Introduction

Titanium dioxide (TiO2) is the third largest inorganic chemical, due to its excellent physical and chemical properties, it is widely used in many fields, such as coatings, plastics, rubber, etc.1,2,3. The industrial production of TiO2 mainly includes the sulfate process and the chloride process. The titanium resource reserves in Panzhihua Xichang region of China are the largest in the world. It is more suitable to produce TiO2 by the sulfate process due to its higher calcium and magnesium content. Hydrolysis of industrial TiOSO4 solution is the core step in the preparation of titanium dioxide by sulfate process. It will go through many complex physiochemical processes, including polymerization reactions of hydroxyl bridge and oxygen bridge connection structure to form the nuclei, TiO2+ ionic reaction to promote crystal growth and aggregation, colloidal particle formation, condensation and precipitation, and finally the formation of metatitanic acid (MA) particles4. When TiOSO4 solution is hydrolyzed, a large number of colloidal microcrystals (7–8 nm) were first formed from crystalline TiO2+ ions, then these microcrystals aggregated and formed the primary aggregate agglomerates (60–100 nm). The size and distribution of these primary agglomerates were important key factors determining TiO2 structure and pigment properties. The primary agglomerates re-aggregated to form the secondary aggregates with a particle size of 1–2 μm, which was called MA precipitation5,6. As there was a large amount of ferrous sulfate in the industrial TiOSO4 solution, the mass ratio of Fe-to-TiO2 (Fe/TiO2) had strict requirements in order to meet the TiO2 quality requirements. The ferrous sulfate played an important role in increasing the relative density, viscosity and total ion concentration of the solution, and it would also affect the hydrolysis process and the precipitation of MA. The impurities such as ferrous ion (Fe2+) and SO42- affected the hydrolytic conversion, structure and composition of MA, they would substantially affect the impurities adsorbed by the hydrated TiO2, the adsorption capacity increased with increase of impurity concentration, and at the same time, the agglomeration of the primary particles would increase7. Ferrous ion concentration influenced the conversion degree of TiOSO4 to hydrated TiO2, but did not influence the mean size of the hydrated TiO2 crystallites8,9. Researchers had widely investigated the influences of hydrolysis conditions of the TiOSO4 solution hydrolysis system on hydrolysates and TiO210,11,12,13,14. Titanium dioxide white pigment with narrow particle size distribution and good performances could also be prepared from the un-enriched industrial TiOSO4 solution by adjusting the hydrolysis conditions15,16. Whiteness was a very important index of TiO2 pigment, and ferrous ion content was the most important influencing factor. The less the ferrous ion content was, the higher the whiteness of TiO2 was. The aims of adjusting hydrolysis conditions, filtration, rinsing and other processes were to control the ferrous ion content in MA, so as to obtain good quality of TiO2. The increase of ferrous ion concentration would affect the formation of MA, reduce the reacting activity of TiO2+ ion, and influence the crystallization and particle size of MA. The composition and structure of MA would change with hydrolysis process, which would also affect the adsorption of water, sulfate and impurities, affect the crystal transformation and crystal growth in the subsequent calcination process, ultimately affected the purity of MA and product performances of TiO2. The problems were that the structural changes of MA during the hydrolysis process and its impact on the content of ferrous ions made it difficult to accurately control the quality and impurities content of MA. However, there were few reports about this.

The aim of this work was to investigate the effects of the structural evolution of MA during hydrolysis of industrial TiOSO4 solution on the ferrous ion content in MA, which would be of great significance to control the hydrolysis conditions to obtain hydrolyzed MA with good composition and structure, so as to ensure the quality and performances of titanium dioxide pigment.

Experimental

Hydrolysis

Thermal hydrolysis for the industrial TiOSO4 solution as raw material was carried out by extra-adding seeded hydrolysis method. The TiOSO4 solution was an industrial grade raw material, and the water used was deionized water. The typical composition of industrial TiOSO4 solution was with the total TiO2 concentration of 194 g/L, F value of 2.03 (F value meant the mass ratio of free sulfuric acid and sulfuric acid combined with Ti4+ (TiOSO4) to TiO2, as with free sulfuric acid concentration of 1.59 mol/L), Fe/TiO2 ratio of 0.30 (as with ferrous ion content of 1.04 mol/L), Ti3+ concentration of 1.5 g/L. The hydrolysis reacted in a four port round bottom flask with heating, stirring and condensation reflux, the typical hydrolysis was conducted as the following procedure. The extra-adding seed, with amount of 2.1% and concentration of 144 g/L, was added to the above industrial TiOSO4 solution which was pre-heated to 96 °C, and the starting point of hydrolysis time was the completion of feeding. After adding, the hydrolysis slurry was uniformly mixed and heated to the first boiling point at 107 °C, then kept in a slightly boiling state. When the hydrolysis slurry turned into gray color at 60 min after feeding, the hydrolysis reaction entered the ageing stage by stopping heating and stirring for another 30 min. After ageing, the slurry was heated again to the second boiling point at 108 °C in 17 min, then also keeping in slightly boiling state. The dilution water was slowly added to the hydrolysis system with volume ratio of 2% in 15 min at hydrolysis time of 227 min. The hydrolysis reaction was stopped at 287 min. At different hydrolysis time, 200 mL hydrolysis slurry was taken out from the hydrolysis system and filtered without washing, the obtained sample called wet MA. Metatitanic acid powders were obtained by grinding after drying the MA filter cake at 100 °C for 6 h, and the obtained MA powders were called dried MA.

Characterization

The hydrolysis degree was determined by measuring the residual TiO2 in the filtered TiOSO4 solution, TiO2 content by determining the TiO2 content of the dried MA, according to the standard ISO 591-1:2000 ‘Titanium dioxide pigments for paints’, by using the ammonium ferric sulfate oxidation–reduction titration method. A X-ray diffractometer (X’ Pert3 Powder, PANalytical) was used to determine the crystal structure for the wet MA samples, by using a tube voltage of 40 kV, a tube current of 40 mA, scanning angle from 20° to 70°, with a scanning step size of 0.0133 and 0.1 s/step. And the anatase grain size L(101) for crystal plane (101) was calculated according to the Scherrer equation17. The lattice strain of MA in the C-axis direction was obtained by refining the structure through the reflex module of MS software. The SEM morphologies of the MA samples were obtained by the field emission scanning electron microscopy (Sigma 300, Zeiss, Germany), by using a test voltage of 30.0 kV, WD of 8.0 mm and secondary electronic signal. Particle size distribution (PSD) was determined by a Malvern particle size analyzer (Mastersizer 2000, Malvern), by using a wet dispersion system, helium neon gas laser light source (633 nm) and blue assisted light source (466 nm), with a shading ratio of 13%. BET surface area (SBET) and the pore size distribution was measured on the surface and pore size distribution analyzer instrument (Autosorb-iQ3, Quantachrome, USA), the sample was first degassed under vacuum conditions at 200 °C for 6 h, and then measured using nitrogen as the adsorbate. The Raman spectra for MA samples were obtained from a micro confocal Raman spectroscopy by using laser with wavelength of 532 nm for measurement (inVia, Renishaw, UK), by using a 532 nm laser for testing, static scanning, wavenumber range of 87.93–959.47 cm−1, exposure time of 10 s, exposure intensity of 5%. The FT-IR test and analysis were carried out on the infrared spectrometer (Nicolet-380, Thermo, USA), by using KBr compression method, with a wavenumber range of 400–4000 cm−1, a maximum resolution of 0.05 cm−1, and the scanning speed of 20 cm−1/min. The ICP-OES (iCAP 6300, Thermo Fisher, USA) was used to determine the Fe content of the MA samples, by using the standard curve method, determined the characteristic emission intensity of ferrous ions at 259.940 nm for the tested sample, and then determined its Fe content through the standard curve. The binding energies for MA samples were determined by an X-ray photoelectron spectroscopy (XPS) (XSAM-800, Kratos), by using Al Kα (1486.6 eV) X-ray gun operation, 12 kV × 15 mA, the analytical instrument adopted high magnification, fixed reduction ratio, high-resolution mode, and the analyzer was calibrated with Au and Ag standard samples.

Results and discussion

The MA samples were taken from the hydrolysis slurry for filtration and separation to determine its structure and composition as the hydrolysis proceeding. Since the initial formed MA particles were too small to be separated, the MA samples were taken from the gray point (that is, the gray point was that the hydrolytic slurry turned into steel gray color at 60 min) and after. The hydrolysis degree of industrial TiOSO4 solution, TiO2 content, crystal size (L(101),MA), lattice strain, average particle size (D), BET surface area (SBET), pore diameter, ferrous ion content (Fe %) for the MA samples were listed in Table 1. The variation curve of hydrolysis degree at different hydrolysis time was showed in Fig. 1.

Table 1 Hydrolysis degree, compositions and structures for metatitanic acid at different hydrolysis time.
Full size table
Figure 1
figure 1

The variation curve for the hydrolysis degree.

Full size image

The variation curve of hydrolysis degree could be divided into two stages, as shown in Fig. 1. The first stage from the gray point to 125 min was the rapid hydrolysis stage by first derivative of the curve, characterized by its large change in the hydrolysis degree and large precipitation of MA. The second stage from 125 min to the end of hydrolysis was the maturation stage, the hydrolysis degree changed slowly and tended to be stable and balanced gradually, and reached at the highest value of 96.65% at the end of hydrolysis process. The dilution water was added at 227 min in order to improve the hydrolysis degree, which could make the overall hydrolysis degree increase by 2%, so as to meet the hydrolysis degree requirements. The change of hydrolysis degree conformed to Boltzmann model, and the correlation coefficient R2 was of 0.98866, indicating that the results had a good fitting degree. The fastest hydrolysis rate was at 74.0 min when the hydrolysis curve was differential treated, and then the hydrolysis rate gradually slowed down. In the first hydrolysis stage, it was mainly controlled by the surface reaction of MA particle and grain growth as there were a lot of TiO2+ ions in the high concentration of TiOSO4 solution to form the MA crystals. While in the second hydrolysis stage, the reaction was mainly controlled by diffusion process due to the low concentration of TiO2+ ions to increase the crystal growth, and the reaction rate was relatively slow.

As the hydrolysis time increased, the TiO2 content of the dried MA samples increased rapidly at first and then slowly after the hydrolysis time of 107 min, as shown in Table 1. After that it was stable at about 73–80% in the subsequent process. This indicated that the structure of MA became more and more compact with the extension of hydrolysis time, and the adsorbed water content in MA decreased, showing that the colloidal properties of MA was weakened. At the later hydrolysis stage, some fine MA particles were precipitated due to a small amount of TiOSO4 solution remaining in the hydrolysis slurry and the dilution water adding. These fine MA particles would absorb more water and impurities, and aggregate with the previously precipitated MA particles because they were ultra-fine and had high surface activity, resulting in TiO2 content decreasing. In addition, the structure and TiO2 content would be also affected by the re-aggregation and particle adjustment of the precipitated MA particles during the hydrolysis process.

The XRD patterns for the wet MA samples were showed in Fig. 2, clearly agreeing with the main diffraction peaks of the standard anatase phase (JCPDS 21-1272), and anatase structure of would be obtained after calcination process. This was because the metatitanic acid formed in the SO42- system was prone to obtaining anatase TiO2 like crystal structure. As the hydrolysis time increased, the diffraction peak intensity gradually increased and the diffraction peak became narrow and sharp, indicating that the MA crystallinity gradually increased. The calculated crystal size of anatase face (101) for the MA samples ranged from 12.7 to 23.8 nm, also showed the L(101) increased with the increasing of the hydrolysis time, consistent with the aforementioned XRD analysis. The crystal size increased significantly from 167 to 197 min, indicating that the crystal size evidently grew when the concentration of TiOSO4 solution was low, consistent with the above analysis of the slow hydrolysis stage. The high angle diffraction peak gradually shifted to the low angle as the hydrolysis proceeding in Fig. 2, indicating that the unit cell parameters gradually became larger and the crystal plane spacing increased, mainly a reflection of lattice distortion caused by macro residual stress, which might be caused by the tensile stress during the crystal growth of MA. This also showed that the initial formed MA had loose structure and high stress. With the crystal growth and hydrolysis process, the crystal structure of MA was constantly adjusted, and the corresponding stress gradually decreased. The lattice strain for the MA samples showed a slightly decreasing trend, mainly fluctuating at about 0.54%. At the late hydrolysis stage, the lattice strain increased due to the newly formed ultra-fine MA particles. The minimum lattice strain of MA was at the end of hydrolysis, with the value of 0.524%, which indicated that the MA crystal was constantly adjusted to gradually reduce stress and maintain low energy during the hydrolysis process.

Figure 2
figure 2

XRD patterns for the metatitanic acid samples.

Full size image

The D for the wet MA samples was also listed in Table 1, ranged from 1.31 to 1.75 μm, increased firstly and then decreased as the hydrolysis proceeding. The diameter distance ratios of the MA particles were about 1.00, and the overall changes were not significant, indicating that the width of particle size distribution was relatively close during different hydrolysis time. Metatitanic acid particles were easy to aggregate due to their smaller crystals and higher surface energy, and the measured D mainly corresponded to the secondary aggregates, which was formed by the primary agglomerates composed of many MA crystal particles. With the crystal size of MA increasing, the primary agglomerates would be smaller, and the secondary aggregates would be larger, which was consistent with the changing trend of D. Due to the addition of dilution water, some fine MA particles were formed, and their structure and aggregation state were changed, resulting in the reduction of the D for the MA samples.

The SEM photographs for the dried MA samples were showed in Fig. 3. After ultrasonic dispersion in ethanol solvent, the MA samples mainly existed in the form of aggregates, while containing some small particles, and the aggregates were formed by finer particles. The size of dispersed fine particles was at about 50–70 nm, which corresponded to the primary agglomerates of the MA. The photographs also showed the different sizes and aggregation states of the precipitated MA particles at different hydrolysis time. The secondary nucleation promoted the formation of crystal clusters, then formed the primary agglomerates through surface nucleation, and formed the micron aggregates by physical forces5.

Figure 3
figure 3figure 3

SEM photographs for the metatitanic acid samples (50kX).

Full size image

The SBET of the dried MA samples ranged from 268 to 311 m2/g in Table 1. With the hydrolysis time increasing, the SBET first gradually increased from 268 m2/g at the gray point to 311 m2/g at 137 min, then gradually decreased to 285 m2/g at the end of hydrolysis. As the hydrolysis proceeding, the pore diameter of MA samples mainly fluctuated at about 3.4 nm, and the overall change was very small. The N2 adsorption–desorption isotherms for sample 9 was close to the type IV in Fig. 4, there was a hysteresis loop because the adsorption desorption process was not completely reversible, indicating that there were pore structures in the metatitanic acid aggregates. As the relative pressure range of hysteresis loop was wide, which indicated that its pore size distribution range was wide, distributing in micropores and mesopores. The pore size distribution curve by DFT method for sample 9 was shown in Fig. 5, the average pore diameter was 3.1 nm, and the most probable pore size in the microporous range was 1.4 nm, the most probable pore size in the mesoporous range was of 2.7 nm, consistent with the previous analysis of the hysteresis loop.

Figure 4
figure 4

Nitrogen isotherms for sample 9.

Full size image
Figure 5
figure 5

Pore size distribution curve for sample 9.

Full size image

The Raman spectra for sample 7 and sample 9 were showed in Fig. 6, and spectral deconvolution was used to determine the peak position and width, and the peak positions were consistent with the Raman shift of anatase TiO2. And the phonon confinement model was used to explain the broadening and shifts of the Raman line shapes18,19. As hydrolysis proceeding, the peak positions shifted to the higher wave number position, and its corresponding peak width also narrowed, indicating that the crystal size of the MA sample increased, consistent with the previous XRD analysis. The infrared absorption of MA samples at different hydrolysis time was similar, and the FT-IR spectrum for sample 9 was showed in Fig. 7. The absorption peak at wave number 1632 cm−1 corresponded to the bending vibration of the physically adsorbed molecular water H–O–H, and the absorption shoulder peak with a wide range of 3400–3600 cm−1 corresponded to the stretching vibration of the sample surface hydroxyl O–H20. The peaks at 1049 cm−1 and 1131 cm−1 were characteristic absorption peaks of bidentate coordination sulfate ion21,22, which would combine with Ti–O bond to form the corresponding pore structure. The micropores and mesopores formed by MA were mainly formed by the aggregation of primary agglomerate particles in different sizes, bonding and filling with sulfate and water.

Figure 6
figure 6

Raman spectrum for sample 7 and sample 9.

Full size image
Figure 7
figure 7

FT-IR spectrum for sample 9.

Full size image

The XPS fitting spectrum for sample 9 was showed in Fig. 8. The ferrous ions were oxidized to ferric ions in MA samples after long-term exposure to air. The Fe2p fitting spectrum consisted of two peaks in Fig. 8a, corresponding to Fe3+2p3/2 and Fe3+2p1/2 energy levels, the binding energies were 710.9 eV and 724.4 eV respectively. Compared with standard pure Fe2O3 (the binding energies of Fe3+2p3/2 and Fe3+2p1/2 energy levels were 710.7 eV and 724.3 eV), the positive shifts of 0.2 eV and 0.1 eV might be the result that Fe3+ entered the crystal lattice of H2TiO3, formed the Ti–O–Fe bond and changed the chemical potential and polarity23. The Fe (II) ions in the near surface region of the system had lower potential energy for both dry and hydrated surfaces24, which was easy to enter the metatitanic acid surface for bonding, thus forming the corresponding bond valence structure. During the hydrolysis process, ferrous ions first diffused to the interface layer of MA precipitation through the main TiOSO4 solution, then diffused to the surface of MA through the interface layer, and formed the corresponding Ti–O–Fe structure after adsorption and bonding. And these ferrous ions in the bonding state were difficult to remove. There was only a single peak in the S2p energy level fitting map in Fig. 8b, and its binding energy was 169.1 eV, indicating that sulfur only existed in the form of S6+. The two peaks of the O1s energy levels in Fig. 8c were located at 530.4 eV and 532.1 eV, corresponding to the lattice oxygen, surface hydroxyl, and the O1s energy level in the chemisorbed water25. The two peaks of the Ti2p in Fig. 8d were located at 459.1 eV and 464.6 eV, corresponding to the Ti4+2p3/2 and Ti4+2p1/2 energy levels. Compared with the pure anatase TiO2 (458.1 eV and 463.8 eV), there was a positive shift, indicating that the Ti4+ ion for sample 9 had greater positive electricity due to the strong induction effect of the SO42−/TiO2 chelating bidentate coordination structure26.

Figure 8
figure 8

XPS fitting spectrum for sample 9 (a) Fe2p; (b) S2p; (c) O1s; (d) Ti2p.

Full size image

The ferrous ion content (Fe %) in MA samples decreased significantly with the hydrolysis time increasing, as listed in Table 1, and the ferrous ion content basically remained at about 0.10% at hydrolysis time of 167–227 min, mainly due to the newly precipitated fine MA particles in this period were few, the adsorbed ferrous ion content changed little, and reached the smallest value of 0.09% at 227 min. After adding dilution water at 227 min, the amount of impurities adsorbed increased, and the ferrous ion content increased, due to the formation of some smaller MA particles with strong adsorption capacity. There was a nearly linear relationship between the ferrous ion content and the TiO2 content of the dried MA samples, as shown in Fig. 9, and the correlation coefficient R2 was of 0.98263, indicating that it had a good linear relationship, and the water content in MA was the key factor causing the change of ferrous ion content, because most ferrous ions were dissolved in the water adsorbed by MA.

Figure 9
figure 9

Relationships between ferrous ion content and TiO2 content.

Full size image

The MA colloidal particles absorbed a large amount of water and impurities due to their bigger SBET, while impurities such as Fe2+ ion would be dissolved in the adsorbed water except for a small amount of them were adsorbed and bounded to MA. It could be seen that, with the extension of hydrolysis time, the grain size increased, the TiO2 content of MA gradually increased, the structure of MA became more compact, and its colloidal properties weakened. On the one hand, the adsorption ability of ferrous ion for MA on its surface was weakened, resulting in a decrease in the amount of ferrous ion adsorbed on its surface. On the other hand, the amount of water absorbed by MA was also reduced, resulting in a decrease in the number of dissolved ferrous ion in the adsorbed water. The two effects of structural changes in MA resulted in a decrease in its Fe content. The research results were consistent with the structural changes of MA reported in literature27. In order to obtain high quality MA, ultrasonic washing could be used and the pH value of the washing water could be adjusted to further save water and remove impurities such as ferrous ions28. In practice, to improve the purity of MA and the quality of titanium dioxide pigments, a feasible method is to properly increase the crystallinity of MA and the solid TiO2 content of filter cake, weaken the colloidal properties, and use higher pressure filtration device to reduce the moisture content of metatitanic acid and impurity content. The reduction of impurity content in MA would greatly save the water consumption in the subsequent washing and impurity removal process, and the lower moisture content of MA would also greatly save the energy consumption in the calcination process. And this would provide new ideas and ways to improve the clean production level of titanium dioxide production.

Conclusions

Metatitanic acid was prepared from industrial TiOSO4 solution by extra-adding seeded thermal hydrolysis method. The hydrolysis process after gray point could be divided into the rapid hydrolysis stage and the slow hydrolysis maturation stage, and the hydrolysis degree changes conformed to Boltzmann model with good fitting. The structural evolution of the precipitated MA changed obviously as the hydrolysis proceeding, the content of TiO2 for MA gradually increased due to its more compact structure and weaker colloidal property, which was determined by the aggregation and particle adjustment of MA particles. The crystal size of MA increased significantly at the lower TiOSO4 concentration, and its lattice strain gradually decreased with the crystal growth and structure adjustment, which led that the average particle size was constantly adjusted and reduced. The aggregating and stacking of the precipitated MA particles formed the microporous and mesoporous structures, with the BET surface area larger than 265 m2/g. The Ti–O bond of metatitanic acid was mainly combined with sulfate and hydroxyl, and the pores were filled with sulfate and water. Ferrous ions mainly existed in MA by dissolving in the adsorbed water, a small amount existed in the form of adsorption and bonding. And the ferrous ion content decreased linearly with the increase of TiO2 content.