Degradation of a Toxic Pollutant (Pyridine) in a Flat-Bed Reactor Using Ultraviolet Light: Comparison of Two Catalysts (ZnO-Au and TiO₂-Au)

1. Introduction

Currently, there is less access to water suitable for human consumption, which is since more fresh water is being contaminated by discharges of effluents contaminated with different types of substances harmful to living beings [1,2].

Industrial development and the creation of new products are causing this type of pollution, putting conventional treatment systems at risk, since they only manage to eliminate dissolved organic matter and some nutrients, but persistent organic pollutants are not eliminated, and these are incorporated into streams, rivers or bodies of fresh water [3,4,5].

This can cause public health problems since many of these emerging contaminants tend to cause various issues, including skin, eye, and nose allergies, and even different types of cancer [6,7,8].

Emerging pollutants involve a wide variety of chemical compounds, which are mainly produced by industries and released into effluents in different concentrations [8,9,10].

Among these emerging contaminants is pyridine. Pyridine is a basic heterocyclic organic compound with the chemical formula C₅H₅N. It is structurally related to benzene, with a methyl group (=CH−) substituted by a nitrogen atom (=N−). It is a highly flammable, weakly alkaline liquid, miscible in water, with a characteristic and unpleasant fishy odor [12,13,14].

Pyridine is colorless, but older or impure samples may appear yellow. The pyridine ring is found in numerous commercial compounds, including agrochemicals, pharmaceuticals, and vitamins. Historically, pyridine was produced from coal tar. In 2016, its global synthesis reached approximately 20,000 tons per year [15,16].

Prolonged exposure to pyridine can cause damage to the liver, heart, and kidneys [17,18,19]. There is evidence that it is a carcinogen in animals, and from these studies, there is sufficient evidence that pyridine is possibly carcinogenic to humans (Group 2B).[20]

Therefore, it is essential to eliminate these types of pollutants from industrial effluents before they reach freshwater bodies. Advanced oxidation treatments exist to remove persistent organic pollutants. These are classified as photochemical and non-photochemical. Photochemical treatments share the common characteristic of producing the hydroxyl radical (OH), which is a strong oxidant, second only to fluorine, which is a much stronger oxidizing radical than the hydroxyl radical [21,22].

The radical oxidizes pyridine to simpler organic compounds and, in some cases, degrades it to complete oxidation. However, in others, it produces intermediate compounds that are far more polluting than the original compound. Therefore, it is appropriate to monitor photodegradation reactions using the total organic carbon technique to achieve complete mineralization of the contaminant and the intermediate compounds formed during the photodegradation process [22,23,24].

Currently, most research focuses on creating active catalysts under natural light (sunlight or white light lamps), which reduces treatment costs. Research also focuses on the design of different types of reactors, including fluidized bed reactors, continuous reactors, bath reactors, suspended catalysts, and catalysts fixed on supports [25,26,27,28].

Heterogeneous photoreactors are the most useful reactors for wastewater treatment, where photocatalysts can be present in suspension (sludge reactor) or immobilized on a transparent support (fixed-bed reactor). Most early photoreactors employed a TiO₂ suspension due to its large surface area for reactions. Disadvantages of sludge photocatalysis include: (1) the difficulty and time required to separate or filter the photocatalyst after the photocatalytic process; (2) particle aggregation and clumping at high photocatalyst concentrations; and (3) the difficulty of using the photocatalyst in suspension in continuous processes [29,30].

To overcome these drawbacks, immobilized photocatalysts are often recommended. Photocatalysts can be immobilized on various supports, such as glass [31,32], silica [33], polymers [34,35] and clays [36].

This work proposes the use of a flat-bed reactor where the catalyst was supported on clay bricks. The catalyst was modified to be active under white light lamps or natural light [35].

2. Materials and Methods

2.1. Chemicals Products and Reagents

The reagents used for the synthesis of the catalysts were: Titanium dioxide (TiO₂) Degussa P25, Zinc oxide (ZnO), Gold chloride trihydrate (HAuCl₄.3H₂O), and silver chloride (AgCl), which were purchased from Sigma-Aldrich. Double-distilled water, purchased from Fermont, was also used.

For the photocatalytic experiments, the following reagents were used: Pyridine (C₅H₅N), purchased from Sigma-Aldrich, and distilled water, purchased from Fermont.

The reagents used to carry out the high-performance liquid chromatography (HPLC) analyses were methanol (CH₄O), purchased from J. T. Baker; sodium 1-hexensulfanate (CH₃(CH₂)₅SO₃Na), obtained from Fischer Sc; and phosphoric acid, also from J.T. Baker.

2.2. Description of the Falling Film Reactor

The degradation of pyridine was carried out in a falling-film photoreactor (Figure 1). This system is a metal structure consisting of a metal plate coated with polyester resin. Clay plates were placed on this plate to support the catalysts. At the top of the plate is a polymer pipe with small holes through which water flows in a cascading manner, creating a film that descends onto the clay plates. This pipe is fed by a 50-psi pump, which recirculates the solution to be analyzed from a 10 L container. Above the plate (where the plates were placed) is a metal structure with a 90° opening to which the UV radiation lamps were attached.

2.2.1. Characteristics of the Plates

To support the catalyst on a porous surface, ceramic plates made of clay, measuring 30 cm wide by 30 cm long and 2 cm thick, were used. These plates are composed of cement, sand, water, and clay.

2.2.2. Plate Treatment

Prior to impregnating the catalysts and metallic nanoparticles, the ceramic plates underwent treatment. They were manually sanded to obtain a porous surface suitable for the photocatalytic process. Subsequently, they were washed with water to remove all dust particles generated during the sanding process. Finally, they were placed in a drying oven at 100°C for 24 hours.

2.3. Synthesis of ZnO/Au, and TiO₂/Au

To carry out the impregnation synthesis of both TiO₂ and ZnO catalysts, it was necessary to prepare a solution of 5 g of titanium dioxide in 100 mL of double-distilled water, and another solution of 5 g of zinc oxide in 100 mL of double-distilled water. Each solution was applied uniformly to different plates until the surface was completely covered and then placed in a muffle furnace at 500°C for 4 hours.

Once the semiconductors were mounted on the plates, the doping process was carried out. First, the zinc oxide plates were impregnated with gold nanoparticles by preparing a solution of 0.9 g of gold trihydrate chloride in 3 L of double-distilled water. The plates were placed on the falling-film reactor to recirculate the gold chloride solution while being irradiated with ultraviolet light for 4 hours. Afterward, the plates were removed from the reactor and placed in a muffle furnace at 500 °C for 1 hour. All the above steps were also applied in the same way to the plates impregnated with TiO₂.

2.4. Catalyst Characterization

In this work, the characterization of photocatalysts was carried out by UV-vis Spectroscopy using a Shimadzu UV240PC model UV-vis diffuse reflectance spectrophotometer with a barium sulfate cell at a wavelength of 200 nm to 800 nm.

This characterization provides the band gap (Eg) of the doped semiconductor, and with this data, the wavelength (λg) at which the catalysts used are activated can be determined.

To obtain this result, it is necessary to take a UV-Vis spectrum with a diffuse reflectance cell within the wavelength range of 200 to 800 nm.

Scanning Electron Microscopy (SEM) analysis was also performed to determine the morphology and particle size of the catalysts. For this purpose, a FEI QUANTA 200 scanning electron microscope was used with a voltage of 5 kV and magnifications of 35,000x to 350,000x.

2.5. Photocatalytic Activity Tests

The reaction system proposed in this research was designed with the mineralization of organic contaminants in mind in large volumes, unlike conventional methods (such as a Batch reactor), for which solutions of 5, 10, 20, 30, 40, 50 ppm of pyridine were prepared in 6 L of water.

The process began with evaluating the degradation flow rate at three ranges: 0.2 L/min, 0.5 L/min, and 1 L/min. Once the appropriate flow rate was obtained, a new series of reactions was performed to assess photocatalytic degradation at different pH levels: a) natural pH (without any modification), b) acidic pH (pH 5.5, constant throughout the reaction), and c) basic pH (pH 8.5, constant throughout the reaction). Samples were taken every 2 h for 24 h.

To monitor pyridine degradation, the Cary 60 UV-Vis spectrophotometer (Agilent Technologies) was used on each sample.

Total Organic Carbon (TOC) analysis was also performed using the Shimadzu TOC-5000A instrument.

A Thermo Fisher Scientific Surveyor high-performance liquid chromatograph was used to identify pyridine degradation and its respective concentrations. The chromatographic column responsible for retaining the sample components through chemical and physical interactions with the stationary phase was an Agilent Zorbax Eclipse XDB-C18, measuring 4.6 x 150 mm with a particle size of 3.5 µm. The eluent was a mobile phase of 60% 1-hexesulfanate and 40% methanol. The selected wavelength was 257 nm, corresponding to the maximum absorbance of pyridine.

Finally, the kinetic parameters were determined using the Langmuir-Hinshelwood and Hougen-Watson (LH-HW) model.

3. Results

3.1. Catalyst Characterization

3.1.1. Spectroscopy Electron Microscopy (SEM)

Zinc oxide exhibits different morphologies depending on the temperature to which it is subjected. In the literature, authors such as [37,38], indicate that in temperature ranges of 350 to 400 °C there is only an agglomeration of zinc oxide particles, at temperatures of 400 to 450 °C very small nanobarn are formed, and in the range of 450 to 550 °C an increase in the size of the nanobarn is observed, which tend to present a large homogeneous thickness.

Figure 2 shows the formation of prismatic structures or pyramidal terminations exhibiting the hexagonal wurtzite crystalline phase.

Under the conditions used for the synthesis of the doping of the gold nanoparticles, at a calcination temperature of 550°C, the result obtained was the adhesion of the gold nanoparticles to the ZnO particles; however, as shown in Figure 2c, the adhesion of the particles was not uniform, although some gold particles are dispersed on the ZnO particles, most are located on a specific area of the ZnO particles.

The uniform distribution of the dopant on the surface of the photocatalyst is a problem that occurs very often. Özgür et al. [39] mention that this difficulty can arise from a wide variety of causes, such as dopants being compensated for by native low-energy defects of the semiconductor (such as Zni or Vo or H-type background impurities), the low solubility of the dopant in the host material, or deep-level impurities can also be a source of doping problems, causing significant resistance to the formation of the surface acceptor level.

Figure 3 shows micrographs of the TiO₂/Au photocatalyst with an improved distribution of Au nanoparticles on the TiO₂ nanoparticles. Unlike ZnO/Au, the agglomeration of Au nanoparticles was not concentrated in a single site on the semiconductor. This suggests that the synthesis and specifications employed were adequate.

Hajos et al. [40] mention that the distribution and particle size of the doping is of utmost importance, as they are factors for the increase of photocatalytic activity since when there are clusters of dopant particles on the surface of the semiconductor they prevent the passage of light photons, contrary to the fact that the dopant particles are distributed on the semiconductor.

3.1.2. Diffuse Reflectance UV-Visible Spectroscopy

The energy of the band gap was calculated by considering the value of the intersection of the line tangent to the curve with the wavelength axis (λ).

The value obtained is applied within Planck's equation (Eq. 1), which relates photon energy and frequency.

$${\mathrm{E}}_{\mathrm{g}}=\mathrm{h}{\mathrm{v}}_{\mathrm{g}}=\left(6.63\mathrm{x}{10}^{-34}\mathrm{J}.\mathrm{s}\right)\left(3.0\mathrm{x}{10}^8{\displaystyle \frac{\mathrm{m}}{\mathrm{s}}}\right)\left({\displaystyle \frac{1\mathrm{e}\mathrm{V}}{1.60\mathrm{x}{10}^{-19}\mathrm{J}}}\right)$$

(1)

where:

${\mathrm{E}}_{\mathrm{g}}=$ Band- gap ($\mathrm{e}\mathrm{V})$

$\mathrm{h}=$ Planck constant$\left(6.63\mathrm{x}{10}^{-34}\mathrm{J}.\mathrm{s}\right)$

${\mathrm{v}}_{\mathrm{g}}$ = Frequency

Equation 2 shows Equation 1 in a simplified form:

$${\mathrm{E}}_{\mathrm{g}}={\displaystyle \frac{\mathrm{h}\mathrm{c}}{{\mathsf{\lambda }}_{\mathrm{g}}}}={\displaystyle \frac{1240\mathrm{e}\mathrm{V}.\mathrm{n}\mathrm{m}}{392\mathrm{n}\mathrm{m}}}=3.1632\mathrm{e}\mathrm{V}$$

(2)

where:

$\mathrm{c}=$ Light constant in a vacuum$\left(3.0\mathrm{x}{10}^8{\displaystyle \frac{\mathrm{m}}{\mathrm{s}}}\right)$

${\mathsf{\lambda }}_{\mathrm{g}}=$ Activation wavelength measured in nanometers (nm)

Figure 4 shows the UV-Vis diffuse reflectance spectrum of zinc oxide. The inflection point, located at 392 nm, is marked with an arrow. At this wavelength, an Eg of 3.163 eV was obtained. The results of this analysis demonstrate that the impregnation synthesis was suitable for the adhesion of gold particles to the surface of the ZnO particles, thereby modifying the ZnO crystalline structures and causing changes in energy levels. Montalvo (2009) characterized ZnO using UV-Vis diffuse reflectance spectroscopy to determine the bandwidth energy and found that this photocatalyst had an Eg of 3.17 eV at a wavelength of 390 nm, confirming the influence of the dopant on the ZnO used.

3.2. Degradation Photocatalytic

3.2.1. Analysis of UV-Vis Spectra for Pyridine Degradation with ZnO/Au and TiO₂/Au Photocatalysts

The spectra of pyridine degradation were obtained for each degradation run to analyze the compound's behavior under different operating conditions. The maximum characteristic wavelength of pyridine was found to be 257 nm.

Figure 5a shows the photocatalytic activity of ZnO/Au for pyridine degradation, with samples taken every 3 hours for 24 continuous hours.

Figure 5b shows the degradation of pyridine in solution using the TiO₂/Au photocatalyst under UV irradiation, monitored by taking samples every 3 hours over 24 continuous hours of degradation.

Degradation spectrum of the pyridine compound with Co = 10 ppm during 22 hours of degradation, under the conditions of 1L/min (HRT), pH = natural (7.8) and irradiated by UV light, using TiO₂/Au as a photocatalyst.

The decrease in concentrations during the first 9 hours shows variations, suggesting the compound may be exhibiting resistance to degradation, as the decrease is not uniform. However, in the first 3 hours, 60% of the contaminant is removed.

Figure 6A shows the concentration vs. time graph for the degradation of the pyridine compound at 50 ppm with the ZnO/Au catalyst under the established optimal operating conditions. Up to hour 6, a nearly constant decrease in concentration is observed. After this time, the decrease slows, and by hour 16, another significant change in the degradation rate is seen, making the degradation process much slower.

The results obtained from the degradation of pyridine at 50 ppm for the TiO₂/Au catalyst are shown in Figure 6B.

In Figure 6B, degradation accelerates within the first 9 hours; However, after 12 hours, degradation remains constant, showing some resistance to complete removal of the contaminant. These results generally indicate that when ZnO/Au is used, degradation within the first 9 hours proceeds more slowly than when TiO₂/Au is used. After 12 hours, the effect is reversed because TiO₂/Au does not achieve concentrations as low as those obtained with ZnO/Au.

3.2.2. Effect of Hydraulic Residence Time on the Photocatalyst Surface

The solution's descent on the flat surface of the clay plates was due to the reaction system being designed with a 60° angle of inclination. Starting from this angle, flow rates of 0.2, 0.5, and 1 L/min were evaluated for the degradation of the pyridine compound with each of the selected photocatalysts. This analysis was performed for samples with the same initial concentration.

Similarly, having very large water parcels with very thick boundary layers on the surface where the catalyst is supported can impede or reduce the arrival of photons to the photocatalyst surface.

Figure 7 shows a very similar behavior among the three TRHs used. At hour 2, all three TRHs show an increase in concentration, which then tends to decrease, exhibiting a fluctuating pattern throughout the 18-hour reaction. It is important to note that the most stable behavior occurred during the reaction when the flow rate was 1 L/min.

Similarly, the hydraulic retention time (HRT) for the TiO₂/Au photocatalyst was compared. The graphs of behavior in relation to the study variable are shown in Figure 8B, where the concentration decreases steadily for the three HRTs. Degradation accelerates at the 0.5 L/min flow rate during the first 3 h and shows the same trend from 24 hours onwards, with the lowest concentration occurring at the 1 L/min flow rate. It should be noted that the difference between concentrations is minimal among the three HRTs.

The behavior observed in the graphs suggests using a very short hydraulic retention time, since with a recirculation flow rate of 0.2 L/min (higher HRT), there is a high concentration variation as well as an increase in degradation times. The graphs also show that recirculation flow rates of 0.5 and 1 L/min (lower HRT) favor photocatalytic activity, exhibiting constant degradation of the compound, which indicates that a higher flow rate promotes mass transfer.

The behavior observed at recirculation flows of 0.2 L/min is a phenomenon that occurs due to laminar flow over a flat surface. At this speed, a boundary layer forms, and as this layer moves along the body, the continuous action of shear stresses tends to decelerate additional fluid particles, causing the layer thickness to increase [41]. With the increased boundary layer thickness, mass transfer decreases, impacting degradation efficiency due to the need to consider diffusive effects.

Increasing the recirculation flow optimizes the process by generating greater mass transfer within the reaction system. Within heterogeneous photocatalysis, there are two types of reactions classified according to their mechanism: the first is direct photoreduction by photogenerated electrons and photooxidation by generated holes (heterogeneous reactions); the second is radical-mediated reactions, which involve reactive species such as the •OH radical and the •O radical (homogeneous reactions).

In the absence of mass transfer, only heterogeneous reactions, such as electron-hole pairs, will occur. This is not optimal due to the complexity of heterogeneous reactions caused by all the factors involved. However, increasing mass transfer promotes the generation and release of radicals, suggesting simpler and faster homogeneous-phase reactions. For this degradation system, shorter hydrostatic reactions (HRTs) generate more turbulent flows and, therefore, a smaller boundary layer with turbulent characteristics. This is beneficial because it leads to greater generation and release free radicals into the aqueous medium.

3.2.3. Effects of pH on the Solution

The effects of pH in the aqueous medium significantly influence the photocatalytic degradation rate because it generates significant changes on the degradation surface. The pH of the medium determines the speciation of the contaminant, thus determining its interaction with the surface of TiO₂ or ZnO [43]. Therefore, the degradation rate was evaluated at the following pH levels: acidic (5.5), basic (8.5), and natural (7.8).

Figure 8A shows the behavior of the ZnO/Au photocatalyst with respect to the degradation efficiency of the pyridine compound when the solution pH is modified. The concentration at the three different pH levels in Figure 8A suggests that the natural pH allows for the lowest concentration values. After 4 hours, the pH of 8.5 results in lower degradation values, while after 14 hours, the natural pH and the pH of 8.5 again converge.

The degradation times of the pyridine solution vary with pH. Similarly, semiconductor types, even with very similar Eg values, will exhibit different behavior due to the semiconductor's orientation towards more negative or positive potentials.

Figure 8B shows the degradation times with respect to pH for the ZnO/Ag catalyst.

The graphs for all catalysts show that the photocatalytic degradation efficiency of pyridine is favored at pH values of 7.8 (natural). This is related to the charge conditions prevailing on the surface of the semiconductor material. Generally, the pH of the solution plays an important role in the photocatalytic degradation of compounds. The point of zero charge (PZC) technique allows the measurement of the pH value at which the surface of a metal oxide has a neutral charge. In the case of titanium dioxide, the PZC value is in the range between 5.6 and 6.8 [42], and for zinc oxide, the PZC value is 9 [43].

The interaction of the ceramic clay plates supporting the photocatalyst, as well as the metals used as dopants, may have caused a shift in the zero charge point towards more neutral pH values. Therefore, when pH 5.5 was used, lower than pH 7.8, the catalytic surface may have become positively charged, and when pH 8.5 was used, the catalytic surface may have become negatively charged. These two states will affect the adsorption and desorption of the reactants and reaction intermediates on the catalyst surface, and thus the photocatalytic efficiency.

3.3. Kinetic Analysis of the Photocatalytic Degradation Of Pyridine

Due to the characteristics of pyridine degradation, the Langmuir-Hinshelwood and Hougen-Watson (LH-HW) model was used. The first step of the photocatalytic degradation of organic compounds follows pseudo-first-order kinetics [43]. Analysis of the degradation results shows that the reaction rate is strongly dependent on the reactant concentration; therefore, it can be deduced that our reactions follow first-order kinetics.

$$-r_a={\displaystyle \frac{dCa}{dt}}=K{C}_a$$

(3)

Due to the above, in which the experimental concentration profile data are normalized, it is possible to plot C/Co as a function of reaction time, which is presented in Figure 10.

Figure 9a shows the C/Co vs. time graphs for the degradation of pyridine using ZnO-Au as a catalyst. The reaction with the initial concentration of 10 ppm is closest to the Y-axis, exhibiting first-order behavior. The other concentrations are in the middle of the graph, close to the 10-ppm run, but tend toward pseudo-first-order reactions. Comparing this group of reactions with the subsequent data reveals that no reactions tending toward zero-order behavior appear here.

Figure 9b shows the ln C/Co versus time plots for the degradation of the pyridine compound carried out with TiO₂ and doped with Au. The reactions with low initial concentrations, such as 5 and 10 ppm, exhibit zero-order behavior, as they are positioned horizontally at the top of the graph. The reactions with initial concentrations of 20, 30, and 50 ppm are positioned in the middle of the graph, suggesting pseudo-first-order reactions. Finally, the reaction at 40 ppm is closer to the Y-axis; however, it could also be considered pseudo-first order. Based on this behavior, it could be assumed that the reactions best fit the model used.

Integrating the velocity equation yields the following function:

$$\ln {\displaystyle \frac{\mathrm{C}}{{\mathrm{C}}_0}}=-{\mathrm{K}}_{\mathrm{a}\mathrm{p}\mathrm{p}\mathrm{a}\mathrm{r}}\mathrm{t}$$

(4)

Which is equivalent to Equation 5:

$$\mathrm{C}={\mathrm{C}}_0{\mathrm{e}}^{-{\mathrm{K}}_{\mathrm{a}\mathrm{p}\mathrm{p}\mathrm{a}\mathrm{r}}\mathrm{t}}$$

(5)

Using the logarithm of the normalized concentration (ln(C/C₀)) versus time plots shown in Figure 9a,b, the apparent reaction constant can be obtained through linear regression.

Table 1 shows the values obtained from the linear regressions of the apparent reaction constants.

Table 1 provides a more detailed view of the reaction behavior through the apparent reaction constants for each initial concentration. The results for ZnO/Au under UV light, with respect to the apparent reaction constants, do not show uniform behavior between higher and lower initial concentrations. The results for TiO₂/Au for the apparent reaction constants suggest that these constants increase with higher initial concentrations, and the opposite is true for lower initial concentrations.

3.4. Determination of Reaction Rate Constant and Adsorption by HPLC

The data necessary to carry out the elaboration of the linear regression graphs was obtained by means of HPLC for the different types of catalysts. The results show that, when increasing the initial concentration, degradation constantly decreases, thus fitting into photocatalytic oxidation reactions that follow a Langmuir-Hinshelwood and Hougen-Watson (LH-HW) type kinetics.

$$-{\mathrm{r}}_{\mathrm{a}}=-{\displaystyle \frac{\mathrm{d}\mathrm{C}}{\mathrm{d}\mathrm{t}}}={\displaystyle \frac{{\mathrm{K}}_1\mathrm{C}}{\mathrm{K}2\mathrm{C}+\mathsf{\Sigma }{\mathrm{K}}_{\mathrm{i}}\mathrm{C}\mathrm{i}}}$$

(6)

where:

K₁ C: Represents the kinetic term of the rate equation.

K₂ C: Represents the adsorption term of the reactant.

ΣK_i C_i: Represents the adsorption term of all intermediate products of the organic compound degradation reaction.

If the experimental data are analyzed at very short reaction times, it is possible to neglect the adsorption term of the intermediate products. Then, equation (6) becomes:

$$-{\mathrm{r}}_{\mathrm{a}}=-{\displaystyle \frac{\mathrm{d}\mathrm{C}}{\mathrm{d}\mathrm{t}}}={\displaystyle \frac{{\mathrm{K}}_1\mathrm{C}}{1+{\mathrm{K}}_2\mathrm{C}}}$$

(7)

If equation (7) is analyzed for very low concentrations of pyridine where K₂ C≪1, Equations 8 and 9 will be obtained:

$$-{\mathrm{r}}_{\mathrm{a}}=-{\displaystyle \frac{\mathrm{d}\mathrm{C}}{\mathrm{d}\mathrm{t}}}={\displaystyle \frac{{\mathrm{K}}_1\mathrm{C}}{1+{\mathrm{K}}_2\mathrm{C}}}={\displaystyle \frac{{\mathrm{K}}_1\mathrm{C}}{1}}={\mathrm{K}}_1\mathrm{C}$$

(8)

$$-{\mathrm{r}}_{\mathrm{a}}={\displaystyle \frac{\mathrm{d}\mathrm{C}}{\mathrm{d}\mathrm{t}}}={\mathrm{K}}_1\mathrm{C}$$

(9)

As can be seen, equation (9) is a first-order rate equation with kinetic coefficient K₁. For high concentrations (K₂ C ≫ 1), the result is:

$$-{\mathrm{r}}_{\mathrm{a}}=-{\displaystyle \frac{\mathrm{d}\mathrm{C}}{\mathrm{d}\mathrm{t}}}={\displaystyle \frac{{\mathrm{K}}_1\mathrm{C}}{1+{\mathrm{K}}_2\mathrm{C}}}={\displaystyle \frac{{\mathrm{K}}_1\mathrm{C}}{1+{\mathrm{K}}_2\mathrm{C}}}={\displaystyle \frac{{\mathrm{K}}_1}{{\mathrm{K}}_2}})$$

(10)

$$-{\mathrm{r}}_{\mathrm{a}}={\displaystyle \frac{\mathrm{d}\mathrm{C}}{\mathrm{d}\mathrm{t}}}={\displaystyle \frac{{\mathrm{K}}_1}{{\mathrm{K}}_2}}$$

(11)

where equation (11) is a zero-order rate equation, with a kinetic coefficient ${\displaystyle \frac{K_1}{K_2}}$.

Based on the foregoing, it can be demonstrated that the general kinetic form can be represented by equation 12:

$$-{\mathrm{r}}_{\mathrm{a}}=-{\displaystyle \frac{\mathrm{d}\mathrm{C}}{\mathrm{d}\mathrm{t}}}={\displaystyle \frac{{\mathrm{K}}_1{\mathrm{C}}^{\mathrm{m}}}{1+{\mathrm{K}}_2{\mathrm{C}}^{\mathrm{n}}}}$$

(12)

If the exponents m and n have a value of 1, the constants K₁ and K₂ can be determined directly from the plot of reaction rate versus concentration. Equation 12 can be linearized in the form recommended by Fogler [44], Moctezuma et al. [45], and Montalvo [43]., using the following initial conditions: t = 0, C = Co, and reaction rate (${\left.-r_a\right|}_{t=0}$).

Equation 13, which corresponds to the initial reaction rate, is transformed by taking its reciprocal, as shown in equation 14.

$$-{\mathrm{r}}_{\mathrm{a}}{{{|}_{\mathrm{t}}}_{=}}_0={\displaystyle \frac{{\mathrm{K}}_1{\mathrm{C}}_0}{1+{\mathrm{K}}_2{\mathrm{C}}_0}}$$

(13)

$${\displaystyle \frac{1}{-r_a{{{|}_t}_{=}}_0}}={\displaystyle \frac{1+K_2{C}_0}{K_1{C}_0}}={\displaystyle \frac{1}{K_1{C}_0}}+{\displaystyle \frac{K_2{C}_0}{K_1{C}_0}}$$

(14)

$${\displaystyle \frac{1}{-r_a{{{|}_t}_{=}}_0}}={\displaystyle \frac{1}{K_1{C}_0}}+{\displaystyle \frac{K_2}{K_1}}$$

(15)

The initial reaction rate is determined from the experimental concentration-versus-time data for each run (using different initial concentrations). This is done by multiplying the reaction constant (K_appar) by the initial concentration. Subsequently, is plotted against (1/-ro) vs (1/Co).

Table 3 and Table 4 show these tabulated values for the ZnO/Au and TiO₂/Au catalysts, respectively, which are subsequently plotted.

Thanks to the R² value, we can see how the points are distributed relative to the line, noting that the ideal R² value is 1, where all points would lie exactly on the line. In the case of Figure 12A—specifically for the ZnO/Au catalyst—the values are observed to be very close to the line, yielding an R² value of 0.9712.

, based on the linear regression of the Langmuir–Hinshelwood model for the photocatalytic degradation of pyridine using ZnO/Au–UV light (a) and TiO₂/Au–UV light (b).

In Figure 10b, we can see this behavior for the TiO₂/Au case. We can observe that the data are somewhat more dispersed, which is reflected in the R2 value, lower than that of zinc oxide with gold, which is 0.9499.

It is thanks to the linear regression plots that it is possible to obtain the K₁ and K₂ values for the degradation reaction coefficients of pyridine with ZnO/Au and TiO₂/Au. The K₁ and K₂ values are shown in Table 5.

Table 5 shows the reaction and adsorption constants (K₁ and K₂) for the TiO₂/Au photocatalyst irradiated with UV light. K₁ was higher than when the reactions were carried out with the other TiO₂/Au catalysts. Similarly, the adsorption constant K₂ was higher in the TiO₂/Au-UV light reactions.

If the initial reaction rate is plotted against the initial concentration, the graphs conform to the Langmuir-Hinshelwood and Hougen-Watson (LH-HW) model.

Figure 11a shows the reaction rate originating from the photocatalytic degradation of the pyridine compound for the ZnO/Au catalyst using UV light. This figure does not fully show the development of the typical LH-HW model curve; however, it is noticeable that the reactions with smaller contractions, such as 5 and 10 ppm, are very close to the first-order area, while the rest of the concentrations are in the central part of the graph, suggesting pseudo-first-order reactions.

Figure 11B shows the reaction rate originating from the photocatalytic degradation of pyridine for the TiO₂/Au catalyst.

Figure 11B indicates that the 5-ppm concentration is the only one close to first-order behavior, while the 10-ppm concentration suggests pseudo-first-order behavior. Finally, the 20, 30, 40, and 50 ppm concentrations are associated with zero-order reaction behavior

3.5. Mass Balance in the Photocatalytic Degradation Process of Pyridine

Performing a mass balance on the results obtained from the photocatalytic degradation of pyridine aims to quantify the amount of pyridine being transformed into CO₂ and intermediate organic products (IOPs). The results of this mass balance determine whether the photodegradation process, along with the photocatalysts used, is effective in the complete mineralization of the compound. In this study, a mass balance could only be performed on the ZnO/Au catalyst and not on the TiO₂/Au catalyst, as the results obtained were neither adequate nor consistent.

Palmisano [46] considered that, if the intermediates are not fully identified in number and identity, the adsorption term of equation 13, for all organic molecules, can be approximated by the following equation:

$$K_{2}C+\sum _i{K}_i{C}_i=K_2{C}_0$$

(16)

where C_O is the initial concentration of the original reactant. The adsorption constant of the intermediates present in the mineralization is assumed to be the same for each of them and equal to the adsorption constant of the reactant. The mass of the reactant and the organic products completely transformed to CO₂ is also disregarded. Substituting equation 16 into equation 13, we have:

$$-ra={\displaystyle \frac{K_1\ast C}{1+K_2{C}_0+\sum _i{K}_i{C}_i}}$$

(17)

$$-ra={\displaystyle \frac{K_{1}C}{1+K_2{C}_0}}$$

(18)

Zamarripa [47] demonstrated that Palmisano's assumption fails to predict the system's behavior when a large portion of the organic carbon mineralizes to CO₂. This assumption can be corrected by performing a total organic carbon balance.

The adsorption term for all intermediate organic reaction products can be explained by the equation:

$$\sum _i{K}_1{C}_i=K_3{C}_{ti}$$

(19)

where C_ti is the sum of the concentrations of all intermediate organic products. In this case, it is assumed that all intermediate organic products (IOPs) have the same adsorption constant, but a different value than the adsorption constant of the original reactant.

The concentration C_ti is determined by a mass balance as follows:

$${\left[TOC\right]}_{t=0}={\left[Total organic carbon in pyridine\right]}_{t=0}$$

(20)

The total organic carbon (TOC) in the reaction mixture at any instant is given by:

$${\left[TOC\right]}_{t=t}={\left[\begin{array}{c}Totalorganic\\ carboninpyridine\end{array}\right]}_{t=t}+{\left[\sum \begin{array}{c}Totalorganiccarbonfromorganic\\ productsintermediaries\end{array}\right]}_{t=t}$$

(21)

The first term on the right-hand side of eq. 21 is determined by analyzing the reaction samples by HPLC.

Therefore, the sum of the organic carbon of all compounds is calculated by a simple subtraction:

$$C_{\mathrm{I}\mathrm{O}\mathrm{P}\mathrm{s}}={C_{TOC}}_{t=t}-{C_{HPLC}}_{t=t}$$

(22)

The amount of CO₂ formed at any given time will be given by:

$$C_{C{O}_2}={TOC}_{\left(t=0\right)}-{TOC}_{\left(t=t\right)}$$

(23)

To illustrate the balance, the photocatalyst ZnO/Au were used, from which the results of experiments on the photocatalytic degradation of pyridine at high concentrations (50 ppm) and low concentrations (5 ppm) were taken. The HPLC and TOC data used are those used, with millimoles (mM) of pyridine as the basis for calculation.

To perform the mass balance in millimoles, we start with the molecular weight of pyridine, which is 79.0999 g/mol, of which 60.0535 g/mol corresponds to the amount of organic carbon present in the molecule, 14.0067 g/mol to nitrogen, and 5.0397 g/mol to hydrogen. For a concentration of 50 ppm of pyridine, there are 0.6321 millimoles of pyridine.

Column 4 of Table 6 shows the TOC data in ppm; this data is used in relative concentrations (column 5). The column labeled present TOC (column 6) results from multiplying the TOC content in millimoles at t = 0 by the relative concentration from the previous column.

The present TOC is determined by considering that one molecule of pyridine at a concentration of 200 ppm contains 151.8987 ppm of TOC. To convert to millimoles, we divide this by 12 (the molecular weight of carbon), resulting in 12.6582 millimoles.

The stoichiometric equation for the photocatalytic pyridine degradation reaction can be written as follows:

$${\mathrm{C}}_5{\mathrm{H}}_5\mathrm{N}\stackrel{\mathrm{u}\mathrm{V},\mathrm{T}\mathrm{i}{\mathrm{O}}_2}{\to }\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{m}\mathrm{e}\mathrm{d}\mathrm{i}\mathrm{a}\mathrm{t}\mathrm{e} \mathrm{o}\mathrm{r}\mathrm{g}\mathrm{a}\mathrm{n}\mathrm{i}\mathrm{c} \mathrm{p}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{s} \left(\mathrm{I}\mathrm{O}\mathrm{P}\right)\to 5\mathrm{C}{\mathrm{O}}_2+{\displaystyle \frac{5}{2}}{\mathrm{H}}_2\mathrm{O}+\mathrm{N}{\mathrm{O}}_3$$

(24)

From this it follows that for every mole of pyridine (equation 24) the photodegradation reaction of pyridine produces 5 moles of CO₂ for every mole of pyridine, so dividing 12.6582 moles by 5 moles produced gives us 2.5316 millimoles of pyridine as TOC (present TOC).

The data for CO₂ produced (column 7) is determined by the difference between the initial TOC concentration and the TOC determined at a given time, equation 25.

$${\mathrm{C}}_{\mathrm{C}{\mathrm{O}}_2}={\mathrm{T}\mathrm{O}\mathrm{C}}_{\mathrm{t}=0}-{\mathrm{T}\mathrm{O}\mathrm{C}}_{\mathrm{t}=\mathrm{t}}$$

(25)

The concentration of intermediate organic products (IOP) is determined by the difference between the total concentration of TOC and the carbon concentration by HPLC (column 8).

$${\mathrm{C}}_{\mathrm{I}\mathrm{O}\mathrm{P}}={\mathrm{C}}_{\mathbf{T}\mathbf{O}\mathbf{C}}-{\mathrm{C}}_{\mathrm{H}\mathrm{P}\mathrm{L}\mathrm{C}}$$

(26)

These data are also shown for the ZnO/Au catalyst and a concentration of 50 ppm in Table 7.

Figure 12 graphically shows the mass balance for the degradation tests of high and low concentration pyridine solutions (5 ppm = 0.0632 mM and 50 ppm = 0.632 mM), comparing the ZnO/Au photocatalyst using UV light.

Determination of the average adsorption constant of the intermediates (K₃) and the reaction constant of the intermediates (K₁₁).

The photocatalytic degradation process of pyridine can be modeled by a series reaction mechanism described by equation (27).

$$P\to POI\to 5C{O}_2+{\displaystyle \frac{5}{2}}{H}_{2}O+N{O}_3$$

(27)

This is done because the intermediate organic products have not been fully identified and quantified.

As demonstrated above, the reaction rate for pyridine degradation is given by equation 27.

$$r=-{\displaystyle \frac{dC}{dt}}={\displaystyle \frac{K_1\ast C}{1+K_2\ast C+K_3\ast C_T{,}_i}}$$

(6.27)

The rate of formation and consumption of intermediate organic products can be written as follows:

$${\mathrm{r}}_{\mathrm{T}\mathrm{i}}={\mathrm{r}}_{\mathrm{a}}-{{\mathrm{r}}_{\mathrm{T}}}_{,\mathrm{i}}$$

(6.28)

Therefore, the kinetic model for the photodegradation of intermediates is:

$${{\mathrm{r}}_{\mathrm{T}}}_{\mathrm{i}}={\displaystyle \frac{\mathrm{d}{\mathrm{C}}_{\mathrm{T}\mathrm{i}}}{\mathrm{d}\mathrm{t}}}={\displaystyle \frac{{\mathrm{K}}_1\ast \mathrm{C}}{1+{\mathrm{K}}_2\ast \mathrm{C}+{\mathrm{K}}_3\ast {\mathrm{C}}_{\mathrm{T}\mathrm{i}}}}-{\displaystyle \frac{{\mathrm{K}}_{11}\ast {\mathrm{C}}_{\mathrm{T}\mathrm{i}}}{1+{\mathrm{K}}_2\ast \mathrm{C}+{\mathrm{K}}_3\ast {\mathrm{C}}_{\mathrm{T}\mathrm{i}}}}$$

(6.29)

To determine the K₁₁ and K₃ constants of equation 29, the optimization program was used using Statistica TM version 7 software, performing nonlinear least squares estimation analysis using the Levenberg-Marquardt estimation method.

For the above, the experimental results of the runs with solutions of 5, 10, 20, 30, 40, and 50 ppm and the material balance results presented in Table 6 and Table 7 were used.

The results for the determined values of K₁₁ and K₃ are reported in Table 8.

These values indicate that the adsorption constant for the intermediates (K₃) is nearly half the value of the reaction constant for the intermediates (K₁₁); consequently, this explains the observed increase in CO₂ production during the final hours of the reaction.

5. Conclusions

The ZnO/Au and TiO₂/Au photocatalysts were characterized using scanning electron microscopy (SEM). For the ZnO/Au photocatalyst, the dispersion of the dopant species (Au) was not uniform across the surface of the ZnO semiconductor. In contrast, for the TiO₂/Au photocatalyst, the dopant distribution was more homogeneous across the surface of the TiO₂ particles.

The ZnO/Au photocatalyst was also characterized using diffuse reflectance UV-Vis spectroscopy (DRS), yielding a band gap of 3.1632 eV, corresponding to a wavelength of 392 nm. Comparing this result with the DRS analysis of pure ZnO, which has a wavelength of 390 nm, demonstrates that doping with gold nanoparticles modified the energy levels of ZnO.

Finally, using UV-Vis spectroscopy techniques, the ability of both photocatalysts to promote the degradation of pyridine in aqueous solution at different concentrations was verified, confirming their effectiveness in photocatalytic processes.

The results of the photocatalytic degradation tests were corroborated using high-performance liquid chromatography (HPLC) at different pH values, ranging from acidic pH = 5.5, natural pH = 7.2, and basic pH = 8.5. Similarly, the reaction performance was compared with different hydrostatic reactions (HRRs) of 0.2 L/min, 0.5 L/min, and 1 L/min.

In the case of the ZnO/Au photocatalyst using UV light, it was demonstrated that the best degradation times were obtained when the degradation parameters were pH = natural (7.2) and HRR = 1 L/min. These same parameters were also observed when using the TiO2/Au catalyst. Therefore, it has been concluded that both pH and hydraulic residence time significantly influence photocatalytic degradation efficiencies.

The LH-HW model was used to determine the reaction behavior for the different photocatalysts.

For the ZnO/Au – UV light photocatalyst, it was observed that at lower concentrations, corresponding to 5 and 10 ppm, the reactions suggest first-order behavior, while at concentrations of 20, 30, 40, and 50 ppm, the model suggests pseudo-first-order behavior.

For the TiO2/Au – UV light photocatalyst, it was observed that at lower concentrations, corresponding to 5 and 10 ppm, the reactions suggest first-order behavior, while at concentrations of 20, 30, 40, and 50 ppm, they suggest zero-order behavior. The results of the mass balance performed for the selected photocatalysts show the generation of intermediate organic compounds, along with CO₂ production.

For the ZnO/Au – UV light photocatalyst, the mass balance showed a direct relationship between CO₂ production and compound degradation. The generation of intermediate organic compounds increases in the first few hours and then decreases as CO₂ levels rise, indicating complete pyridine mineralization.

It was observed that ceramic clay sheets are a good option as a support for the photocatalysts used.

The porosity of the plate surface prevents the catalyst from being washed away by the drag forces of the surrounding liquid medium.