Effects of Structural Factors of Hydrated TiO2 on Rutile TiO2 Pigment Preparation via Short Sulfate Process.

The structural factors such as crystal structure, particle size distribution and impurity content of hydrated TiO2 had great effects on the structures and pigment properties of the rutile TiO2. The rutile TiO2 white pigment was prepared via the Short Sulfate Process from low concentration industrial TiOSO4 solution. In order to produce rutile TiO2 pigment with good structures and excellent pigment properties, the crystal size of the hydrated TiO2 should be controlled less than 8.9 nm and as close as possible to 7.9 nm, which could effectively promote the phase transformation and crystal growth of the rutile TiO2. The appropriate particle size distribution of hydrated TiO2 had obvious effects on obtaining rutile TiO2 with narrower particle size distribution and near 0.20 µm. It was best to adjust the hydrolysis conditions to reduce the specific surface area of the hydrated TiO2 so as to reduce the iron ion impurity adsorption.

saving was obvious, while the lattice parameters of TiO 2 had no effect on rutile and product quality 18 . The effects of ball milling process parameters and stress energy on the particle size distribution of TiO 2 were investigated, and the results showed that particle size distribution had greatly influenced the pigment performances 19 . The particle size and its distribution of TiO 2 had prominently affected its spectral reflectance and color coordinates, and a model was established to investigate the effect of particle size on the aesthetic and thermal properties of poly-dispersed titanium dioxide pigments coatings 20 . By investigating the precipitation and growth behavior of hydrated TiO 2 hydrolyzed from titanyl sulfate solution, the hydrolysis temperature was the most important factor affecting the particle size of hydrated TiO 2 , and the hydrated TiO 2 particles were easy to aggregate in the preparation process, which could be described by an empirical expression 21 . The physicochemical properties of anatase TiO 2 nanoparticles could also be changed by using surface treatment 22 . However, there were few reports about the effects of the structural factors of hydrated TiO 2 on the pigment properties. The short sulfate process refers to the titanium dioxide preparation by using unenriched low concentration TiOSO 4 solution as titanium source via sulfate process, cancelling the concentration section of diluted TiOSO 4 solution, having the advantages of short process, low cost and low energy consumption, which could promote the technological innovation of traditional sulfate process for TiO 2 pigment production 23,24 .
The structures of the hydrated titanium dioxide had great impacts on TiO 2 production. Herein, hydrated TiO 2 was prepared via short Sulfate Process by using the unenriched low concentration TiOSO 4 solution as raw material to produce rutile TiO 2 pigment. It was important to investigate the influences of the crystal structure, particle size distribution and impurity of the hydrated TiO 2 on the TiO 2 pigment preparation.

Experimental
Rutile TiO 2 pigments were prepared from different low concentration industrial TiOSO 4 solution (the total TiO 2 concentration ranging from 155 g/L to 180 g/L, weight concentration) as titanium sources, through thermal hydrolysis by authigenic seed method via the Short Sulfate Process. The typical hydrolysis process was carried out as listed in our literature 23 , and the pre-adding water volume ratio (as water to TiOSO 4 solution) was of 0.18:1, the hydrolysis time after the second boiling point was of 2.5 h, then finished the hydrolysis process and obtained the hydrated TiO 2 . The as-prepared hydrated TiO 2 was washed with water, bleached and filtered, then whipped to slurry with the deionized water uniformly. The slurry was doped with the rutile calcining seed (5%, as to TiO 2 , wt %), zinc salt (ZnO of 0.26%), potassium salt (K 2 O of 0.50%) and phosphate salt (P 2 O 5 of 0.11%), then calcined in a muffle furnace in the air atmosphere. The calcining conditions was as the following: firstly from room temperature raising to 420 °C in 60 min and holding for 30 min at 420 °C, secondly from 420 °C to 780 °C in 60 min and holding for 60 min at 780 °C, and lastly from 780 °C to 870 °C in 120 min and holding for 40 min. Then the rutile TiO 2 pigment powder was obtained after cooling and grinding by the three head grinder. The different concentration of industrial TiOSO 4 solution was conducted at 155 g/L, 161 g/L, 167 g/L, 173 g/L, 180 g/L, and the obtained hydrated TiO 2 samples were marked as A, B, C, D, E, the rutile TiO 2 samples after salt treatment and calcination were denoted as A1, B1, C1, D1, E1, respectively.
The crystal structures of hydrated titanium dioxide and rutile TiO 2 were determined by the XRD analysis (X' Pert3 Powder, PANalytical), and the crystal size L (101) for the anatase crystal plane (101) of hydrated TiO 2 and L (101) for rutile TiO 2 crystal plane (110) was calculated according to Scherrer equation (Eq. A), where K was the constant (0.8900), λ was the wavelength of CuKα 1 (0.15418 nm), β was the full width at half maximum intensity (FWHM) of crystal plane for XRD peak in radians, and θ was the Bragg's diffraction angle, respectively. The rutile content (X R ) was calculated according to Eq. B, where I A and I R represented the integrated intensity of the anatase (101) main peak and the rutile (110) main peak, respectively. Particle size distribution (PSD) test was carried out on a Malvern particle size analyzer (Malvern Zetasizer Nano ZS90). The specific surface area of hydrated TiO 2 was measured on the surface and pore size distribution instrument (3H-2000PS1, Beishide, China). The S BET of the hydrated TiO 2 samples were calculated by the BET multi-point method according to the N 2 adsorption-desorption curves. The UV-vis diffuse reflection spectra were obtained on a ultraviolet visible spectrophotometer with integral ball accessories (U-4100, Hitachi). The particle morphology was observed on a JEOL scanning electron microscopy (JSM-7100F). The surface morphology was carried out on a field emission transmission electron microscopy (Tecnai G2 F20S-TWIN) at 200 kV. The impurities of the rutile samples was determined on an ICP-AES (ICAP 6300, Thermo Scientific Co. Ltd). The ultra-precise colorimeter (LabScan EX, American Hunter) was used to determine the pigment properties, such as the chromatic power (TCS), blue phase (SCX), the brightness (Jasn) and the relative scattering force (Rs), by using the R930 (Ishihara Sangyo Kaisha, Ltd.) as the standard reference sample.

Results and Discussions
Crystal structure. The crystal structure of the obtained hydrated TiO 2 had great impacts on the crystal structure of titanium dioxide pigment, and would ultimately affect the pigment properties of titanium dioxide. The XRD patterns of the as-prepared hydrated TiO 2 series were showed in Fig. 1, and the XRD patterns for rutile TiO 2 in Fig. 2. The crystal size for anatase L (101) of hydrated TiO 2 , and the crystal size for rutile L (110) , the rutile content X R and pigment properties for rutile TiO 2 white pigment were listed in Table 1.
In the hydrolysis system to produce titanium dioxide pigment, the hydrolysis intermediates (also named hydrated TiO 2 , or metatitanic acid) would absorb a large amount of water and sulfate anion to form the crystalline structure with anatase phase, due to the presence of a large number of sulfate anion in the hydrolysis system, with (2020) 10:7999 | https://doi.org/10.1038/s41598-020-64976-4 www.nature.com/scientificreports www.nature.com/scientificreports/ the molecular structure as H 2 TiO 3 ·H 2 O. In Fig. 1, all the XRD patterns of the hydrated TiO 2 samples were clearly consistent with the standard anatase TiO 2 phase (JCPDS 21-1272), without any other crystal phase diffraction peaks, showing with only the anatase phase. The hydrated TiO 2 was anatase phase with low crystallinity due to their wide flat diffraction peaks and low diffraction intensity. The crystal size L (101) for hydrated titanium dioxide ranged from 8.42 nm to 7.10 nm. Hydrated titanium dioxide was transformed into rutile TiO 2 structure after salt treatment and calcinations, and the rutile crystal phase structure was consistent with the standard rutile TiO 2 (JCPDS 21-1276). Due to the difference of crystal size, particle size distribution and impurity content for metatitanic acid, there was a small deviation in the rutile crystallization process during the calcination process, resulting in a small deviation in the XRD pattern. The rutile content (X R ) increased gradually with the decreasing of the crystal size L (101) of hydrated titanium dioxide (as listed in Table 1). The negative linear correlation between X R and L (101) were shown in Eq. (1). In the equation, R represented as the correlation coefficient and SD represented as the standard deviation.   When the crystal size of hydrated TiO 2 was small, it was easier to transform into rutile TiO 2 crystalline by surface atom diffusion during calcination process due to its higher crystal surface energy. In the calcination process, the phase transformation from anatase to rutile was mainly through the surface atomic diffusion, in order to use the atomic reconstruction to reduce the energy of the crystal and form a stable rutile structure. When rutile crystal nucleus was formed, the TiO 2 crystal ions continuously aggregated and grew on the newly formed rutile crystal nucleus through surface atom diffusion, which would make the rutile titanium dioxide crystal growing. And at the same conditions, the smaller crystal size of hydrated TiO 2 needed the lower calcination conditions and strength for crystal phase transformation from anatase to rutile. In order to meet the requirement of rutile content for rutile TiO 2 production (X R > = 98%), the anatase crystal size of the hydrolyzed hydrated TiO 2 should be controlled below 8.89 nm. However, when the crystal size of hydrated TiO 2 was too small and with obvious colloidal properties, the calcined particles would be easily sintering to form larger rutile titanium dioxide particles, resulting in deteriorating the rutile TiO 2 pigments properties.
The negative linear mathematical relationship of crystal size between L (110) for the rutile TiO 2 and L (101) for the hydrated TiO 2 was as the following, Eq. (2).
From Eq. (2), as crystal size decreasing of hydrated TiO 2 , it would be easy to transform hydrated TiO 2 to rutile structure and promote the atomic diffusion in the calcination process basing on the high surface energy of hydrated TiO 2 , which would make the phase transformation and crystal growth of rutile TiO 2 easier and obtain the larger rutile crystals. On the other hand, it would be easier to cause the sample sintering and agglomerate to form larger particles when the crystal size of hydrated TiO 2 was too small.
The pigment properties of rutile TiO 2 pigments (also named initial product) were mainly determined by its crystal structure, particle size and its distribution. The chromatic power (TCS) of the calcined rutile TiO 2 products ranged from 1700 to 1840. Without coating post-treatment, the TCS values for the rutile samples were lower than the reference sample R930 which was with coating post-treatment. The TCS of the samples increased firstly and then decreased as the crystal size of hydrated TiO 2 decreasing gradually. The mathematical relationship between TCS and L (101) was as the following equation, Eq. (3). Equation (3) showed quadratic linear relationship with high fitting degree. After taking the derivation, it showed that the maximum value of TCS was of 1836.2 (about 1840) when the crystal size of hydrated TiO 2 was of 7.90 nm, indicating that the pigment properties of rutile TiO 2 were partly determined by the structure of hydrated TiO 2 . And the crystal size of hydrated TiO 2 could be controlled in an appropriate range by adjusting the hydrolysis conditions such as the number and quality of the hydrolysis seeds, concentration of TiOSO 4 solution and hydrolysis time, which could effectively improve the product pigment properties. It could also approximately predict the pigment properties according to crystal size of hydrated TiO 2 under certain conditions.
The blue phase (SCX) of the rutile pigment showed a similar changing trend as that of TCS, which was gradually increasing from 2.25 to 3.18, and then decreasing to 2.19. It was harmful for phase transformation from anatase to rutile and rutile crystal growth during calcination process when the crystal size of hydrated TiO 2 was too large or too small. Because of the inconsistent phase transformation and growth of rutile crystal, it was easier to cause uneven growth of rutile TiO 2 crystalline grain, ultimately reducing its sintering resistance and the SCX value. The brightness index (Jasn) of the calcined products was better and higher than that of the reference sample R930 (94.52). The Jasn also showed the same changing trend as that of TCS, increasing gradually to 95.17 at first and then decreasing to 94.68.
Appropriate crystal structure and suitable crystal size of hydrated TiO 2 were helpful to promote the phase transformation from anatase to rutile and the rutile crystal growth during the calcination process to obtain good rutile TiO 2 crystalline structure, resulting in proper crystal size of the rutile titanium dioxide. The integrity of rutile crystal structure ensured its high refractive index and scarcely any sintering particles, which could improve the calcination process and obtain rutile TiO 2 with suitable particle size distribution and better crystal structure. All these influencing factors were helpful to improve the rutile pigment properties.
Particle size distribution and impurity iron content. The particle size distribution (PSD) of the rutile titanium dioxide also had important influences on the pigment properties. Because the TCS of the rutile TiO 2 was related to the light scattering coefficient (S) and the light absorption coefficient (K), the larger of S value and the smaller of K value, the larger of TCS, and the better of the covering power and whiteness of the titanium dioxide pigment 25 . Generally, in order to improve the S value, the particle size of the pigment should be controlled in the range of 0.15-0.35 µm in the visible light range, and the amount and content of titanium dioxide particles should be maintained as higher as possible in the range about 0.2 µm. The average particle size (D AV ), particle polydispersity (Pdi), relative scattering force (R S ) and impurities content of the calcined rutile TiO 2 pigments were showed in Table 2.
The D AV of hydrated TiO 2 ranged from 0.782 µm to 1.130 µm. The polydispersity (Pdi) was used to characterize the particle size distribution of the mono-dispersity particles, and the smaller the value was, the more concentrated particle size distribution was. The Pdi value was the index indicating the wide and narrow of particle size distribution, the smaller the Pdi value was, the narrower the PSD was. The Pdi of hydrated TiO 2 ranged from 0.132 to 0.475, and sample C was the smallest with the narrowest PSD. The D AV of the calcined rutile TiO 2 ranged from www.nature.com/scientificreports www.nature.com/scientificreports/ 0.238 µm to 0.372 µm, with the Pdi value ranging from 0.107 to 0.387, of which sample C1 had the narrowest PSD with Pdi of 0.107. It showed that the Pdi value of titanium dioxide prepared by hydrated TiO 2 with smaller Pdi value was also smaller than the others (Table 2), which indicated that narrower PSD of hydrated TiO 2 was beneficial to obtain TiO 2 powders with narrower PSD. The Pdi values for hydrated TiO 2 and rutile TiO 2 were met with the following mathematical relationships. The relationships of Pdi showed positive linear correlation (Eq. (4)). Narrower PSD of hydrated TiO 2 was beneficial to control the crystal phase transformation and crystal growth of rutile TiO 2 in a suitable uniform range during the calcination process. When the PSD of hydrated TiO 2 was narrower, as the particle size was more uniform in the calcining process, and the properties and compositions of the active sites of anatase TiO 2 in the phase transformation and crystal growth were closer, by atomic diffusion and crystal growth on the particle's surface, the obtained calcined rutile particles were uniform. Ultimately, the obtained calcined rutile TiO 2 was with narrower and more uniform PSD, which could effectively improve the pigment properties. While when the PSD of hydrated TiO 2 became wider, the PSD of the calcined rutile TiO 2 became wider due to the inconsistent diffusion process and crystal growth during the calcining process, which ultimately deteriorated the pigment properties. The mass content of titanium dioxide with particle size distribution in the range of 0.15-0.30 µm was listed in Table 2. This further proved that when the PSD was narrower, the higher the mass ratio of particles in the suitable size range for rutile TiO 2 products, the better the corresponding pigment properties. The relative scattering force (Rs) refers to the ratio of the scattering ability of a pigment to the incident light in a certain medium compared with the reference pigment, and the larger the ratio, the better the pigment performance. The relationship between R S and D AV,TiO2 was as the following equation. The relationship between R S and D AV,TiO2 showed negative linear correlation (Eq. (5)). In order to improve the R S value, it was necessary to control the product particle size near 0.20 µm calculated from Eq. (5) and maintain it in a narrow PSD range, so as to obtain rutile TiO 2 with excellent pigment properties.
The UV-vis spectra for the rutile TiO 2 pigments were showed in Fig. 3. The absorption spectra for all the samples were consistent, there was a strong absorption in the wavelength region less than 402.4 nm, and the difference of absorption intensity was not significant. This part of absorption corresponded to the intrinsic absorption of rutile titanium dioxide, which was the energy absorbed by the electron transition from the valence band of  www.nature.com/scientificreports www.nature.com/scientificreports/ titanium dioxide to the conduction band (the intrinsic forbidden band width, 3.08 eV). The fine fluctuation of the absorption spectra might be caused by different sizes of the rutile crystals.
The SEM photographs for the rutile samples were shown in Fig. 4. All the particles showed rutile TiO 2 morphologies with clearly crystal contour, and the high crystalline was consistent with the XRD analysis. The particle size ranges from 70 nm to 300 nm, and the main average particle size was of about 120 nm. During the calcination process, the smaller particles might congregate together to form a larger one due to re-crystallization because its higher surface energy and crystal growth drive, resulting in larger particle size and larger D AV , as showing in Table 2. It could be seen that the larger and wider the particle size of hydrated TiO 2 was, the wider PSD of rutile TiO 2 products was, resulting in poor pigment properties, as listed in Tables 1 and 2. And sample C1 was with the narrowest PSD and the smallest Pdi, and it confirmed that the appropriate PSD for the hydrated TiO 2 would improve the crystal growth and PSD for rutile TiO 2 . This also proved that appropriate particle size and narrow PSD of hydrated TiO 2 were conducive to obtaining narrower and uniform particle size distribution of rutile TiO 2 , and reducing sintering phenomena for the particles, which would lead to improve the pigment properties. The regular crystallographic perfection, good and complete morphology, proper PSD for TiO 2 was all beneficial to improve pigment properties. The TEM photograph for sample C1 was showed in Fig. 5. The crystal profile of sample C1 was clear, and the particle uniformity was good. The average size was about 240 nm, which was consistent with the particle size test results. The size of particles was different, which might be related to the slow hydrolysis rate and uneven hydrolysis process of TiOSO 4 solution. Better particle morphology and particle size distribution would contribute to the improvement of the pigment performances.
Because the hydrolyzed hydrated TiO 2 was with small size of crystal and aggregates, large specific surface area and stronger colloidal properties, the impurities were easy to be adsorbed on and brought out by hydrated TiO 2 . As the adsorbed amount of colored impurities was one of the key factors affecting the structure of titanium dioxide, it was of great importance to improve pigment properties by controlling the adsorbed impurities in an appropriate range. The iron impurity was the main colored impurity, and when the content exceeded the limited range (commonly less than 30 ppm for rutile TiO 2 and less than 90 ppm for anatase TiO 2 ), it would cause the rutile samples yellowing and sintering, which would seriously worsen the pigment performances. The main impurities content including Fe, ZnO, K 2 O + Na 2 O and P 2 O 5 of the rutile TiO 2 pigments were listed in Table 2, and the content of these impurities was low, which had little effect on the properties of the rutile TiO 2 pigments. The specific surface area (S BET ) was a key factor that affecting the adsorbed amount of impurities. The absorbed content of iron ions gradually increased with the increasing of the S BET for hydrated TiO 2 (as showed in Table 2). The relationship between the adsorption amount of impurity iron (% Fe) and the S BET of hydrated titanium dioxide was as the following equation. The % Fe had positive linear correlation with the S BET of hydrated TiO 2 . The S BET of the hydrated TiO 2 should be as small as possible in order to control the % Fe. The S BET of hydrated TiO 2 was connected with the hydrolysis conditions and operational parameters. When the hydrolysis conditions were well controlled, the hydrolysis reaction was conducted more uniform, the precipitated hydrated TiO 2 particles were more well-distributed and the formed aggregates were with narrower particle size distribution and relative smaller S BET . At the same time, the colloidal properties of hydrated TiO 2 would be weakened by increasing the hydrolysis temperature and prolonging the hydrolysis time, which could reduce the S BET and adsorption amount of impurity iron, resulting in better pigment properties. In addition, it was also great important that hydrated TiO 2 with narrower particle size distribution and smaller S BET , which could be conducive to reduce the subsequent washing water consumption, shorten the washing time and reduce the washing strength and cost.

Conclusions
The rutile TiO 2 pigments were prepared through thermal hydrolysis by authigenic seed method via Short Sulfate Process. The structural factors such as crystal structure, particle size distribution, impurity content of the iron ion and specific surface area of the hydrated titanium dioxide had great important impacts on the crystal structure, pigment properties and PSD of the rutile TiO 2 , and there had also an internal influencing relationship among these factors, these factors influenced and determined each other. Suitable crystal size and crystal structure of the hydrolyzed hydrated TiO 2 were helpful to promote the phase transformation from anatase to rutile and crystal growth of rutile TiO 2 , and it was also related to the rutile content, crystal size and pigment properties of rutile TiO 2 satisfying with mathematical regression correlation. It was advisable to control the crystal size of hydrated to be less than 8.9 nm and close to 7.9 nm, which could obtain rutile TiO 2 with good crystal structure which could reduce the sintering of the particles and enhance the pigment performances for rutile TiO 2 pigments. The appropriate particle size and particle size distribution of hydrated TiO 2 had obvious effects on the particle size distribution, polydispersity and relative scattering force of the rutile TiO 2 particles. The adsorption amount of