Titania nanotubes (TNTs) prepared through the complex compound of gallic acid with titanium; examining photocatalytic degradation of the obtained TNTs

3.1.1 ¹H NMR Analysis

¹H NMR spectra of gallic acid as a ligand and its complexes with titanium (IV) is presented in Fig. 1.

Close

Fig. 1

¹H NMR spectra of gallic acid as (a) the ligand (L) and (b) its Ti(IV) complex.

Export to PPT

The intense peak at 3.41 ppm (s, 2H) in the ¹H spectrum of the ligand belongs to two equivalent protons of the benzene ring. At coordination with the titanium ion, the protons become non-equivalent and the signal splits into two.

The vocative protons of the hydroxyl-groups involved in H-bonding of different types (both intra- and intermolecular interactions) are detected as a peak with the maximum at 8.83 ppm. It also can be seen in the spectrum of the complex compounds that not all the OH groups are ionized at complexation.

The peaks at 4.28 and 1.33 ppm in the spectrum of the complex are related to H₂O and impurities of n-butanol.

3.1.2 Quantum chemical modeling

This study is devoted to quantum-chemical modeling of the structure and some properties of gallic acid (H₄L), as well as its metal complex (Fig. 2). H₄L molecule has four mobile hydrogen atoms, however, according to chemical analysis, this molecule is involved in complexation as a dianionic ligand. Therefore, in the following, we will denote this molecule as H₂L.

Close

Fig. 2

Molecular structure of the isomers of the H₂L molecule, dianion L²⁻ and metal complexes according to the DFT calculation.

Export to PPT

The number of intermolecular H bonds determines the stability of the H₂L molecule. In this case, the H (O1) atom can form a hydrogen bond with both the O2 atom of the carbonyl group and the O3 atom of the hydroxyl group. The isomer formed in the latter case is 17 kJ/mol less stable than that shown in Fig. 2 as isomer H₂L, in which the length of the H-bond O2 ∙∙∙ H (O1) is 1.703 Å.

For the H₂L² molecule (Fig. 2), three (L²⁻) dianions are possible, in which the deprotonated oxygen atoms are able to form a chelate bond with the metal cation. These dianions are formed by ionization of the H (O1) and H (O3) atoms, the H (O1) and H (O5) atoms, the H (O4) and H (O5) atoms. The first of them is 43 and 64 kJ/mol more stable than the second and third, and its structure is given in Fig. 2. The dianion is stabilized by the O1∙∙∙ H(5) intramolecular H bond with the length as 1.610 Å. The coordination polyhedron of the Ti(L)₂ complex is a distorted tetrahedron: the O1TiO1A and O3TiO3A angles are 115 and 120°, and the O1TiO3 and O1TiO3A angles are 94 and 117° (Fig. 2). Table 1 includes the calculated bond lengths in the H₂L molecule, its dianion (L²⁻), and the metal complex Ti(L)₂ shown in Fig. 2. In Table 2, some atomic charges (ē) calculated by the NBO method in the H₂L molecule, its dianion (L²⁻), and in metal complexes are presented.

From Table 2, it is evident that the charges on the Ti-atom in the complex equals + 1.760 instead + 4. This indicates the essentially covalent nature of the metal – ligand bonds. The data presented in Table 3 indicate the values ​​of electron density transfer to metal cations Δq (Mⁿ⁺), as well as the values ​​of electron density transfer from the dianion of the ligand Δq (L²⁻), and water molecules Δq (H₂O). Also the electronic configurations of central atoms in the molecules of the complexes is illustrated.

It is obvious from the Table 3 that some part of the electron density of gallic acid dianion transfers to the metal cation. The charge on the O1 atom changed from the value of −0.776 in the dianion to the values ​​of −0.822, −0.963 in the doubly charged complexes. Similarly, the charge on the O3 atom changed from a value of −0.787 in the dianion to values ​​of −0.817, −0.974 in doubly charged complexes.

The intense long-wavelength bands (DP) calculated by the DFT method in the ESP in the H₂L molecule, (L²⁻) dianion, and Ti (L)₂ titanium (Table 4) .

When a molecule going from neutral to a dianion, a small bathochromic shift of the DP in the ESP is observed. In Table 5 shows the results of the calculation of the IR spectra of the H₂L molecule and in Ti (L) titanium using the DFT method.

The bands in the region of 3100–3500 cm^{− 1} are related to O— H group vibrations, and the position of the bands is determined by the degree of participation of the hydrogen atom of this group in the intermolecular H bond.

The ν (C⚌O) absorption band shifts to the low-frequency region when going from the H₂L compound (1768 cm⁻¹) to the titanium complex (1797 and 1775 cm⁻¹) there is a shift to the high-frequency region.

3.1.3 IR and elemental analysis

Fig. 3 illustrates the experimental FT IR spectra of gallic acid as ligand (L), and complex compound of titanium. Based on the result obtained from the FTIR spectroscopy, the C⚌O stretching, which shows itself as a strong absorbance, appeared at 1712 cm⁻¹ in the ligand. This position (C⚌O absorption band) illustrates that an intermolecular H bond went on between C⚌O and the adjacent OH group. Overlapping of ν(C⚌O) vibrations with the C⚌C aromatic stretching occurred at lower frequency shoulder at 1616 cm⁻¹. Another signal appeared at 3511 cm⁻¹, which related to ν(O-H) band. In the complex compound mode, the ν(C⚌O) band shifted to 1670 cm⁻¹. There are also two other signals at 1515 and 1628 cm⁻¹, which correlated to anti-symmetrical and symmetrical stretching of the carboxylate anion COO⁻ respectively (Fig. 3). According to the Δ value (Δ = ν^as - ν^s) (113 cm⁻¹), monodentate coordination happened from the carboxylic group of the ligand. Elemental analysis with accompany of IR illustrated that titanium ion was coordinated by the carbonyl group and its adjacent hydroxyl group through their O-atoms (Table 6).

Close

Fig. 3

FT IR of the ligand, complex, and nanoparticle compounds.

Export to PPT

In the case of titanium oxide nanoparticle (TiO₂), after raising temperature, all signals corresponding to the organic substances and H₂O (between 800 and 3500 cm⁻¹) disappeared. Several interactions between water’s molecule and the surface of the titanium oxide nanoparticle (H-bond) occurred, and it is observable by splitting of the band into various components. There are also a few weak signals around 1640 cm⁻¹ (δ(OH)), which is due to the hydration on the surface of TiO₂. In addition, two sharp and broad signals appeared at around 675 and 790 cm⁻¹ respectively (Ti—O stretching band), which related to TiO₆ octahedrons in the lattice with several different distortion.

3.1.4 Thermal decomposition

Fig. 4 provides information about the optimum temperature to synthesize titanium oxide nanoparticles by the thermal analyzing method. Mass loss started at low temperature (35 °C). Then, decreasing mass continued in four different steps. Step A related to the removal of water, which terminated around 80 °C. Step B was a transient state from the complete removal of water to start removing organic substances, which also involved three next steps of C, and D to loss 80.55% of total mass by around 550 °C, which after that the mass fixed with no remarkable changes. The most mass reduction occurred between 445 and 550 °C, which belonged to the loss of ligands, surrounded the titanium ion, and according to the final mass (80.55 g), synthesizing nanoparticle of TiO₂ was proved regarding the initial mass of the complex (Ti(L)₂, 384 g). Furthermore, the DSC curve (red line) provided information about the thermodynamic effects. The initial mass loss started with endo effects at temperatures 70 and 221 °C, and then the first exo-effects started at higher temperatures by starting removing gallic acid as ligand compound at around 330 °C, and then it continued by newt exo-effect at 445 °C, which was the point that the organic substance was decomposed, and continued to obtain an oxide of titanium removed at 557 °C.

Close

Fig. 4

Thermal decomposition of TiL_2.

Export to PPT

3.1.5 XRD Pattern

Fig. 5 provides information about the XRD of ligand, complex and nanoparticle. Some new peaks appeared after formation of titanium complex compound, and proved some other phases were created. There are two new peaks in the areas 25.26 (1 0 1), 27.42 (1 0 0), which related to anatase and rutile phases respectively and indicated that “Ti” ion was coordinated by the ligand.

Close

Fig. 5

XRD pattern of ligand (L), Ti(L)₂, and TiO₂ (anatase, rutile phases).

Export to PPT

In the case of titanium nanoparticle, two different phases of anatase and rutile phases appeared at the temperature of 500 and 600 °C. Based on the XRD pattern, the JCPDS card Number are equal 21–1272 , 21–1276 for anatase and rutile respectively.

Scherer formula was used for measuring the size of the nanoparticles: $$D=\frac{k\lambda }{\beta cos\theta }$$

According to the results, the average sizes of anatase and rutile phases obtained around 21 and 39 nm respectively.

3.1.6 Microscopic Image

The morphology of the obtained nanoparticle was studied by SEM image accompanied with EDX analysis (Fig. 6).

Close

Fig. 6

Microscopic image of Rutile phase; (a) SEM images, (b) EDX pattern, and Anatase; (c) SEM images, (d) EDX pattern.

Export to PPT

According to the result obtained from the rutile nanoparticle (Fig. 6a), it showed that the shapes had not regular structure which was also predicted because titanium oxide in the phase of rutile are obtained usually rock-like, however, the surface of the nanoparticles showed nanocrystalline domains

According to the result, the spherical shape of the anatase phase after calcination complex compound of titanium was formed. The size measured by the microscopic image is matched with XRD analysis. Uniformity of the nanoparticles is quite satisfied, which related to the proper decomposing of the complex compounds. EDX pattern of the nanoparticle also was studied (Fig. 6b). Based on the analysis, the nanoparticle was pure with no impurity (Table 7).

3.2.1 XRD pattern

XRD pattern of the nanoparticles in different steps was recorded. According to Fig. 7a, the XRD pattern of rutile in all steps had some changes, although, all of them are still rutile. It demonstrated that the procedures, which were applied in each step did not effect on chemical composition of titanium oxide nanoparticle and the structure of the corresponding nanoparticle had no remarkable changes. The sizes of the nanoparticles obtained by Scherer equation are almost between 41 and 42 nm (sample I), 40–41 nm (sample II), and 39–40 nm(sample III).

Close

Fig. 7

XRD pattern of samples; (a) Rutile phase; (I) after synthesizing to rutile, (II) after treating by HCl, (III) after annealing at 500C. (b) Anatase phase; (I) after synthesizing to anatase, (II) after treating by HCl, (III) after annealing at 500C.

Export to PPT

On the other hand, Fig. 7b illustrates XRD pattern of the different samples obtained from anatase phase. Fig. 7b illustrates that the XRD pattern related to the titanium oxide, which treated by NaOH. According to the result, the samples is matched with Na₂Ti₃O₇ with JCPDS 00–059-0666, the main peaks can be observed at around 11, 25.5, and 37, which the second signal related to anatase phase (25.5). The reaction between sodium hydroxide and titanium oxide can be written as the following Eq. (1):

By water treatment, some Ti-O-Ti bonds were broken to form [Ti(OH)₆]₂. However, this structure is not so stable and due to combination with oxides, nuclei were formed and became thermodynamically stable because of increasing the size of the crystal. Thin nanosheets was formed and combined while the process was grown. The obtained Na₂Ti₃O₇ nanoparticle had layer with octahedral sites and cations of Na⁺ were located between the octahedral layers.

Fig. 7 b (II) is related to the sample which was treated by HCl. Na⁺ in Na₂Ti₃O₇ was replaced by H₃O⁺ when Na₂Ti₃O₇ reacted with HCl:

The XRD pattern of the sample showed signal at 25.3°, which illustrated anatase phase ‘s signal is still exist and the JCPDS was matched with 01–089-4921. However, because the crystalinity of the sample is low the profile fitting could not be determined very well. In common researches, exchanging Na⁺ ions are the reason for collapsing titania nanotubes, but in this experiment, preparing nanosheets was the main reason for being this type of crystalinity (low crystalinity).

Fig. 7b (III) belongs to the titania nanotubes after annealing at 500 °C. The XRD pattern determined that titanium oxide nanotube in anatase phase was prepared by being signals at 25.36, 37.9, 48.17, and 55°.

Comparing to rutile phase, anatase showed that the structure was flexible and it is possible to obtain titania nanotubes from anatase phase, while rutile phase because of its very strong and stable structure had no remarkable changes. The size of the samples obtained from the same method, and obtained 23, 10, and 18 nm.

3.2.2 Microscopic image

Fig. 8 provides information about the SEM image of the obtained nanoparticles in different steps. According to the result obtained from the rutile nanoparticle, it showed that the morphology of the nanoparticle has not been changed a lot and there were no special changes in the shape of the nanoparticles, even after all the processes. The microscopic images illustrated that the shapes had no regular structure which was predicted, because titanium oxide in the phase of rutile are obtained usually rock-like, however, the surface of the nanoparticles showed nanocrystalline domains. Rutile phases due to having high stability, which obtained as the last phase of titanium oxide at high temperature are usually so refractory and resist different conditions, in addition, its irregular shape did not allow obtaining a regular shape such as tube.

Close

Fig. 8

SEM image of rutile phase; (a) sample I, (b) sample II, (c) sample III, and anatase phase; (d) sample I, (e) sample II, (f) sample III, (g) TEM image of the nanotube obtained from the anatase phase.

Export to PPT

In the case Anatase phase, it’s obvious that how the spherical shape (Fig. 6c) changed to the tube shape with the 5 nm and 8 nm of inner and outer diameter respectively (Fig. 8f). According to the result, we assumed the following procedure. When the sample was treated by alkali, Na-O-Ti bonds appeared on the surface of the nanoparticles (TiO₂). However, there are some microscopic particles, which Na-O-Ti bonds did not appear on their surfaces, so they made charge imbalance (Fig. 8d) (Kasuga et al., 1999) . After treating by HCl, the charged, which appeared on the surface of the nanoparticles and subsequently the electrostatic repulsion vanished. In addition, next treating by distilled water let the charged slowly disappeared and changed Ti-O-Na bonds to Ti-OH (Fig. 8e). The mentioned bonds (Ti-OH) made the structure to be sheet like, which is existed in the tube structure. Also, because Ti-O-Ti or Ti-O---H-O-Ti hydrogen bonds were produced by the dehydration of Ti-OH bonds due to treating by chloric acid, the interval bond between two atoms of titanium (Ti) reduced, which made two sides of the sheet attracted to each other. While this procedure was happening, remaining Ti-O-Na bonds and their electrostatic repulsion made a connection between the extremities of the sheet, so the tube structure formed (Fig. 8f). The most differences between this method and the anodization methods, which has been already carried out, is that, in the current method there is no need to use Teflon-lined autoclave and it can be converted to tube just by a simple chemical reaction (which was mentioned). In addition, the highest temperature required for this reaction, before calcination, is 110C, while in other reported methods, temperature was around 160-200C (Qiu et al., 2018; Tan et al., 2019; Zhao et al., 2019; Liu et al., 2019; Wang et al., 2019).

TEM image of the last product (Fig. 8g) showed the shape of product clearly, and nanotube shape is observable. The inner and outer diameter obtained by TEM image was found to be around 3–4 nm and 8–10 nm respectively, which is matched with the data resulted from SEM image.

3.2.3 Raman Spectroscopy

Fig. 9 Provides information about the Raman spectroscopy of rutile and anatase phases. According to the result, a remarkable change did not occur when rutile phase was used as precursor. There are two peaks at around 602 and 450 cm⁻¹, which related to Ti-O-Ti. After treating by NaOH and HCl, which are observable as I and II sample respectively, still the peaks remain the same as rutile phase. After annealing (sample III) the peaks had no any changes and it showed that the bonds between titanium and oxygen in rutile phase were not broken and nothing happened.

Close

Fig. 9

Raman spectroscopy of samples; (a) Rutile phase, (b) Anatase phase.

Export to PPT

On the other hand, the figure related to anatase phase (Fig. 9 b) illustrated some changes in the Raman spectrum. In anatase phase (before any treatment), there are two peaks at around 619 and 445 cm⁻¹, which related to Ti-O-Ti bonds. After treating by hydroxide sodium (sample I), the mentioned peaks almost vanished, while two new peaks at 290 cm⁻¹ and 908 cm⁻¹appeared, which related to Na-O-Ti and four-coordinate Ti-O, respectively (Kasuga et al., 1999). It is matched with the result obtained from the XRD and it confirms that the Ti-O-Ti was broken to form Ti-O-Na and Ti-OH. The spectrum regarding sample II, the peak at 908 cm⁻¹ disappeared, while a small peak at 158 cm⁻¹ appeared. It is believed that the titania nanotubes was formed in this stage. In the spectrum, which related to sample III, a sharp peak at 155 cm⁻¹, and a new broad small peak at 623 cm⁻¹ appeared. All the mentioned peaks are due to anatase phase (Kasuga et al., 1999).

3.2.4 Photoluminescence analysis

The photoluminescence of the samples was studied and it illustrates in Fig. 10. Rutile phase illustrated the following characterization; photoluminescence of the samples was measured at 370 nm excitation wavelength with an optical filter placed before detector (370 nm cut-off wavelength). Fig. 10.a shows a number of weak surface-state emissions at 400–480 nm visible range with maximums at 417,438,470 nm clearly expressed for samples rutile phase I and less for II and III with ± 2 nm offset. Green emission peak at 480–570 nm ranges with the maximum at 514 nm is clear and twice more intensive for rutile phase I samples (40–33 r.u. vs. 19–16 r.u. for II-III samples). Separated photoluminescence bands could be explained by different nature of electronic transitions – free holes (Rex et al., 2014, 2016) or trapped holes (Nakamura and Nakato, 2004; Nakamura et al., 2005) or can be a consequence of surface oxidation which also explains weak green emission for samples II and III.

Close

Fig. 10

Photoluminescence spectroscopy of samples; (a) Rutile phase, (b) Anatase phase.

Export to PPT

On the other hand, Fig. 10 b provides information about anatase sample and its samples derivative from the different steps of the process. Although the shape of the samples is different, the spectra for all the samples were obtained almost the same as each other, however, the intensities are different. The intensity related to the III sample (pink line) showed the greatest intensity, while in contrast the sample, related to treating by acid (II, black line) had the lowest intensity. It can be realized that, the sample III due to having the shape of tube and greater BET surface activity than previous samples, which were spherical (anatase phase) or irregular shape, could trap the electrons more. The sample treated by hydroxide (sample I) and acid (sample II) in opposite, had the lowest spectra. This is because of two reasons of the shape and some new bonds. The irregular shape of the samples, which obtained in these steps did not illustrate large BET surface activity, in addition, forming some new bonds (Ti-OH and Ti-O-Na) act as coordinator led the sample destabilize (0 0 1) the surface by their coordination, while in anatase phase and in the last sample (III) due to annealing there were just TiO₂ in spherical and tube shape respectively and trapping electron related to red photoluminescence were because of under-coordinated Ti by proper element.

3.2.5 UV–Vis absorption investigation

Fig. 11 provides information about the UV–Vis diffuse reflectance spectra of rutile and anatase. Tauc’s plot method was used to obtain the energy band gap of the substance: $$\alpha \text{hv}=\text{A}{{(\text{hv}-{\text{E}}_{\text{g}})}^n}^{=0.5-2}$$

Close

Fig. 11

Absorbance of Rutile phase sample (a);(I) before modifying processes, (IV) after modifying processes, and anatase (b); (I) before modifying processes, (IV) after modifying processes.

Export to PPT

In the case of rutile (Fig. 11a), samples showed a very weak absorbance at 340 nm (3.65 eV) which is close to normal absorbance of rutile but a huge absorbance at 252 ± 2 nm for all measured samples. According to Tauc’s plot, direct transitions zone was calculated as 4.46 ± 0.02 eV and indirect was about 3.91 eV. Such huge band gaps were discussed by authors (PiotrKowalski et al., 2009; Li et al., 2015; Mercado et al., 2012) with the use of DFT method of analysis to explain the oxygen positions on the TiO₂ (1 0 1), (1 0 0) and (0 0 1) surfaces. As it’s clear from the figure, both first and last samples were almost the same with no particular changes.

On the other hand, two 1st and last samples showed a huge gap (almost 20 nm), which belonged to the anatase phase and nanotube respectively (Fig. 11b). It illustrated that when the nanoparticle converted to tube shape, due to having larger active surface, it could absorb UV light at higher wavelength. And because the substance was the same with no extra modification, it can be realized that the only parameter effected on the properties of the nanoparticle was the shape of it, which turned to tube from the spherical shape. Based on the result, obtained from Tauc’ plot, the energy band gap of anatase was obtained 4.08 eV, while the energy bandgap of the nanotube decreased by 3.8 eV.

3.2.6 BET surface analysis

Fig. 12 provides information about BET surface analysis of

Close

Fig. 12

Adsorption-desorption isotherm N2 (77 K) of nanoparticles.

Export to PPT

“A”, “AM”, “R”, and “RM” samples to investigate the specific surface activity, total pore volume, and average pore diameter was studied by BJH method. The pressure range (P/P₀) between 0.4 and 0.95 was applied to obtain the surface area and texture parameters (Absalan et al., 2018).

Table 8 illustrates the mesoporosity of the samples. The result obtained from the graph proved that the hysteresis loops were kind of type 4 (Yahya Absalan et al., 2017), which provides micropores related to mesoporous. In addition the size of the nanoparticles were between 10 and 40 nm (based on the result of the XRD), and it illustrates presence of mesoporous. While process of calcination was carrying out, the obtained coarser mesoporous degradation and crystal development let forming micropores. In addition, aggregates of plate like nanoparticle, especially in “A”, appeared to thin slit-shaped pores, which make forming type 4 of hysteresis loops.

The specific surface of the mesoporous was obtained from the result of the adsorption-desorption isotherm. Based on the result, it suggested that the nanoparticles with smaller size had larger specific surface activity (Table 8); the advantage of the larger surface activity is based on this fact that they can react better in photocatalytic degradation’s reaction. In addition, the results associated with the designation of the larger mesoporous diameter, which is the result of the pore volume’s structural curves (Table 8).

Photocatalytic investigation. The ability in photocatalytic degradation of bromophenol blue by the obtained titania nanotubes is illustrated in Fig. 13. bromophenol blue are among the organic harmful substances, which is so difficult to be degraded. However, all the synthesized nanoparticles could remove bromophenol blue from the environment with the different ratio under UV light; “R”, “RM”, “A”, and “AM”. According to the result shown in Fig. 13, photocatalytic activity of “AM” was more than others, while sample “A” had the least ability to remove bromophenol blue after “R” and “RM” which were able to remove almost the same. Based on the result extracted from BET surface activity, the surface active of “R”, “RM”, “A”, and “AM” were 20.36, 19.33, 16.12, and 43.52 cm². Thus, it was predictable that “AM” could remove bromophenol blue under weak UV light more than others. A better irradiation by the amorphous “R” and “RM” than “A” (spherical shape) could be explained that the nanomicroporous, which existed in “R” and “RM” higher than “A”, were able to prevent the structure of “R” and “RM”, which are immobilized in the matrix of TiO₂ (Zeng et al., 2010). In addition, the cyclic photodegradation curves of bromophenol blue by “AM” are provided by Fig. 14. “AM” was used to degrade bromophenol blue for three times to test its photo-stability, and it showed that its activity was not reduced remarkably even after three times. Indeed, it suggests that the immobilization of “AM” nanoparticles in its structure made it resist to photo-corrosion (Long et al., 2006). The time of photodegradation was calculated by the intensity of UV light and the first concentration of bromophenol blue. The UV light, which was set for the experiment, was 0.55 mW/cm² (255 nm) as a very weak UV light. The result showed that it could reduce the concentration of bromophenol blue very well even under weak UV light, which was held at 25 °C. The result showed a high ability for the obtained titania nanotubes; according to our best knowledge and under the same situation, other reported titania nanotubes could decrease concentration of the bromophenol blue by almost 75% (Liu et al., 2008)

Close

Fig. 13

Photodegradation curves of bromophenol blue with “AM”,”R”, “RM”, and “A” (a), and the bromophenol blue degradation concentration (mg/L) under UV irradiation 140 h (b).

Export to PPT

Close

Fig. 14

Cyclic photodegradation curves of bromophenol blue with the “AM”.

Export to PPT

The mechanism of degradation can be considered as follow: