A simple preparation route for polysilicate titanium salt from spent titanium solutions

It is necessary to remove organic matter from water, which could eliminate potential carcinogenic, disinfection-byproducts (DBPs) and trihalomethanes (THMs). The main part of the organic matter in natural freshwater is humic acid (HA, about 50%–90%), which might be converted to DBPs and THMs (Hu et al. 2006; Kazpard et al. 2006). HA is composed of polycyclic aromatic macromolecules and a plurality of oxygen-containing functional groups. The concentration of HA in drinking water ranges from 3 to 5 mg/L regarding chemical oxygen demand (COD) according to standards for drinking water quality (Zhao et al. 2011b). Coagulation and flocculation play important roles in water treatment, which could promote the sedimentation of suspended solids from small particles to large flocs and then reduce HA as a result (Lyon et al. 2014; Xia et al. 2018).

The flocculants are mainly composed of inorganic and organic chemical electrolytes. Inorganic flocculation is widely used in water treatment due to its low cost and good effect. The inorganic flocculation used in water treatment contains iron and aluminium salts, which can be hydrolyzed in water to form a positively charged colloid such as Al(OH)₃ or Fe(OH)₃ (Duan & Gregory 2003; Wei et al. 2018; Yang et al. 2019). However, hydrolyzation is too fast to control the species of floc. To resolve this problem, polysilicic acid (PSiA) chained by Si-O-Si, such as Si₂O₃(OH)₄²⁻ and Si₃O₅(OH)₅³⁻, is added to the iron and aluminium salt solution to prevent iron ions and aluminium ions from precipitating, and at the same time successfully increasing floc weight and size (Zouboulis & Moussas 2008; Qiu et al. 2011). Nevertheless, aluminum is related to some neuropathological illness such as Alzheimer's disease (Lévesque et al. 2000; Camacho et al. 2017; Santore et al. 2017; Katrivesis et al. 2019). Compared with aluminium-based flocculants, the flocculation performance of iron-based flocculants is usually environmentally friendly. Unfortunately, the iron-based flocculation treated water has a higher chromaticity, which will limit its wide application (Zhao et al. 2013).

To overcome these difficulties, many researchers have paid attention to titanium-based flocculants. Titanium(IV) salts show good flocculation performance due to their high valence (Wu et al. 2011; Wang et al. 2019). Upton & Buswell (1937) first found that Ti(SO₄)₂ could be used as a flocculant in water treatment in 1937. Compared with Al₂(SO₄)₃, Ti(SO₄)₂ has better performance in removing fluoride and chromaticity color in water treatment. Shon et al. (2007) first reported that the coagulated sludge of titanium-based flocculants could produce TiO₂, which is applied in paints, photocatalysts, electronic paper, cosmetics, and solar cells (Obee & Brown 1995). However, there are some certain drawbacks for titanium salts during water treatment, such as high hydrolysis rate and difficulty in storage and transportation, which are similar to iron and aluminium salts. Also, the pH value of the effluent is low after titanium salt treatment (Huang et al. 2016b). Hence, researchers have set out to prepare polytitanium salt flocculants to improve the stability of titanium salts. Zhao et al. (2011a, 2011b) prepared polyaluminum chloride (PACl) to treat surface water with an effective flocculant. Furthermore, PSiA includes a large number of electronegative silica species, which could improve the aggregation ability of titanium salt flocculants and also restrain the hydrolysis of Ti⁴⁺ to stabilize titanium salts (Dietzel 2000; Zouboulis & Moussas 2008). Huang et al. (2014a, 2014b) have studied the flocculation performance and floc properties of polytitanium-silicate-sulfate and polysilicate titanium salt (PST) with different Si/Ti molar ratios during water treatment. However, a large amount of titanium salt is used in the above flocculants. It is well known that titanium salt is more expensive, and the increasing cost of titanium-based flocculants will limit their application in water treatment. Moreover, presently there is no study on the preparation of titanium-based flocculants with spent titanium solutions.

For the first time, we prepare PST with spent titanium solutions directly. The flocculation performance of the PST with different Ti/Si molar ratios is studied in sodium humate–kaolin synthetic water and domestic wastewater treatment. The physical and chemical properties of PST with different Ti/Si molar ratios are characterized by scanning electron microscopy (SEM), Fourier transform infrared spectroscopy (FT-IR) and X-ray photoelectron spectroscopy (XPS). The flocculation performance is investigated through turbidity removal efficiency and COD removal efficiency. Floc properties are evaluated by floc size, flocculate speed, sedimentation ability, rapid sweep netting ability and stability.

The spent titanium solutions used in this work are provided by Yunnan Titanium Industry Co., Ltd, of China, and come from the cold-rolled titanium plate pickling process, and the composition is mainly metal ions such as tetravalent titanium ions and strong acid solution. The composition of the spent titanium solutions is shown in Table 1. Na₂SiO₃·9H₂O, H₂SO₄ and other chemicals in this work were all purchased from Shanghai Aladdin Biochemical Technology Co., Ltd, of China. The solution used in the experiment was prepared by deionized water. The PST was prepared at room temperature.

Preparation of PSiA: First Na₂SiO₃·9H₂O was dissolved in deionized water. Then the Na₂SiO₃ solution was slowly dripped into 0.15 M H₂SO₄ solution under magnetic stirring and aged for 4 h. Finally, the PSiA solution (0.12 M as SiO₂) was acquired.

Preparation of PST: Three kinds of PST with different Ti/Si molar ratios of 2, 1 and 0.5 (named as PST2, PST1, and PST0.5, respectively) were prepared by mixing PSiA with different volumes of spent titanium solutions. The experimental process is shown in Figure 1.

Sodium humate–kaolin synthetic water was synthesized by sodium humate (purchased from Sinopharm Group Co., Ltd, Shanghai) and kaolin; 50 mg sodium humate and 5 g kaolin were dissolved into 5 L deionized water under magnetic stirring for 1 h and allowed to stand for 30 min. The sodium humate–kaolin synthetic water contained 10 mg/L sodium humate with initial turbidity of 143 ± 0.5 NTU (nephelometric turbidity unit) and was adjusted by kaolin. The d50 values of kaolin used in this experiment were 1.63–3.89 μm. The COD and pH value of the sodium humate–kaolin synthetic water were 36.4 ± 0.1 mg/L and pH = 7.89 ± 0.02, respectively.

Rural domestic wastewater was collected from the Dianchi valley. The properties of the domestic wastewater were as follows: turbidity = 126 ± 1 NTU, COD = 61.1 ± 1 mg/L, pH = 7.13 ± 0.02.

The flocculation experiment was used to evaluate the flocculation performance of PST. Initially, 250 mL of sodium humate–kaolin synthetic water and domestic wastewater was transferred to a beaker, respectively. Then, the different Ti/Si molar ratio PST was added in each beaker and the pH value was controlled from 3 to 12 with stirring (first, stirred rapidly at 200 rpm for 3 min; second, stirred slowly at 40 rpm for 10 min; finally settled for 30 min). Finally, collection was from 2 cm below the solution surface for subsequent measurements. Turbidity was directly measured without any handling using a 2100P turbidimeter (Hach, USA). COD was measured by potassium permanganate titration.

The surface morphology of freeze-dried PST was observed by scanning electron microscope (TESCAN VEGA3, CZE) at an accelerating voltage of 20 kV. The PSiA and PST with different Ti/Si molar ratios were characterized by a KBr pressed disc with Fourier transform infrared spectroscopy (Thermo Scientific Nicolet iS10, USA). The chemical valence states of Ti and Si on the surface of the PST were investigated by using X-ray photoelectron spectroscopy (Thermo Fisher Scientific, K-Alpha⁺) with a monochromatic Al X-ray source (1,486.6 eV). The particle size distribution of kaolin was tested by Mastersizer 3000 (Malvern, UK), as shown in Figure 2 and Table 2. The variation of floc size in situ during the treated sodium humate–kaolin synthetic water process was measured by Mastersizer 3000. We tested the particle size every 16 seconds and tested a total of 200 times with 500 rpm stirring speed. In this study, the floc size is represented by the median volumetric diameter (d50).

The morphology and surface structure of PST2, PST1, and PST0.5 were investigated by SEM as shown in Figure 3. It is clear that the Ti/Si molar ratio directly affects the overall morphology of the PST. When the Ti/Si molar ratio is 2:1, the prepared PST2 displays a rod-like structure with a width of 0.21–4.08 μm. When the Ti/Si molar ratio is 1:1, the prepared PST1 reveals a spherical shape with agglomeration and the particle diameter is about 0.61–3.31 μm. Once the Ti/Si molar ratio drops to 1:2, the morphology of PST0.5 transforms into spherical secondary particles with a diameter of 2.66–8.35 μm. In summary, the Ti/Si molar ratio determines the morphology and surface structure of PST. The PST tends to be a one-dimensional linear or two-dimensional sheet shape when the content of titanium increases. Once the content of titanium reduces, the PST will transform into a three-dimensional spherical shape with larger secondary spherical particles.

Figure 4 shows the FT-IR spectrum of PSiA, PST2, PST1, and PST0.5. The strong and broad peaks around 3,448 cm⁻¹ and 1,635 cm⁻¹ in the four samples can be attributed to the stretching vibration peak of -OH and the vibration of H₂O molecules on the material surface (Huang et al. 2014a; Chen et al. 2015a; Sun et al. 2017; Ajao et al. 2018). The vibrational bands at 1,136 cm⁻¹ with strong absorptions can be ascribed to the sulfate. The absorption peaks at 1,059, 884 and 848 cm⁻¹ could be caused by the asymmetric stretching vibration of Si-O-Si (Moussas & Zouboulis 2008; Xu et al. 2009; Sun et al. 2017). The peak at 997 cm⁻¹ corresponds to the bending vibration of Ti-O-Si, while the observed absorption band at 771 and 735 cm⁻¹ could be assigned to the Ti-O-Ti stretching mode (Sun et al. 2010; Li et al. 2013; Chen et al. 2015b; Huang et al. 2016a). The absorption peaks at 601 and 462 cm⁻¹ could owe to the Si-O out-of-plane bending mode and symmetry stretching mode, respectively, which stems from hydrolysis of PSiA in general (Chindaprasirt et al. 2009; Sun et al. 2017; Yan et al. 2017).

To analyze the bonding structure of the prepared flocculants, we detected the variation of the absorption peaks of the Si-O-Si and Ti-O-Si bonds. The results show that the intensities of both Si-O-Si and Ti-O-Si bonds experience a process of rising and then falling with the titanium content increasing. So, we believe that the polymerization of samples is strengthened then weakened. In addition, we could also confirm the variation of Ti-O-Ti and Si-O absorption peaks from the FT-IR results, which indicates that the inhibition of hydrolysis is strengthened and then weakened as well. The similar pattern may be attributed to the titanium-containing waste liquid promoting the formation of Si-O-Ti bonds and inhibiting the hydrolysis of PSiA. However, when spent titanium solution increases, the above effects are weakened, and then Ti⁴⁺ starts to hydrolyze (Dietzel 2000; Duan & Gregory 2003; Zouboulis & Moussas 2008; Qiu et al. 2011; Li et al. 2013; Chen et al. 2015a). It is well known that the flocculant will display an excellent bridging absorption capability and sweep netting ability with a larger floc size and a higher polymerization degree (François 1987; Li et al. 2013; Huang et al. 2016a, 2016b; Shahadat et al. 2017). Weighing the polymerization and hydrolysis, we can reach a conclusion that PST1 has the optimal polymerization and moderate hydrolysis while the polymerization of Ti and Si for PST0.5 is not sufficient.

The C, O, Ti, and Si XPS energy spectra of the prepared PST samples are shown in Figure 5. Since there is no C in the PST sample, contaminated C (CH_x) is added as a spectral alignment on the binding energy (BE) scale (284.8 eV). The C components at ∼285.8 eV and ∼289.9 eV of Figure 5(a) could be assigned to epoxy and carboxyl bonds, respectively (Chinh et al. 2018). The binding energy for the O1s peak of different Ti/Si molar ratio PST samples is mainly 532.1 eV and 533.1 eV as shown in Figure 5(b), which matches well with the state of O in TiO₂ and SiO₂ or water/OH⁻ (Mohai et al. 1990; Green et al. 1997). As plotted in Figure 5(c), the binding energy for the Ti2p_3/2 and Ti2p_1/2 peak of PST with different Ti/Si molar ratio samples is almost 459.9 eV and 465.6 eV, which could be contributing to the oxidation state of Ti(IV) in TiO₂, suggesting that titanium in PST is almost in a Ti(IV) state (van Hassel & Burggraaf 1991; Green et al. 1997; Kurtz & Henrich 1998; Hashimoto & Tanaka 2002; Oktay et al. 2015; Chinh et al. 2018). However, it is a problem that the binding energy of Ti2p is higher than in the published literature, which indicates that there is a higher valence state of Ti (TiO_x) in PST (Naitabdi et al. 2016; Guo et al. 2017; Köymen et al. 2019). In addition, the binding energy intensity of Ti2p XPS increases with Ti/Si molar ratio increasing, which could also indicate the Ti content rising in the PST. As shown in Figure 5(d), the binding energy for the Si2p peak of PST2, PST1 and PST0.5 is 104.1 eV, 104.1 eV and 103.8 eV, respectively, which matches well with the oxidation state of Si(IV) in SiO₂ (Oswald 2010; Philippe et al. 2012; Nie et al. 2013; Dasog et al. 2014).

In situ particle size (d50) change of flocs is shown in Figure 7, and it is obvious that the flocculant particle size first increases rapidly then decreases, and finally becomes stable during the test process. In general, the in situ particle size detection of flocs shows the following rules in water treatment: first, the particle size of flocs increases rapidly to the maximum particle size (d_max), which can be attributed to the aggregation of flocculant particles; second, the particle size of flocs gradually decreases and tends to a constant value over time, which is related to the stirring (Jarvis et al. 2005; Zhao et al. 2011a, 2011b; Huang et al. 2014a; 2014b; Vajihinejad & Soares 2018). In order to evaluate the performance of flocculants more efficiently and quickly in our experiment, we believe that excellent flocculation and sedimentation ability will be obtained with the higher max floc size (d_max). A short time (Δt₁) for getting the max obtained floc size and a fast floc size increasing rate (k₁) are both beneficial for rapidly bridging absorption and sweep netting of suspended particulates. At the same time, smaller variation (k₂) of floc size indicates excellent stability of flocculant during treatment of contaminants. According to the results from Figure 7 and Table 5, we can reach a conclusion that PST1 has a higher d_max, k₁ and lower k₂, which indicates PST1 will have excellent flocculation and sedimentation ability, rapid physical entrapment/adsorption ability and high stability. This conclusion is in good agreement with the actual flocculation effect evaluation as above.

PST with different Ti/Si molar ratios (2, 1 and 0.5) was prepared with spent titanium solutions. The preparation of PST described in this article could not only solve the problem of how to treat spent titanium solutions but also reuse the valuable resources. The morphology and surface structure of different Ti/Si molar ratio PST are influenced by the content of titanium. With the decrease of titanium content, the morphology of the PST changes from a rod-like shape and finally transforms into a sphere-like shape, while spherical particles are larger when the titanium content is low. In addition, FT-IR results show that the polymerization of Si-O-Si and Si-O-Ti for PST is first strengthened and then weakened with the decrease of titanium content, and the hydrolysis of PSiA could be effectively inhibited. The change of C, O, Ti, and Si XPS energy spectra of the PST samples are basically similar, therein, the valence states of Ti and Si were IV. The performance of the flocculant is evaluated by in situ laser particle size analyzer. The related result indicated PST1 has excellent flocculation and sedimentation ability, rapid physical entrapment/adsorption ability and high stability. Moreover, PST1 also shows the best flocculation performance in sodium humate–kaolin synthetic water and domestic wastewater treatment (turbidity removal efficiency was 96.5% and 89.29%, respectively), and high COD removal efficiency (COD removal efficiency was 89.29% and 61.70%, respectively). The excellent performance of PST1 could be attributed to the special structure, which has a large number of stable Si-O-Ti bonds.

We gratefully acknowledge the National Natural Science Foundation of China (No. 51604132 and 21965017), the Yunnan Provincial Science and Technology Major Project (No. 2018ZE002), and the Provincial Natural Science Foundation of Yunnan (No. 2017FB085 and 2018FB087).