). In addition, a loss of function is to be feared in the permanent reactor operation due to the retention of a highly viscous, almost inert mineral oil phase. With regard to the development of an effective technical application, the investigation of permanent modes of operation represents an essential step and should possibly be taken into account within future research projects.

A comparison of the tests with mineral oil and methylene blue revealed significant differences in the activity of the catalysts investigated. The catalysts with steel grids as carrier material showed almost no activity for diesel fuel, while they degraded methylene blue at a very high rate. The catalysts with glass foam and glass fiber as carrier material showed the same ranking in activity for both groups of substances. In addition to varying solubility, the investigated substance groups differed substantially with respect to their molecular charge. Catalysts fixed on conductive stainless steel appear to show an affinity for the degradation of polar substances. An increased positive potential on the catalyst would cause an increased adsorption of polar substances and thus explain their higher reaction rate according to the Langmuir–Hinshelwood kinetic. Furthermore, by discharging excited electrons to a counter electrode, the recombination rate of the electron-hole sites is reduced, which can increase the quantum efficiency of the catalyst. Increased degradation rates of photoelectric catalysis using an external voltage source are described in Song et al. (2019); Collivignarelli et al. (2020); Wang et al. (2020). It is also known that electrical potentials are generated by excitation of the semiconductor material titanium dioxide (Sclafani & Herrmann 1996; Li & Shen 2006; Suhadolnik et al. 2016). The influence of the carrier material on this electrochemical system has not yet been sufficiently investigated for carrier-based photocatalytic applications. Within this work, overlapping effects of solubility and electric charge were found, which is why an exact quantification of each process is not possible. Therefore, there is a need for further research on the influence of the carrier materials on the obtainable reaction rates for different groups of substances.

Apart from the main objective of the study, it can be stated that carrier-bound solar photocatalysis offers an energy-neutral and low-maintenance option for the oxidation of dissolved substances. While in most applications it is not possible to achieve a continuously operating treatment stage due to varying solar radiation, this technology could provide a supporting purification stage which, in the case of solar irradiation, allows energy and more cost-intensive conventional processes to be relieved. For these applications, the floatability of the catalysts only influences the absorption of solar radiation by the water layer above the catalyst. Due to the higher activity, the use of glass wool or steel grit catalysts, equipped with floating frames or fixed installations, seems to be reasonable. An application idea would be to use the free surfaces of existing post clarification plants for solar photocatalysis as a supplementary measure for the elimination of micro pollutants. The use of carrier-bound catalysts as a sole purification stage for micro pollutants or other organic compounds is also possible. In this case, an artificial irradiation source is required to achieve a permanent and efficient system operation. UV-LEDs have proven to be an efficient source of UV light and electricity consumption.

All relevant data are included in the paper or its Supplementary Information.