Porous Ceramic ZnO Nanopowders: Features of Photoluminescence, Adsorption and Photocatalytic Properties

Abstract

The grainy and porous ZnO powders were synthesized by thermal decomposition of zinc nitrate and polymer-salt method. The comparative study of the crystal structure, morphology, luminescence, adsorptive and photocatalytic properties of ZnO powders was carried out. The addition of PVP in initial aqueous solutions of zinc nitrate determines the remarkable change of powder morphology and decreases the average size of ZnO nanocrystals. Luminescence spectra in the visible spectral range indicate the significant difference of structural defects types in grainy and porous powders. Porous powders demonstrate high ability for singlet oxygen photogeneration and photocatalytic properties. The kinetics of diazo dye adsorption on both powders is described successfully by the kinetic equation of pseudo-second order. Kinetic dependencies of photocatalytic oxidation of Chicago Sky Blue diazo dye using as grain ZnO powder so as porous ZnO powders are described by the Langmuir–Hinshelwood model but process rates are different. Porous ZnO powder demonstrates strong ability for photogeneration of singlet oxygen under visible irradiation and high photocatalytic properties (rate constant 0.042 min⁻¹).

1. Introduction

The development of effective ceramic photocatalysts for different modern practical applications (hydrogen energy, environmental applications, medicine, sensors) is a real problem. Many different oxide photocatalysts have already been developed and tested in many practical applications [1,2,3].

Photocatalytic activity of oxide materials strongly depends on their morphology [4,5,6]. Furthermore, 1D- and 2D-nanomaterials (nanowires [7], nanorods [5,8], fibers [9], sheets [10], etc.), “flowers”-like particles [11] and porous matrices [10,12,13], i.e., materials that have high specific surface areas, demonstrate high photocatalytic properties.

Structural defects in ceramic materials can play an important role in their photocatalytic properties. These intrinsic defects can optimize characteristics of photocatalysts, including the shifting of band edges, dispersion of energy bands, creation of charge-trapping centers, creation of surface catalytic sties and surface sensitization [14].

Adsorption is one of the key processes in photocatalysis [15,16,17,18]. The adsorption of organic contaminants from liquids on the photocatalyst’s surface is the initial stage of their subsequent photocatalytic decomposition. The organics adsorption and their ensuing photodecomposition are often considered as successive stages of the photocatalytic process. Clearly, the adsorption kinetics affect the effectiveness and kinetics of photocatalytic decomposition of organic pollutants. This fact determines the importance of the study of adsorption kinetics in the optimization of photocatalysts.

The generation of reactive oxygen species (ROS) by different oxide nanoparticles under external light irradiation is an important stage of the photocatalytic processes and their antibacterial activity [19]. This process includes the following stages: (a) light excitation of oxide crystals; (b) electron-hole pairs formation; (c) ROS generation on the materials surface and (d) the decomposition of surface organic contaminations. The generation of ROS can be interpreted by the electronic structures and surface defects of the metal oxide nanoparticles and the ROS redox potentials [20]. The study of ROS generation by photocatalysts is important for their optimization and development.

ROS formation proceeds on the surface of photocatalysts and the increase in their specific surface area enhances photocatalytic properties significantly. High photocatalytic activity of different porous photocatalysts having high specific surface areas was observed in [16,17,21]. Furthermore, the influence of surface structural defects on the electronic structure is high in porous materials.

The features of adsorption and photocatalysis processes in different porous materials were reviewed in [16,17,21]. Linear dependencies of photocatalytic degradation rates of organics decomposition from the visible light intensity were found in [21]. In [17], two stages of adsorption on porous N-doped TiO₂ photocatalyst were observed: (1) fast stage (duration of this stage about 1 h) and (2) second slow stage.

The study of features of the porosity influence on materials’ photocatalytic and adsorption properties should be carried out on the single component material to minimize the effect of chemical composition on these characteristics. Therefore, in this work we study the influence of the porosity of single component ZnO powders on their structural, photoluminescence, adsorption and photocatalytic properties.

It is known that ZnO is one of the most effective photocatalyst [9,22,23,24]. The relatively wide band gap of this semiconductor (E_g~3.36–3.37 eV [25,26,27]) limits its photocatalytic activity under sunlight irradiation. Solar radiation contains only 4% UV but visible light accounts for about 42% solar light [22,25]. Doping ions (Co²⁺, Eu³⁺, etc. [28,29,30]) or different structural defects in the crystal lattice [12] modify the material structure and form isolated energy levels in the band gap of the semiconductor and affect its photocatalytic activity under visible irradiation.

Porous ZnO materials with enhanced photocatalytic activity were prepared and studied in [10,12,13]. At the application of the porous materials, the enhancement of photocatalytic properties is related to their higher specific surface areas and improved adsorption properties compared with non-porous analogues [18]. It was found [18] that the ratio of surface to bulk structural defects in porous ZnO plays an important role in materials’ photocatalytic characteristics. Thus, experimental results described in [10,12,13,18] showed the high photocatalytic effectiveness of porous ZnO. However, the complex study of the porosity effect on the ability of photogeneration of singlet oxygen and luminescent, adsorptive and photocatalytic properties of ZnO ceramic powders is still a real problem.

Different methods (sol–gel [12], microwave hydrothermal [13]) processes were applied for the synthesis of porous ZnO photocatalysts. In [31], porous ZnO-MgO-Ag nanocomposites with enhanced ability of singlet oxygen photogeneration and high bactericidal properties were synthesized by polymer–salt method.

The polymer–salt method is simple, cheap and is applied for the synthesis of different oxides nanoparticles (NPs) [11,15,31,32,33,34]. Solutions containing soluble polymer and thermally decomposable metal salts are used as oxide NPs precursors in this method. After drying, the obtained polymer–salt composite is calcined for the polymer and metal salts decomposition and formation of oxide NPs.

The aim of this work is the comparative study of features of the structure, luminescent spectra, ability for photogeneration of singlet oxygen, adsorptive properties and photocatalytic activity of grainy and nanoporous ceramic ZnO powders. In the present work we consider the main differences in the structure, morphology, luminescent, adsorptive and photocatalytic properties of grainy and porous ZnO particles. We guessed that the obtained data could be useful for the development and application of another ceramic photocatalyst.

2. Materials and Methods

Aqueous solutions of Zn(NO₃)₂ (0.05 g/mL) (analytical grade; Neva Reactive Co., Saint-Petersburg, Russia) were applied as the raw materials for ZnO powder preparation. After drying the solution, the obtained salt material was calcined at 550 °C for 2 h and ZnO powder 1 was obtained.

In polymer–salt synthesis, the predetermined volumes of aqueous solutions of Zn(NO₃)₂ and polyvinylpyrrolidone (0.05 g/mL) (PVP, M_w = 25,000 ÷ 35,000) were mixed in predetermined volumes. Polyvinylpyrrolidone is a non-toxic organic polymer which is used for the stabilization of different nanoparticles in liquids [23,32,35]. Furthermore, PVP can react at high temperatures with oxidizing agents, such as metal nitrates, that increases the local temperature of the reaction mixture and leads to gas product formation [31]. The prepared mixture was dried in an air atmosphere at 70 °C, and then the obtained organic-salt composite was calcined at 550 °C for 2 h. The obtained ZnO powder 2 was light and fine.

The crystal structure of the prepared composites was studied by XRD analysis using a Rigaku SmartLAB 3 (CuKα, 40 kV, 44 mA). Based on obtained XRD data, crystal sizes were estimated using the Scherrer formula:

$$\mathrm{d}=\frac{\mathrm{K}\times \mathsf{\lambda }}{\mathsf{\beta }\times \cos \mathsf{\theta }}$$

(1)

where d is the average crystal size, K is the dimensionless particle shape factor (for spherical particles K = 0.9), λ is the X-ray wavelength (λ(CuKα) = 1.5418 Å), β is the width of the reflection at half height (in radians, and in units of 2θ) and θ is the diffraction angle. Furthermore, the Williamson–Hall method was applied for the calculation of accurate crystallite sizes.

SEM analysis using a Tescan Vega 3 SBH instrument was applied to study the morphology of ZnO powders.

The photoluminescence (PL) spectra measurements were carried out using a Perkin Elmer LS-50B fluorescence spectrophotometer in the spectral range of 380–550 nm.

The photoluminescence (PL) method previously used in [11,33] was applied for the estimation of the material’s ability to generate chemically-active singlet oxygen under light irradiation. The measurements of PL spectra in the near IR spectral range were performed using an SDH-IV spectrometer (SOLAR Laser Systems, Minsk, Republic of Belarus) with luminescence excitation by LED HPR40E (λ_max = 405 nm; power density of light 0.35 W/cm²).

Chicago Sky Blue (CSB) diazo dye (Sigma Aldrich Rus Ltd., Moscow, Russia) was used as the model organic contaminant for the estimation of adsorption and photocatalytic properties of prepared ZnO powders. This dye was used previously for the study of photocatalytic and adsorption properties of different materials [32,33]. Adsorption and photocatalytic properties of powders were studied in their suspensions, which were prepared by the addition of 0.01 g of ZnO to 3 mL of the aqueous solution of Chicago Sky Blue (CSB) dye (0.01 g/L). The quartz cuvettes were filled with these suspensions and kept in darkness during dye adsorption experiments. During photocatalytic experiments, these cuvettes with suspensions were subjected to UV irradiation using a high-pressure mercury lamp.

The specific surface areas of prepared ZnO powders were estimated using experimental data of the long-term CSB adsorption of solutions with high dye content (4.2 mmol/L). These adsorption experiments were carried out for 97 h at room temperature in darkness. The application of concentrated dye solutions provided the full saturation of powders’ surfaces by dye molecules. The surface area of CSB dye molecule is 423 Ǻ² [36] and, assuming monomolecular adsorption, we calculated the values of specific surface areas of ZnO powders.

Dye contents in the solutions were determined during adsorption and photocatalytic experiments by periodical measurements of their absorption spectra using a Perkin Elmer Lambda 650 UV/VIS spectrophotometer.

3. Results

3.1. Crystal Structure and Morphology

Figure 1 demonstrates XRD patterns of the prepared powders. The intensive peaks in diffraction patterns correspond to hexagonal wurtzite ZnO phase (PDF#36-1451). The average ZnO crystal sizes obtained using the Scherer equation are 48 and 30 nm for powders 1 and 2, respectively.

Table 1. Average ZnO crystal sizes and lattice strains in these crystals.

Powder Sample	Average Crystal Size, nm		Lattice Strain, %
	Scherer Equation	Williamson–Hall Method
1	48	71.0 ± 2.4	0.081 ± 0.005
2	30	35.5 ± 1.1	0.066 ± 0.008

shows the results of the calculation of average ZnO crystal sizes and lattice strains in these crystals. The average crystal size in powder 1 is significantly bigger than in powder 2. This fact can be related to the spatial separation of forming ZnO crystals due to the significant gas product formation during PVP oxidation during thermal treatment of PVP-containing raw materials. Furthermore, it is worth noting that there are less lattice strains in ZnO crystals synthesized using the polymer–salt method.

SEM images of prepared powders show significant difference in their morphologies (Figure 2). Powder 1 obtained without PVP addition contains dense micron-scale grains consisting of smaller particles (Figure 2a). Figure 2b demonstrates the structure of powder 2, which consists of small ZnO nanoparticles and contains numerous pores of different sizes.

The small size of ZnO crystals and the formation of porous structure in ZnO powder after PVP addition into the initial solution is explained by the chemical processes proceeding during thermal treatment and formation of ZnO particles.

It is known [23] that PVP additions reduce the sizes of ZnO crystals at their precipitation from solutions. The data of FTIR spectroscopy show that PVP molecules can interact with Zn²⁺ ions forming metal–polymer complexes [37]. In this case, the effect of PVP molecules is determined by the interaction with the surface of the formed particles and the obstacle to their growth and aggregation in the liquid phase.

During polymer–salt synthesis the formation of ZnO crystals proceeds during thermal treatment of organic–inorganic composites at high temperatures (400 ÷ 550 °C) [11]. The observed decrease in the size of ZnO at the addition of PVP into the initial solution is related to the significant volumes of gas products formed during thermal treatment that provides the spatial separation of forming oxide particles [31]. A similar effect of the formation of photocatalytic porous g-C₃N₄ aggregates at the addition of Pluronic P-123 polymer modifier into raw materials was observed in [38].

3.2. Spectroscopic Properties

Figure 3 demonstrates reflection spectra of prepared powders in near UV and visible spectral ranges. The strong light absorption is observed at λ < 380 nm for both powders that is in agreement with the band-gap value of ZnO (E_g~3.36 ÷ 3.37 eV [25,26,27]).

The diffuse reflection spectra were analyzed using the Kubelka–Munk function, which is proportional to the material absorption optical coefficient:

$${\mathrm{F}}_{\mathrm{KM}}=\frac{\left(1-{\mathrm{R}}^2\right)}{2\mathrm{R}}$$

(2)

where R is the diffuse reflectance of the material, measured relative to a completely white body.

The band-gap values of the synthesized coatings were estimated using the Tauc equation [39], which can be written for direct semiconductors as:

(F_KM·hν)² = A(hν − E_g)

(3)

where hν is the photon energy, E_g is the band gap value, A is the constant and F_KM is the absorption coefficient. The plotting of the (F_KM·hν)² vs. hν graphs was used for the determination of E_g values. The intersection of the dotted line with the x-axis shows the band gap value.

The (F_KM·hν)² = f (hν) dependencies obtained for powders 1 (b) and 2 (c) are shown in Figure 3b,c. E_g values are 3.18 and 3.26 eV for ZnO powders 1 and 2, respectively. These values are less than the band-gap values of ZnO crystals, which can be related to some defects in powders structures.

3.3. Photoluminescence

Structural defects in ZnO crystals can significantly affect their photocatalytic properties [40,41]. Defects such as interstitial zinc or oxygen atoms, surface oxygen defects, oxygen antisites and oxygen vacancies are considered to be possible causes of the visible photoluminescence (PL) in ZnO [42,43]. Luminescence is one of the negative processes of electron-hole pair recombination that decrease photocatalytic activity of materials. Therefore, PL spectra are often used to study the formation of structural defects in ZnO and identify their types [43,44,45].

The ratio of volume and surface structural defects in ceramic materials depends on their morphology. The porous structure of powder 2 consisting of small nanoparticles (Figure 2b) can result in relatively high content of surface structural defects.

Figure 4 demonstrates normalized PL spectra of ceramic powders 1 (curve 1) and 2 (curve 2). Numerous emission bands of different structural defects of ZnO crystals are observed in these spectra. Luminescence bands attributed to Zn vacancies (V_Zn) (λ_max~407 nm) [46] and interstitial Zn (Zn_i) (λ_max = 420 ÷ 427 nm [47]) are located in the blue spectral range.

Luminescence bands observed in the green spectral diapason (λ_max = 508; 530 nm) are related to oxygen vacancies [18,43,47]. The authors of [48] found that green (λ_em = 510 nm) emission attributed to singly ionized oxygen vacancies in ZnO is strongly influenced by free carrier depletion at the small ZnO particles’ surfaces. They also reported that the energy band is distorted near the surface of small ZnO particles due to surface energy states and green emission is inactive in this region. The energy band bending near the surface of ZnO nanowires was also discussed in [49]. It was shown [50] that the thickness of the free carrier depletion region on the surface of ZnO nanowires can be about 30 nm that is close to the average size of ZnO crystals in powder 2. Thus, the observed relatively weak green emission in this porous material consisting of small ZnO nanoparticles is related to the features of their electronic structures. It is possible to suggest that these features can affect the ability of ROS photogeneration and photocatalytic activity.

PL spectra of ZnO powders in NIR spectral range are presented in Figure 5a. The intensive luminescence band with λ_max = 1270 nm is observed in the spectrum of ceramic powder 2. This band is characteristic for singlet oxygen and corresponds to the electron transition ¹∆_g − ³Σ_g. The absence of this band in the spectrum of powder 1 indicates a weak ability of this material to generate singlet oxygen under blue light irradiation.

It is worth noting that the singlet oxygen photogeneration by powder 2 occurs under visible light irradiation (λ_ex = 405 nm). The photon energy of this excited light is 3.06 eV, which is remarkably less than the band-gap value of ZnO (~3.37 eV) [25,26,27]. This fact suggests some features of the electronic structure of this powder.

The increase in exciting light intensity significantly enhances the generation of singlet oxygen (Figure 5b). The dependence of the luminescence band (λ_ex = 1270 nm) intensity I_SO, from the intensity of excited radiation (λ_ex = 405 nm) I_ex, is linear.

The deviation from the linearity of the dependence I_SO = f(I_ex) was observed in [34] at the relatively high intensity of exciting light (I_ex > 1200 mW/cm²). This phenomenon was explained by the significant increase in the process of photogenerated electron-hole pairs recombination at the growth of exciting light intensity. Non-linear dependence I_SO = f(I_ex) observed in [34] was similar to the dependencies of the photocatalytic activities of different semiconductors from the intensities of exciting light irradiation reported in [51,52,53]. Observed linear dependence I_SO = f(I_ex) for powder 2 indicates the high effectiveness of enhancing singlet oxygen photogeneration by increasing the exciting light intensity.

3.4. Kinetics of Dye Adsorption

Experiments showed the significant effect of the powder’s morphology on the kinetics of CSB dye adsorption. Figure 6a shows the general view of kinetic dependencies of the solutions discoloration during a relatively long-term adsorption process. As can be seen from this figure, the adsorption capacity of porous powder 2 is significantly higher (more than three times) than the analogous value of powder 1. During the first 40 min of the process the percentages of CSB molecules adsorbed on the surface of powders 1 and 2 were about 2 and 7%, respectively. After 40 min, the surface of powder 1 is completely filled with dye molecules, while powder 2 continues to adsorb the dye. At the end of long-term adsorption process (duration 1 week), powder 2 removed about 30% of CSB molecules from the solution. Observed difference of adsorption capacities was expected, taking in account the features of these powder morphologies.

The experimental results of the long-term CSB adsorption from the solutions with high dye content are given in

Table 2. The experimental results of the long-term CSB adsorption from the solutions with high dye content.

Sample	Powder 1		Powder 2
Weight of powder sample, g	0.01		0.01
Duration of adsorption t, hours	Optical density of solution D	CSB content, mol/L	Optical density of solution D	CSB content, mol/L
0	2.953	0.0042	2.953	0.0042
12	2.447	0.0042	2.950	0.0035
97	1.833	0.0041	2.884	0.0026

. The dye adsorption on the surface of powder 1 is weak and slow. On the contrary, the adsorption on the surface of porous powder 2 is high and this process proceeds quickly.

The values of specific surface areas calculated using experimental data of dye adsorption during 97 h are 76 and 1222 m²/g for powders 1 and 2, respectively. These values ectis corrfor specific surface areas are very high and comparable with the values of specific surface areas of microporous carbons [54,55]. This can suggest that CSB adsorption on the prepared ZnO powders is not monomolecular.

Another difference is that process durations required to achieve the adsorption/desorption equilibrium are 40 min for the dye adsorption on powder 1 and about one week for the adsorption on powder 2. Figure 6a shows that CSB adsorption on the surface of porous powder 2 proceeds quickly during the first 40 ÷ 60 min and then the process rate is reduced remarkably. A two-stage kinetic dependence such as this for the adsorption on porous powder 2 is similar to that described in [17] for the adsorption on porous N-doped TiO₂ photocatalyst.

Many different models (pseudo-first and pseudo-second order models, intraparticle diffusion model) are used for the description of dye adsorption kinetics on the surface of photocatalysts. The selection of the adsorption kinetic model which most adequately describes the experimental results was carried out in many works [11,15,56,57,58]. However, the combination of various kinetic models which were applied at the different adsorption stages provides the best correspondence to the experimental results in some cases [57,58].

In the pseudo-first order kinetic model the adsorption rate is described by the expression [59]:

$$\frac{{\mathrm{dq}}_{\mathrm{t}}}{\mathrm{dt}}={\mathrm{k}}_1\times \left({\mathrm{q}}_{\mathrm{e}}-{\mathrm{q}}_{\mathrm{t}}\right)$$

(4)

where q_t (mg/g) is the amount of dye adsorbed by 1 g of sorbent in time t, q_e is the equilibrium adsorption capacity of the sorbent, k₁ (min⁻¹) is the adsorption rate constant and t is the duration of the adsorption (min). The adsorption rate decreases as the surface is filled with dye molecules.

The pseudo-second order equation is also applied for the description of the adsorption kinetics on photocatalysts [57,58,60]. In integral form of this equation for the adsorption on the external surface of the porous material can be written as:

$$\frac{\mathrm{t}}{{\mathrm{q}}_{\mathrm{t}}}=\frac{1}{{\mathrm{k}}_2\times {\mathrm{q}}_{\mathrm{e}}^2}+\frac{\mathrm{t}}{{\mathrm{q}}_{\mathrm{e}}}$$

(5)

where k₂ is the second-order adsorption rate constant, q_e is the maximum equilibrium adsorption capacity of the material (mg/g) and q_t is the amount of dye adsorbed by 1 g of sorbent in time t (min).

The intraparticle diffusion model can be applied by using the equation that is represented by:

q_t = k_id t ½ + C

(6)

where k_id (mg/g·min) is the intra-particle diffusion rate constant and C is the thickness of boundary layer [58,59].

Figure 7 demonstrates dependencies ln(q_{e −} q_t) = f(t) (a), t/q_t = f(t) (b) and q_t = f(t^1/2) (c) for log-term (adsorption duration 7200 min) CSB adsorption on the surface of powders 1 and 2. This figure shows that kinetic models of pseudo-first order and intraparticle diffusion do not correspond to experimental results (R² < 0.9). The pseudo-second order kinetic model describes the experimental long-term adsorption data successfully with the k₂ constant rate values of 7.32 g/mg·min and 0.13 g/mg·min for powders 1 and 2, respectively.

With the application of effective photocatalysts, the organic dye decomposition proceeds quickly and the duration of the photocatalytic process does not often exceed 60 ÷ 120 min. Therefore, more detailed study of the dye adsorption during this first stage of the process could be important for the consideration of its photocatalytic decomposition features.

The fast first stage of the process can correspond to the dye adsorption on the external surface of the porous material. Figure 8a,b demonstrates the results of the application of pseudo-first (a) and pseudo-second order kinetic models for the description of the first, short-term stage of dye adsorption. Both models correspond to experimental data successfully.

Figure 8b shows dependencies t/q_t = f(t) for the first stage of dye adsorption on the powders. These dependencies are linear for the adsorption on both powders but their slopes are different. Figure 8a,b demonstrates that both kinetic models (pseudo-first order and second-order models) describe experimental results successfully.

The authors of [56] compared the adequateness of different kinetic models of CSB adsorption during 120 min on the surface of ZnO nanoparticles. The results of [59] showed that the pseudo-second order kinetic model is the most correct for describing CSB adsorption on ZnO nanoparticles. Our results (Figure 8b) fully agree with the data [56]. The pseudo-second-order model suggests that the adsorption involves a chemisorption process in addition to physisorption [61]. The interaction between Zn²⁺ ions and CSB molecules, which was observed in [15], can be responsible for this chemisorption process.

However, it is worth noting that constant rate values are different for long-term and short-term stages of dye adsorption. So, for porous powder 2, the constant rate values are k₂ = 0.13 g/mg·min for the long-term process and k₂ = 0.41 g/mg·min for the dye adsorption on the external material surface. The variation of appropriate kinetic models and constants rates for the most adequate adsorption description at the different process stages is somewhat common [62].

3.5. Photocatalytic Dye Decomposition

Figure 9 shows kinetic dependencies of photocatalytic dye decomposition under UV irradiation in the presence of ZnO powders. The photolysis of CSB molecules under UV irradiation without photocatalysts is a very slow process and the amount of dye decomposed by this mechanism is negligible [15,31].

The photocatalytic dye decomposition kinetics are usually described by the Langmuir–Hinshelwood model and approximated by the kinetic equation [63,64]:

$$-\frac{\mathrm{dC}}{\mathrm{dt}}=\frac{{\mathrm{k}}_{1\mathrm{p}}{\mathrm{K}}_{\mathrm{a}}\mathrm{C}}{1+{\mathrm{K}}_{\mathrm{a}}\mathrm{C}}$$

(7)

where C is the current dye concentration (mM) at time t (min), k_1p is the constant rate of the process and K_a is the adsorption/desorption equilibrium constant. At a low dye concentration (C << 10⁻³ M), Equation (4) simplifies to a pseudo-first order process rate equation:

$$\ln \left({\mathrm{C}|\mathrm{C}}_0\right)={\mathrm{k}}_1{\mathrm{K}}_{\mathrm{a}}\mathrm{t}={\mathrm{k}}_{\mathrm{app}}\mathrm{t}$$

(8)

where k_app is the pseudo-first order rate constant and C₀ is the initial dye concentration (mM). Figure 9 shows that this equation describes experimental data successfully (R² > 0.9).

Photodecomposition rates are significantly different for the application of grainy powder 1 and porous powder 2. The application of porous powder 2 provides significantly faster dye decomposition than grainy powder 1. This fact is related to the faster dye adsorption with the ensuing dye photodecomposition on the surface of powder 2 (Figure 6b) due to more intensive ROS photogeneration by this porous material (Figure 5).

Furthermore, the difference in powders’ electronic structures and structural defects in their lattices can affect their photocatalytic activities. The authors of [42,43] found that the presence of structural defects in ZnO crystals decreases their photocatalytic activity. On the contrary, the enhancement of photocatalytic properties with increased amounts of structural defects in ZnO nanorods was described in [8]. This disagreement can be to the different ratios of volume to surface structural defects in ZnO materials synthesized and studied in [42,43,45].

The comparison of the kinetic dependencies of dye adsorption by porous powder 2 (Figure 6b, curve 2,) and dye photodecomposition (Figure 9, curve 2) shows the significant difference in the rates of these process. The dye photodecomposition proceeds significantly faster than its adsorption. This phenomenon was observed early in [8,10]. It was proposed that the dye photodecomposition proceeds on the surface of nanoporous ZnO as it does in the solution.

The mechanism of photocatalytic dye decomposition in the solution volume described in [26] comprises the following stages: (1) the excitation of a dye molecule by electron or hole generated by the photocatalyst, (2) diffusion of an excited dye molecule into the solution and interaction with another dye molecule and (3) decomposition of another dye molecule in the solution. However, it is necessary to note that excitation transfer from the electrons or holes to dye molecules should proceed on the photocatalyst surface. This means that the dye molecule should primarily be adsorbed on this surface and the photocatalytic dye decomposition rate cannot be higher than dye adsorption rate. Thus, the mechanism proposed in [26] cannot describe the observed difference in rates of photocatalytic dye decomposition and its adsorption (Figure 6b, curve 2 and Figure 9, curve 2).

Table 3. Rate constants of CSB photodecomposition under UV irradiation in the presence of different photocatalysts in solutions.

Sample	Rate Constants of CSB Photodecomposition kapp, min−1	References
Powder 1	0.025	Present study
Powder 2	0.042	Present study
Powder ZnO 80.16 mol.% + ZnAl2O4 19.83 mol.% + CuO 0.04 mol.%	0.021	[63]
Powder ZnO 20.81 mol.% + ZnAl2O4 79.18 mol.% + CuO 0.01 mol.%	0.005	[63]
ZnO-MgO	0.062	[64]
ZnO nanoflowers	0.032	[65]

presents the kinetics parameters of the photocatalytic decomposition of CSB dye in aqueous solutions using different photocatalysts. These data show that powder 1 obtained without PVP addition demonstrates a photodecomposition rate constant (k = 0.025 min⁻¹) comparable with rate constant values reported previously with the application other photocatalysts based on ZnO. A significantly higher constant rate of the photocatalytic dye decomposition (k = 0.042 min⁻¹) is observed with the application of powder 2.

4. Discussion and Conclusions

ZnO powders with different morphologies were prepared with the thermal decomposition of zinc nitrate and with the polymer–salt method. The addition of PVP into initial solutions significantly reduces the size of ZnO crystals (from 48 to 30 nm) forming during thermal treatment and changes the morphology of obtained powders from grainy structure to porous.

Photoluminescence spectra of ZnO in the visible spectral range contain emission bands of structural defects: oxygen vacancies, zinc vacancies, etc. The relative ratio between various types of these defects is remarkably different in grainy and porous powders.

Porous ZnO powder demonstrates the ability for singlet oxygen generation under visible (405 nm) irradiation. The obtained dependence of the intensity of singlet oxygen photogeneration by the porous ZnO from the excited light intensity is linear.

The features of adsorptive and photocatalytic properties of nanoporous ZnO consisting of small nanoparticles are determined by its morphology and crystal structure and include a few factors:

• Small size of ZnO nanoparticles determines high material specific surface area and adsorption capacity.
• Some changes of the electronic structure of the nanomaterial and relatively high ratio between surface and volume structural defects compared with bulk ZnO.
• The kinetics of organic dye adsorption from solutions by the nanoporous material includes a fast covering of the external surface of porous particles by dye molecules and slow dye diffusion and adsorption inside nanopores.
• Photocatalytic dye decomposition is faster than its adsorption on the surface of nanoporous ZnO. Therefore, it is possible to suggest that the dye photodecomposition proceeds as on the surface of nanoporous ZnO as in the solution.

The kinetics of photocatalytic dye decomposition are described by the Langmuir–Hinshelwood model. The presence of porous structure significantly increased the rate of photocatalytic dye decomposition. A high constant rate (0.042 min⁻¹) is observed with the application of porous ZnO photocatalyst.