Photocatalytic degradation of metoprolol in water suspension of TiO₂ nanopowders prepared using sol–gel route

1 Introduction

Pharmaceuticals, as a large group of human and veterinary medicinal compounds, are considered as an emerging environmental problem, due to long term usage worldwide. Its continuous input and persistence to the aqueous system, even at low concentrations, may result in a potential risk for aquatic and terrestrial organisms. Cardiovascular drugs are among most frequently detected pharmaceuticals in wastewaters, where, according to literature, β-blocker metoprolol is listed in concentration of 10–390 ng L⁻¹ [1]. Although in trace concentrations, those pharmaceuticals are persistent against biological degradation and they retain their chemical structure long enough and their presence in the environment is considered potentially dangerous.

Metoprolol tartrate salt (1-[4-(2 methoxyethyl)phenoxy]-3-(propan-2-ylamino)propan-2-ol tartrate (2 : 1), CAS No 56392-17-7, (C₁₅H₂₅NO₃)₂ C₄H₆O₆, M _r = 684.81, MET) is a selective β-blocker that is used to treat a variety of cardiovascular diseases, such as hypertension, coronary artery disease, and arrhythmias, by blocking the action of epinephrine and norepinephrine on the β-adrenergic receptors in the body, primarily in the heart [2]. Wide use of MET in recent years has a consequence in increasing concentrations of this pharmaceutical in aqueous effluents [3, 4]. MET shows slow direct phototransformation and/or hydrolysis [5, 6]. An efficient way to deal with this problem is degradation of the drug by advanced oxidation processes (AOPs) based on the formation of hydroxyl (^•OH) and other radicals that oxidize toxic and non-biodegradable compounds to the different co-products, and, in the course of time, to inert final products [7]. MET contains a secondary amine group and a weakly/moderately activated aromatic ring that are probable targets of molecular ozone and attacks of ^•OH [2]. Yang et al. [8] investigated degradation of selected β-blockers (atenolol, metoprolol, and propranolol) in aqueous suspensions of TiO₂ Degussa P25 and proposed a preliminary mechanism of degradation of these compounds. Romero et al. [9, 10] have also investigated and compared degradation intermediates of metoprolol and propranolol by AOPs.

Titanium dioxide (TiO₂) is recognized as catalyst employed for almost all the pharmaceuticals photocatalytic treatment studies, as inexpensive, commercially available at various crystalline forms and particle characteristics, non-toxic and photochemically stable material [11]. Degussa P25, a commercially available powder is a mixture of anatase and rutile phases (with 80% of anatase approximately [12]), showing exceptional photocatalytic performance, superior to other grades of TiO₂ [13].

We synthesized TiO₂ nanopowders via sol–gel routine to obtain photocatalyst efficient in metoprolol degradation. The sol–gel is solution-based technique, where material structure is created through chemical reactions in liquid state. It has recently received more attention in production of high performance nanostructured materials, where morphology, surface area, average nanocrystallite size, crystallinity, and phase structure are controlled by sol–gel synthesis parameters [14]. The properties of sol–gel synthesized TiO₂ nanopowders are determined by the key sol–gel processing parameters such as the precursor type, pH value in hydrolysis process, duration and temperature of hydrolysis (aging), type of alcogel, duration and temperature of drying, and calcination conditions (heating rate, temperature, duration and cooling rate) [15].

In this paper we investigate the influence of sol–gel synthesis condition on photocatalytic properties of TiO₂ nanopowders, by varying some synthesis parameters (precursor, alcohol, basic solution, and calcination temperature) aiming to obtain photocatalyst for degradation of metoprolol, efficient enough in comparison to Degussa P25. Several methods of characterization, such as the X-ray diffraction (XRD), scanning electron microscopy (SEM), energy dispersive X-ray spectroscopy (EDS), Brunauer-Emmett-Teller (BET) measurements, Fourier transform infrared (FTIR) and Raman spectroscopy, are employed to correlate structural and morphological properties of synthesized TiO₂ nanopowders and their photocatalytic activity under ultraviolet (UV) irradiation.

2 Experimental

2.1 Powder synthesis

The TiO₂ nanopowders used as catalysts in this study are synthesized by the sol–gel method, which typically involves hydrolysis of a solution of a precursor molecule in order to obtain a suspension of colloidal particles—gel (at first in the form of hydrogel and then in the form of alcogel). Titanium tetrachloride (TiCl₄) and tetrabutyl titanate (Ti(OC₄H₉)₄) are the precursors commonly used in sol–gel synthesis, which required different procedures in alcogel production:

(1) In synthesis with tetrabutyl titanate as precursor, ethanol (EtOH) is used as the solvent, hydrochloride acid as the catalyst, and water for hydrolysis. The hydrolysis and polycondensation reactions of Ti(OC₄H₉)₄ were carried out at the room temperature, whereas the measured pH value of solution was 7.0. The reagent molar ratio was Ti(OC₄H₉)₄:EtOH:HCl:H₂O = 1:15:0.3:4 [16]. After the gelation, the wet gels were dried at 80 °C;

(2) In synthesis with titanium tetrachloride as the precursor, Ti(OH)₄ hydrogel was obtained by hydrolysis of TiCl₄ at 0 °C with controlled addition of 2.5 wt% alkalic solution (NH₄OH or NaOH) into the aqueous solution of TiCl₄ (0.3 M) and careful control of the pH value (9.3) of the solution [15]. After aging (fixed time of 5 h) [17], as-prepared hydrogel was filtered and rinsed out with distilled water until complete removal of chloride ions. The absence of Cl^– ions was checked using 0.1 M solution of AgNO₃. Obtained Ti(OH)₄ hydrogel was converted to its alcogel by repeated exchange with corresponding alcohol for several times. Also, different alcohols were used to investigate the influence of the type of alcogel on properties of photocatalysis: methanol, ethanol, 2-propanol, amylalcohol and glycerol. The alcogels were dried at 280 °C during 4 h.

The TiO₂ nanopowders produced by one of the procedures described above are amorphous. During the calcination process at temperature of 350 °C or higher, the crystalline anatase phase is formed [18]. As calcination temperature approaches ~600 °C, the phase transition from anatase to rutile starts [19]. However, high calcination temperature results in the growth of nanocrystalline particles and the rapid decrease of specific surface area. To investigate the influence of temperature of calcination, few samples synthesized with the same alcohol (amylalcohol) were exposed to different calcination temperatures (400, 550 and 650 °C). The duration of calcination process (7 h) and temperature gradient of the heating (135 °C h⁻¹) were the same in all the experiments.

The scheme of the TiO₂ production by sol–gel process can be written as follows:

$${\text{TiCl}}_{4} \;{\text{or}}\;{\text{Ti}}({\text{OC}}_{4}{\text{H}}_{9} )_{4} \to {\text{Ti}}({\text{OH}})_{4} \to{\text{alcogel}}{\mathop{\rightarrow}\limits^{{\text{calcination}}}}{\text{TiO}}_{2}$$

(1)

The samples synthesized by described procedures are listed in Table 1, together with variable parameters of synthesis (precursor, alcohol, basic solution, and calcination temperature).

Table 1 The list of sol–gel synthesis parameters

All chemicals in this experiment were analytical grades, used as received. Tetrabutyl titanate 99% was produced by Acros Organics, TiCl₄ and glycerol by Merck Chemicals, amylalcohol by Lacheme, ethanol by J. T. Backer, methanol and 2-propanol by Centrohem, 25% NH₄OH by Carbo Erba and NaOH by MP Hemija.

2.2 Characterization methods

Powder XRD patterns of TiO₂ nanopowders were recorded by Siemens X-ray Diffractometer (Kristaloflex 500) with Ni filtered CuKα radiation and using Si as an external standard. The measurements were performed in the 2θ range from 20° to 80° in a continuous scan mode with a step width of 0.02° and at a scan rate of 1° 2θ min⁻¹.

Morphology of the catalyst is determined by SEM. Gold coated particles were observed on SEM type JEOL JSM-6460LV with the operating voltage of 20 keV. Composition/quality of TiO₂ was analyzed on SEM with EDS detector. EDS are performed by INCAx-sight detector and “INAx-stream” pulse processor (Oxford Instruments).

The powder specific surface area of the samples was calculated following the multipoint BET procedure on Quantachrome ChemBet-3000 setup. The particle size distribution (PSD) was derived from nitrogen adsorption/desorption isotherms obtained at 77 K.

The IR reflection spectra of TiO₂ nanopowders pressed into pellets were measured at room temperature using a BOMEM-DA8 spectrometer in the spectral range between 400 and 4,000 cm⁻¹.

Raman scattering measurements of TiO₂ samples was performed in the backscattering geometry at room temperature in the air using Jobin–Yvon T64000 triple spectrometer, equipped with a confocal microscope and a nitrogen-cooled charge coupled device detector. The spectra excited by 514.5 nm line of Ar⁺/Kr⁺ ion laser with output power less than 5 mW to avoid local heating due to laser irradiation.

2.3 Photocatalytic activity

The photocatalytic activity of the TiO₂ nanopowders was evaluated by the degradation of the solution of metoprolol tartrate salt (≥99%, Sigma–Aldrich). Photocatalytic degradation was carried out in a cell made of Pyrex glass (total volume of ca. 40 mL, liquid layer thickness 35 mm), with a plain window on which the light beam was focused. The cell was equipped with a magnetic stirring bar and a water circulating jacket. A 125 W high-pressure mercury lamp (Philips, HPL-N, emission bands in the UV region at 304, 314, 335 and 366 nm, with maximum emission at 366 nm), together with an appropriate concave mirror, was used as the radiation source. The output for the mercury lamp was calculated to be ca. 8.8 × 10⁻⁹ Einstein mL⁻¹ min⁻¹ (potassium ferrioxalate actinometry).

In a typical experiment the initial MET concentration was 0.05 mM and the TiO₂ loading was 1.0 mg mL⁻¹, unless otherwise is stated. The total suspension volume was 20 mL. The aqueous suspension of TiO₂ was sonicated (50 Hz) in the dark for 15 min before illumination, to uniformly disperse the photocatalyst particles and attain adsorption equilibrium. The suspension thus obtained was thermostated at 25 ± 0.5 °C in a stream of O₂ (3.0 mL min⁻¹), and then irradiated. During irradiation, the mixture was stirred at a constant rate under continuous O₂ flow. Control experiments were carried out under O₂ flow but by stopping the irradiation showed that there were no losses of volatile compounds during the degradation. All experiments were performed at the natural pH.

For the LC–DAD kinetic studies of MET photodegradation, aliquots of 0.50 mL were taken from the reaction mixture at the beginning of the experiment and at regular time intervals. Aliquot sampling caused a maximum volume variation of ca. 10% in the reaction mixture. The suspensions containing photocatalyst were filtered through Millipore (Milex-GV, 0.22 μm) membrane filter. Lack of adsorption of MET on the filters was preliminary checked. After that, a 20-μL sample was injected and analyzed on an Agilent Technologies 1100 Series liquid chromatograph, equipped with an Eclypse XDB-C18 column (150 mm × 4.6 mm i.d., particle size 5 μm, 25 °C). The UV/vis DAD detector was set at 225 nm (wavelength of MET maximum absorption). The mobile phase (flow rate 0.8 mL min⁻¹) was a mixture of acetonitrile (ACN, 99.8%, J. T. Baker) and water (with the following gradient: 0 min 15% ACN, 5 min 30% ACN and 5 min 30% ACN; post time 3 min), the water being acidified with 0.1% H₃PO₄ (85%, Lachema, Neratovice).

Commercially available TiO₂, Degussa P25 (75% anatase and 25% rutile, specific area of 50 m² g⁻¹, and average particle size about 20 nm, according to the producer’s specification), was used for the purpose of comparison.

3 Results

3.1 XRD

The XRD patterns of chosen TiO₂ nanopowders are presented in Fig. 1. The diffraction peaks present in all samples coincide to anatase TiO₂ structure (JCPDS card 21-1272, space group I4₁/amd). Their positions are slightly shifted, indicating a change in lattice parameters in comparison to the bulk values for anatase (a ₀ = 0.37852 nm, c ₀ = 0.95139 nm), as shown in Table 2. Consequently, the lattice volume of anatase in all synthesized nanopowders is smaller than in its bulk counterpart (V ₀ = 136.3127 × 10⁻³ nm). It has been noticed that positive pressure caused by large surface hydration could be a primary reason for lattice contraction volume in anatase nanostructures [20, 21].

Fig. 1

The XRD diffractograms of some synthesized TiO₂ nanopowders

Table 2 Lattice parameters, crystallite size L (obtained from Scherrer method and structure analysis by Power Cell), and microstrain ε (also from Power Cell) for anatase phase in some TiO₂ nanopowders

The average crystallite size of the anatase nanoparticles, estimated by using Scherrer’s equation, is in range from 13 to 17 nm, as listed in Table 2. A whole powder pattern structure refinement using the structure analysis software Power Cell [22] was also carried out. As expected, structural analysis gives slightly higher values for crystallite size in comparison to Scherrer method, because the effects of microstrain are taken into account.

Note that diffraction peaks of brookite phase are not seen in XRD patterns of TiO₂ nanopowders, whereas the peaks corresponding to the rutile phase (JCPDS card no. 21-1276, space group P4₂/mnm) are observed in the diffractogram of sample T(CAN)₆₅₀. The amount of rutile phase in this sample is estimated to be ~30% according to Power Cell. The lack of the reliable results for crystallite size and microstrain in T(CAN)₆₅₀ is probably related to the presence of rutile phase in this sample.

3.2 SEM/EDS

The SEM images of two synthesized TiO₂ nanopowders (T(CAA)₅₅₀, T(CAN)₆₅₀) and Degussa P25 are presented in Fig. 2. As can be seen from the micrographs, most of the particles in this sample have a spherical shape, with a grain diameter, roughly estimated to be about 23 nm in synthesized samples, and around 29 nm in Degussa P25. Table 3 presents the chemical composition of the same nanopowders, estimated by EDS spectra (Fig. 3, upper row), which correspond to framed area of micrographs (Fig. 3, lower row). In the samples analyzed here, oxygen weight percent is higher than in stoichiometric TiO₂ (40 wt%), with the exception of the sample T(CAN)₅₅₀. Among synthesized samples, the highest value of oxygen weight percent (but still lower than value in Degussa P25) is registered in T(CAN)₆₅₀. The EDS analysis has also shown the presence of sodium (Na) and very low concentration (<1 wt%) of chlorine (Cl) in the samples synthesized from NaOH as basic solution and TiCl₄ as precursor (Table 3). Note here also that hydrogen cannot be detected by EDS method.

Fig. 2

SEM micrographs of TiO₂ nanopowders T(CAA)₅₅₀ (a) and T(CAN)₆₅₀, (b), compared to Degussa P25 (c)

Table 3 EDS results for some TiO₂ nanopowder samples and Degussa P25

Fig. 3

SEM images with corresponding EDS spectra of TiO₂ nanopowders T(CAA)₅₅₀ (a), T(CAN)₆₅₀ (b) and Degussa P25 (c)

3.3 BET

The values of the BET specific surface areas (S _BET ) of some TiO₂ nanopowders are given in the Table 4, together with pore volume (V _p ) and pore diameter (D _p ). The sample T(CAN)₄₀₀ exhibits the lowest specific surface area (32 m² g⁻¹), whereas S _BET ranges between 51 and 63 m² g⁻¹ in the rest of the samples. All nanopowders listed here, as well Degussa P25, are mostly mesoporous (pore diameters of 2–50 nm, S _BET  ≈ S _meso ). However, mean pore diameter of synthesized samples, which is in the range of 6–7 nm, is much lower than the corresponding one of Degussa P25 (31.5 nm [12]).

Table 4 Porous properties of TiO₂ samples: specific surface area (S _BET ), pore volume (V _p ) and pore diameter (D _p )

The PSD of specified synthesized samples are bimodal (Fig. 4), with the first mode peaked in the range 2–3 nm, and the second mode around 4 nm. Among these samples, the T(CAN)₆₅₀ contains the greatest number of pores with diameter higher than 3 nm, whereas the number of such pores in T(CAN)₄₀₀ is the smallest.

Fig. 4

The pore size distribution of some synthesized TiO₂ nanopowders

3.4 FTIR

The IR reflectivity spectra of representative TiO₂ nanopowders in the range of 450–4,000 cm⁻¹ are presented in Fig. 5. The spectra are dominated by the features in the range of 450–1,000 cm⁻¹, which can be related to the surface vibration of Ti–O bonds [23, 24]. It is obvious that more features in this region appear in the samples with mixed anatase/rutile phase (T(CAN)₆₅₀ and Degussa P25), as rutile has one more IR active mode than anatase in specified region [25]. Also, nanostructure of analyzed powders has strong influence on the shape of IR reflectivity spectra [26], but this is beyond the scope of this study.

Fig. 5

The IR reflectivity spectra of representative TiO₂ nanopowders. The inset: enlarged part of the IR spectra in the region from 1,200 to 4,000 cm⁻¹

The reflectivity spectra of the same samples in the region of 1,200–4,000 cm⁻¹ are shown enlarged in the Inset of the Fig. 5. The main spectral features in this region appear around 1,260 cm⁻¹, between 1,400 and 1,800 cm⁻¹, and in the ranges of 2,800–3,000 cm⁻¹ and 3,200–3,800 cm⁻¹. The band at about 1,260 cm⁻¹, which is most prominent in the IR spectrum of Degussa P25, was assigned to carbonate [27] or nitrate [28] species at TiO₂ surface. Small bands in the range of 1,400–1,800 cm⁻¹ most probably originate from carbonates or hydrogen carbonates groups in TiO₂ [29]. The bands in the range of 2,800–3,000 cm⁻¹ can be assigned to the symmetric and asymmetric stretching mode of the C–H bond [30], originating from residues which remain in TiO₂ samples even after calcination.

As TiO₂ surface O–H group nature and distribution are in many cases determining factor of catalyst behaviour, a considerable research has been devoted here to characterizing these groups present at the surface of TiO₂. According to many FTIR studies published in the literature, the vibration band between 3,000 and 3,150 cm⁻¹ is associated with the O–H stretching mode of water molecules [31], while that at about 1,630 cm⁻¹ is assigned to O–H bending modes [23 and references therein]. The bands in frequency range of 3,150–3,500 cm⁻¹ are usually attributed to H-bounded O–H groups, and those in the 3,400–3,600 cm⁻¹ region to mixed O–H groups [31]. The bands in the 3,600–3,800 cm⁻¹ region are due to stretching modes of different types of free hydroxyl groups adsorbed on the crystalline surfaces of titania or on defects located at the surfaces [23 and references therein]. It should be noted that IR reflectivity spectra of TiO₂ nanopowders, shown in Fig. 5, enable identification of specified surface groups, but do not allow determination of the relation between the amounts of different kinds of these groups at the surface of TiO₂ nanoparticles.

3.5 Raman scattering

The Raman spectra of the samples synthesized by sol–gel method, listed in Table 1, were the subjects of analysis, which indicated anatase as a dominant phase in all samples. The spectrum of the sample T(CAA)₅₅₀ shown in Fig. 6a illustrates characteristic anatase Raman modes in the region of 50–1,000 cm⁻¹ and so called C–H and O–H spectral region in the 1,000–4,000 cm⁻¹ frequency range.

Fig. 6

a Raman spectrum of T(CAA)₅₅₀ nanopowder. Spectra of TiO₂ nanopowders divided in three characteristic regions: b around anatase E _g mode, c 210–380 cm⁻¹ (brookite modes), and d 350–700 cm⁻¹ (anatase, rutile and sodium titanate modes). The intensities of the spectra are normalized to 1 in the first region (E _g ) and multiplied by 50 and 7 in following two regions to pronounce corresponding features of the spectra. Anatase, rutile, brookite and sodium titanate are labelled as A, R, B, and ST, respectively

Generally, the intensity and linewidth of Raman modes provide the information about crystalline quality of the material. According to this, among synthesized samples, T(CAN)₆₅₀ and T(BE)₅₅₀, with the most intensive and narrowest characteristic anatase modes in Raman spectra, have the crystalline quality close to Degussa P25.

In order to point out some features related to different TiO₂ phases (anatase, brookite and rutile) and the presence of residuals due to synthesis process, Raman spectra of some nanopowders (T(CAA)₅₅₀, T(CAN)₄₀₀, T(CAN)₅₅₀, and T(CAN)₆₅₀), together with Degussa P25, are shown in Fig. 6b–6d. The spectra are divided in three characteristic regions because of different intensities in each of them.

The first region, around anatase E _g Raman mode (~143 cm⁻¹), is shown in Fig. 6b. The relationship between Raman shift and full width at half maximum (FWHM) of this mode, for all TiO₂ nanopowders, is presented in Fig. 7. These results show that the Raman shift and FWHM of sol–gel synthesized samples are greater than in Degussa P25 nanopowder: Raman shift is positioned between 142.6 cm⁻¹ in T(CAN)₆₅₀ and 144.1 cm⁻¹ in T(CGA)₅₅₀, whereas the variation of FWHM ranges from 9.4 cm⁻¹ in T(BE)₅₅₀ to 12.6 cm⁻¹ in T(CPA)₅₅₀. Beside the crystalline quality, Raman shift and linewidth of E _g mode in anatase nanocrystals at given temperature also depend on confinement effects, due to the nanocrystalline size, microstrain and nonstoichiometry [32, 33]. In order to estimate which of those effects has the greatest influence on blueshift and broadening of Raman E _g mode of synthesized nanopowders in comparison to corresponding values in bulk anatase (141 cm⁻¹ and 8 cm⁻¹ [34]), the relationship between Raman shift and FWHM is calculated by phonon confinement model (PCM) (Fig. 7). The particle size used in the calculation belongs to the range of 17–25 nm, corresponding to the size of nanocrystallites in synthesized powders, estimated from XRD (Table 2). However, for the specified crystallite sizes, the values of Raman shift and FWHM calculated by PCM are lower than experimental ones, even when PCM with inhomogeneous strain is applied (open circles in Fig. 7) [35, 36]. Although widening of E _g mode can be the consequence of lower crystalline quality, it is obvious that, in addition to phonon confinement and microstrain, the nonstoichiometry must be included in the explanation of blueshift and broadening of Raman E _g mode in most of our samples. Namely, it is well known that this mode is blueshifted and broadened due to nonstoichiometry in anatase nanopowders [37].

Fig. 7

The relationship between Raman shift and FWHM of E _g mode of anatase phase in TiO₂ nanopowders. The experimental values are denoted by stars (the calcination temperature is intentionally omitted in the names of all samples calcined at 550 °C for the clarity). Open symbols denote the values calculated by PCM without strain (triangles), and including strain effect (circles)

In order to investigate the presence of brookite phase by Raman spectroscopy, the second region of the Raman spectra, from 210 to 380 cm⁻¹ is analyzed. This part of the spectra is shown in Fig. 6c for previously chosen set of samples. The brookite modes are located as follows: A _1g (~247 cm⁻¹), B _3g (~288 cm⁻¹), B _1g (~322 cm⁻¹), and B _2g (~366 cm⁻¹) [38]. Low intensities and large widths of these modes indicate great disorder and partial amorphization of brookite in all the samples. To estimate the amount of brookite phase, the sum of the integrated intensities of Lorentzian peaks, originating from the brookite modes (ΣI _B ), was compared to the intensity of Lorentzian peak related to the B _1g mode of anatase phase (I _A (B _1g )). Although the estimation of brookite modes intensities is unreliable to some extent, due to mode overlapping, a variation in brookite content with synthesis conditions may be roughly estimated from the intensity ratio of brookite to anatase modes, ΣI _B /I _A (B _1g ), shown in Fig. 8. In such a way, the amount of brookite phase in the samples synthesized with NH₄OH basic solution is higher than in the samples synthesized with NaOH basic solution. Note also that additional mode at about ~275 cm⁻¹, which can be related to the sodium titanate [39], is registered in the Raman spectra of the samples synthesized from NaOH basic solution.

Fig. 8

The ratio of the total brookite modes intensity, ΣI _B , and the anatase B _1g Raman mode intensity I _A (B _1g ) in TiO₂ nanopowders. The inset: Lorentzians of brookite modes in the experimental Raman spectra of sample T(CPA)₅₅₀. (The calcination temperature is omitted in the names of the samples calcined at 550 °C for clarity)

The Raman spectra in the region from 350 to 700 cm⁻¹ (Fig. 6d) of all TiO₂ nanopowders are dominated by anatase modes B _1g (~399 cm⁻¹), A _1g  + B _1g (~518 cm⁻¹), and E _g (~639 cm⁻¹) [40]. However, the modes of rutile phase, E _g (~447 cm⁻¹) and A _1g (~612 cm⁻¹) [40], appear in this region in the spectra of samples T(BE)₅₅₀ (synthesized with tetrabutyl titanate as a precursor) and T(CAN)₆₅₀ (calcined at 650 °C), as well as in Degussa P25.

Compared with IR spectroscopy [41], Raman spectroscopy has rarely been applied to study the surface complex formation at anatase nanoparticles [42]. Namely, in C–H and O–H spectral region, Raman features are hard to resolve as sample luminescence is often too intensive. Nevertheless, we demonstrate here the applicability of Raman spectroscopy to identify specified surface groups and estimate their amount at the surface of TiO₂ nanoparticles.

Raman spectra of representative TiO₂ nanopowders in the region between 1,000 and 4,000 cm⁻¹ are shown in Fig. 9. This region is dominated by the broad feature between 2,800 and 3,800 cm⁻¹ and less intensive mode at about 1,650 cm⁻¹. According to many vibrational studies, carried out to determine the characteristics of water molecules and hydroxyl groups adsorbed on anatase surfaces [29 and references therein], broad Raman structure in the frequency range of 2,800–3,800 cm⁻¹ can be assigned to O–H stretching vibration of water molecules adsorbed on the surface of the particles and condensed in the pores of TiO₂ nanopowders [42]. The intensity of this structure is the highest in the Raman spectra of Degussa P25 and T(CAA)₅₅₀, followed by T(CAN)₄₀₀, implying that the amount of adsorbed water molecules is the greatest in those samples. Also, the feature peaked at the frequency between 1,625 and 1,670 cm⁻¹, which can been assigned to the bending modes of adsorbed water [29], is the most intensive in Degussa P25 and T(CAA)₅₅₀. Complex shape and wideness of the spectral features described above point out to the variety of components in this range, indicating the presence of several types of surface sites bearing hydroxyl groups and adsorbed H₂O molecules [43, 44].

Fig. 9

The Raman spectra of TiO₂ nanopowders in the C–H and O–H spectral region (1,000–4,000 cm⁻¹). The spectra of some powders are upshifted for clarity

In addition, several sharp peaks are noticed in the Raman spectra shown in Fig. 9. A sharp peak at ~1,055 cm⁻¹, registered in the spectra of T(CAN)₄₀₀, T(CAN)₅₅₀ and T(CAN)₆₅₀, most probably belongs to stretching vibration of NO₃ ⁻, due to the presence of NaNO₃ [45], as a consequence of using NaOH basic solution in the synthesis process. The peaks in the frequency range 1,200–1,500 cm⁻¹, most prominent in Degussa P25 and T(CAA)₅₅₀, can be associated with carbonate groups [23], whereas the peak at 2,940 cm⁻¹ can be assigned to the stretching mode characteristic for CH- species [46].

However, low intensity peak located at ~3,700 cm⁻¹ noticeable in all the samples except T(CAN)₄₀₀, is the most interesting one in our analysis. Namely, this feature, which is assigned to the O–H stretching mode of isolated (free) hydroxyls [29], can be related to the ability of the anatase samples to form the surface O–H groups with the important influence on their catalytic activity [47]. According to several authors, the hydroxyl groups with vibration frequency at about 3,700 cm⁻¹ are the most photocatalytically active among all hydroxyl groups at the surface of TiO₂ nanoparticles [23].

The analysis of Raman data presented above has demonstrated the ability of Raman spectroscopy not only to characterize TiO₂ in terms of structure and crystallite size, the presence of different crystal modifications, microstrain and nonstoichiometry, but also to recognize surface O–H groups in TiO₂ nanopowders, important for photocatalytic activity.

3.6 Photocatalysis

The influence of the different synthesis parameters on the photocatalytic activity of TiO₂ nanopowders was studied on the example of the photocatalytic degradation of MET under UV light irradiation. The results are presented in Fig. 10, together with the results obtained using commercially available Degussa P25.

Fig. 10

Kinetics of photocatalytic degradation of MET (c ₀ = 0.05 mM) in the presence of sol–gel synthesized TiO₂ nanopowders and Degussa P25 (1.0 mg mL⁻¹). The influence of different synthesis parameters: a precursor b alcohol and c calcination temperature

The Fig. 10a illustrates the influence of the type of precursor on TiO₂ nanopowders photocatalytic activity in the photodegradation of MET. As can be seen, the catalytic activity of TiO₂ nanopowders synthesized with titanium tetrachloride as the precursor is much higher, but still somewhat lower than of Degussa P25.

The influence of the type of alcogel on photocatalytic activity of TiO₂ nanopowders (Fig. 10b) was examined using different alcohols. As can be seen, the highest photocatalytic activity showed the catalysts obtained using methanol and amylalcohol.

The Fig. 10c shows the influence of calcination temperature on the photocatalytic activity of TiO₂ nanopowders on the MET photodegradation. It can be seen that the TiO₂ nanopowders calcined at ≥550 °C showed higher photocatalytic activity, even slightly higher than Degussa P25.

Since T(CAN)₆₅₀ showed the highest photocatalytic activity in the photodegradation of MET, the effect of TiO₂ loading on the efficiency of MET photodegradation was examined in the loading range from 0.5 to 5.0 mg mL⁻¹ (Fig. 11a). As can be seen, with the increase of the TiO₂ loading up to 2 mg mL⁻¹, the degradation rate increases, and afterwards decreases. Theoretically, the increase in the catalyst loading above an optimum value has no effect on the photodegradation rate since all available light is already utilized. However, higher loading of TiO₂ led to the aggregation of its particles and thus to a decrease in the contact surface between the reactant and photocatalyst particles, which caused a decrease in the number of active sites, resulting in a lower rate of photodegradation^. Also, when TiO₂ was overdosed, the intensity of the incident UV light was attenuated because of the decreased light penetration and increased scattering, which attenuated the positive effect coming from the dosage increment, and therefore the overall performance decreased [48].

Fig. 11

a The effect of T(CAN)₆₅₀ loading on the kinetics of MET (c ₀ = 0.05 mM) degradation; b The influence of TiO₂ loading on MET (c ₀ = 0.05 mM) degradation rate determined for the first 20 min of irradiation

A similar effect was also observed with Degussa P25 (Fig. 11b), but the optimal TiO₂ loading was 1 mg mL⁻¹. It should be pointed out that, T(CAN)₆₅₀ showed higher photocatalytic activity in the whole range of the catalyst loading.

4 Discussion

On the basis of experimental results presented in previous sections, we shall analyze the dependence of TiO₂ nanopowder properties on sol–gel synthesis conditions, and the impact of their structural, morphological and surface properties on photocatalytic degradation of metoprolol under UV irradiation.

Structural analysis of synthesized nanopowders, performed by XRD and Raman spectroscopy, has shown that the sample T(CAN)₆₅₀, with the majority of structural characteristics similar to Degussa P25, has the highest photocatalytic activity. Namely, the mixture of anatase and rutile (with negligible amount of brookite) in this sample is similar to Degussa P25. At the same time, high intensity of Raman signal of T(CAN)₆₅₀, with the position and linewidth of E _g anatase mode (the closest to Degussa P25 among all synthesized powders) points to high crystalline quality of anatase of this sample. However, notice that T(BE)₅₅₀ also has intensive Raman signal, with narrow E _g anatase mode, but its catalytical activity is relatively poor compared to T(CAN)₆₅₀. Despite that, T(CAN)₅₅₀ has lower Raman intensity, with considerably greater blueshift and linewidth of E _g mode, but its photocatalytic activity is close to T(CAN)₆₅₀. From this point of view we may conclude that structural properties of nanopowders have some impact on photocatalytic activity, but not crucial.

On the other side, the results obtained from BET analysis for specific surface area, and especially pore size distribution, lead to unambiguous conclusions. Namely, larger surface area and greater number of pores with dimensions in mesoporus region (>2 nm), produce higher photocalytical degradation of metoprolol, as a material with relatively large molecules (1.7 nm in length). Note that higher mesoporosity is obtained in the samples synthesized in the presence of amyl-alcohol and is getting higher with increase of calcination temperature up to 650 °C.

Special attention must be also paid to analyze the amount of surface O–H groups in TiO₂ nanopowders. Firstly, the lattice contraction, obtained by XRD analysis, points to increased hydration in all synthesized nanopowders. Also, the results for chemical composition obtained from EDS analysis, show higher oxygen content than in stoichiometric TiO₂ for majority of the samples, except the sample T(CAN)₅₅₀, which is obviously nonstoichiometric. However, the EDS results cannot give information about the form of the excess oxygen in TiO₂ (which can be in the form of interstitial oxygen in the lattice, molecular oxygen at the surface or oxygen participating in OH groups). More information on the character and amount of O–H groups in our powders we obtained from IR, and especially from Raman spectra in the C–H and O–H spectral region. These results show that the frequencies and character of surface hydroxyl modes strongly depend on the morphology of the nanoparticles, sol–gel routes and type of TiO₂ crystallites (anatase or rutile) formed during thermal treatment.

According to the presented results, the content of surface O–H groups, mostly related to adsorbed water molecules, is the highest in T(CAA)₅₅₀, followed by T(CAN)₄₀₀. However, in terms of photocatalytic activity, the presence of isolated hydroxyl groups on the surface of TiO₂ is more important than the amount of adsorbed water molecules and surface hydroxyl groups [49]. Namely, it has been recently demonstrated that surface hydroxyl groups, which can trap holes to generate the photocatalytically important active species, i.e. ^•OH radicals, represent decisive factor in high photocatalytic activity of TiO₂ [49]. Reactive hydroxyl radicals ^•OH are the precursors of degradation of any organic and inorganic compound and the first step in the photocatalytic degradation is the formation of the hydroxyl radicals from the TiO₂ surface [50]. Indeed, the Raman results presented in Fig. 9 show that, despite the highest intensity of Raman modes associated to O–H groups related to adsorbed water in samples T(CAA)₅₅₀ and T(CAN)₄₀₀, these samples exhibit worse photocatalytic activity than T(CAN)₅₅₀ and T(CAN)₆₅₀. Note that the presence of high water content in T(CAN)₄₀₀ is surely related to relatively low calcination temperature (400 °C). Moreover, the appearance of mode at about 3,700 cm⁻¹, originating from isolated O–H groups, in the Raman spectra of all samples with relatively high photocatalytic activity (including Degussa P25), and the absence of this mode in T(CAN)₄₀₀ with relatively poor activity, confirm the conclusions proposed above [49, 50].

5 Conclusion

We have demonstrated that highly photoactive nanocrystalline TiO₂ powders can be produced by appropriate choice of sol–gel synthesis parameters. Different mixtures of TiO₂ crystal modifications (anatase, rutile, and brookite) are obtained with proper choice of synthesis parameters: the type of precursor, alcohol and basic solution, as well as the calcination temperature. It is noticed that the content of brookite decreases photocatalytic activity in synthesized nanopowders with other properties comparable. On the other side, the amount of rutile does not have critical impact on the photocatalytic activity. The photocatalytic performances have been attributed to the anatase particle size and crystalline quality, rather than to the presence of rutile phase. Namely, the photocatalytic activity of TiO₂ nanopowders mostly depends on anatase structure (crystallinity, particle size, microstrain, and nonstoichiometry) and porosity, as well as the amount of surface O–H groups, which can produce hydroxyl radicals. Additionally, pore size distribution in TiO₂ nanopowders has crucial role in photocatalytic degradation of metoprolol, as a material with relatively large molecules. According to this, the highest photocatalytic degradation of MET is obtained for the sample synthesized by sol–gel routine with TiCl₄ as a precursor, in the presence of amyl-alcohol and NaOH as basic solution, and calcined at 650 °C.