Microstructure and Morphology Control of Potassium Magnesium Titanates and Sodium Iron Titanates by Molten Salt Synthesis

Titanates materials have attracted considerable interest due to their unusual functional and structural properties for many applications such as high-performance composites, devices, etc. Thus, the development of a large-scale synthesis method for preparing high-quality titanates at a low cost is desired. In this study, a series of quaternary titanates including K0.8Mg0.4Ti1.6O4, Na0.9Mg0.45Ti1.55O4, Na0.75Fe0.75Ti0.25O2, NaFeTiO4, and K2.3Fe2.3Ti5.7O16 are synthesized by a simple molten salt method using inexpensive salts of KCl and NaCl. The starting materials, intermediate products, final products, and their transformations were studied by using TG-DSC, XRD, SEM, and EDS. The results show that the grain size, morphology, and chemical composition of the synthesized quaternary titanates can be controlled simply by varying the experimental conditions. The molar ratio of mixed molten salts is critical to the morphology of products. When KCl:NaCl = 3:1, the morphology of K0.8Mg0.4Ti1.6O4 changes from platelet to board and then bar-like by increasing the molar ratio of molten salt (KCl–NaCl) to raw materials from 0.7 to 2.5. NaFeTiO4 needles and Na0.75Fe0.75Ti0.25O2 platelets are obtained when the molar ratio of molten salt (NaCl) to raw materials is 4. Pure phase of Na0.9Mg0.45Ti1.55O4 and K2.3Fe2.3Ti5.7O16 are also observed. The formation and growth mechanisms of both potassium magnesium titanates and sodium iron titanates are discussed based on the characterization results.


Introduction
The titanates is a group of inorganic compounds consisting of titanium, oxygen, and one or more other metallic elements. Dependent on the linkage method of the structure unit of TiO 6 octahedra, titanates may exhibit cage, tunnel, and layered structures. The commonly employed titanates in industrial applications include CaTiO 3 , BaTiO 3 , SrTiO 3 , and M 2 O.nTiO 2 (M = K/Na, n = 1~8), covering a wide range from the medical to electrical and automotive industries. As opposed to the ternary titanates with well-investigated properties and developed applications, quaternary titanates are still in the exploring stage, partly due to the large range of compositions and the complex structural deviation. Among all quaternary titanates, K 2 O-MgO-TiO 2 [15] studied the structural stabilization of iron containing cathode materials by substituting some iron in α-NaFeO 2 with titanium to produce Na x Fe x Ti 1−x O 2 (0.75 ≤ x ≤ 1.0). The studies above mainly focus on the investigation of compositional and structural variations and the potential applications on specific compositions. Various compositions have been made by different kinds of synthesis methods, such as high-temperature calcination, molten salt synthesis, kneading-drying-calcination (KDC), etc. [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15]. The subject of this study is to achieve a stable production of high-quality quaternary octatitanates with controllable morphology and narrow size distribution for potential applications in inorganic fiber-reinforced composites and sodium ion batteries. The molten salt method and low-cost raw materials have thus been exclusively used for future scalable industry production. Through the adjustment of the content (α) of molten salt in raw materials, the ratio (β) of KCl in KCl-NaCl molten salt, and the reaction temperature and time, we obtained pure phase of lepidocrocite-like K 0. 8  Two NFTO products, namely NaFeTiO 4 needles and Na 0.75 Fe 0.75 Ti 0.25 O 2 platelets, are obtained when T = 900 and 1000 • C; α = 4; β = 0; and t = 4. The products and their intermediate products are characterized by scanning electron microscopy, X-ray diffraction, and thermogravimetric analysis for a better understanding of their formation and growth processes. The current synthesis procedure can be scaled for controllable production of these types of titanates.

Adjustment of Molten Salt Content
The molten salt content α is defined as the molar ratio of molten salt to raw materials, namely α = n molten salt /n raw materials . The raw materials include the molten salt and the starting materials (see 2.3 and 2.4). The molten salt ratio β is defined as the molar ratio of KCl to KCl-NaCl, namely β = n KCl /n mixture of NaCl-KCl .

Preparation of Sodium Iron Titanates (NFTO)
Sodium iron titanates were produced by the same procedure as described in the preparation of KMTO. The starting materials of Na 2 CO 3 and FeTiO 3 , with or without Fe 2 O 3 , were mixed with KCl-NaCl molten salt (β = 0, 0.25, 0.5, 0.75, and 1) at a certain molar ratio (α = 2, 4, and 6). The mixture was calcined at 600, 700, 800, 900, and 1000 • C for 2, 4, or 6 h for the procedure optimization. Two kinds of NFTO products, NaFeTiO 4 needles and Na 0.75 Fe 0.75 Ti 0.25 O 2 platelets, were obtained at two sets of optimum conditions. Without Fe 2 O 3 as the Fe source, NaFeTiO 4 needles were prepared at the condition of Na:Fe:Ti = 1.

Characterizations
The crystalline phase of samples were examined with X-ray diffraction (XRD) by using a D8-Advance, Bruker AXS diffractometer (Cu-Kα radiation, λ = 1.5418 Å) in the continuous scan mode over 5-70 • (2θ) with a scan rate of 0.3 • /s, operating at 40 kV and 40 mA. The morphology and microstructure of samples were characterized by field-emission scanning electron microscopy (FESEM, HITACHI S−4800, Hitachi, Tokyo, Japan) equipped with energy-dispersive X-ray spectroscopy (EDS). Thermogravimetric analysis (TGA) was performed on a NETZSCH 449 STA thermogravimetric analyzer (Netzsch, Sabre, Germany). The samples were heated in N 2 atmosphere from 30 to 1100 • C at a heating rate of 10 • C·min −1 .  Figures 1 and 2 show the SEM images and XRD patterns of the KMTO samples prepared at different molar ratios of the molten salt KCl-NaCl to the raw materials (α = 0.7, 1.5, and 2.5) after calcination at 1050 • C for 4 h. The morphologies of KMTO products are shaped like platelets (α = 0.7, Figure 1a1,a2), boards (α = 1.5, Figure 1b1,b2), and bars (α = 2.5, Figure 1c1,c2), respectively. As shown in Figure 2, the major phase of all the three differently shaped products exhibit the XRD patterns belonging to K 0.8 M g0.4 Ti 1.6 O 4 (PDF#35-0046). A very small amount of impurity peaks of hydrated potassium tianium hydrogen oxide hydrate (K 0.5 H 1.5 Ti 4 O 9 ·0.6H 2 O) is also observable in KMTO boards and bars (Figure 2c,d). This impurity phase may be due to the dissolution of K + and Mg + when the products were washed by DI water. The relative peak intensity of three samples differs from that of the standard XRD pattern and the product synthesized without the presence of molten salts (α = 0, Figure 2a). The deviation of the peak intensity is caused by the preferential growth of samples. The elemental analysis from EDS is consistent with XRD, see Supplementary Figure Figure 1b1,b2), and bars (α = 2.5, Figure 1c1,c2), respectively. As shown in Figure 2, the major phase of all the three differently shaped products exhibit the XRD patterns belonging to K0.8Mg0.4Ti1.6O4 (PDF#35-0046). A very small amount of impurity peaks of hydrated potassium tianium hydrogen oxide hydrate (K0.5H1.5Ti4O9·0.6H2O) is also observable in KMTO boards and bars (Figure 2c,d). This impurity phase may be due to the dissolution of K + and Mg + when the products were washed by DI water. The relative peak intensity of three samples differs from that of the standard XRD pattern and the product synthesized without the presence of molten salts (α = 0, Figure 2a). The deviation of the peak intensity is caused by the preferential growth of samples. The elemental analysis from EDS is consistent with XRD, see Supplementary Figure      The growth mechanism of the KMTO platelets, boards, and bars are proposed as depicted in Figure 3, based on the melting point of the molten salt (675 • C for β = 0.75), calculations on the thermodynamics of the reactions between sodium and potassium cations, and analyses using SEM ( Figure S2), XRD ( Figure S4), and TG-DSC ( Figure S5) on the intermediate products during the entire heating process. As the calcination temperature increases, 4MgCO 3 ·Mg(OH) 2 ·5H 2 O first loses hydration water and then decomposes to MgO and CO 2 . Upon the dissolution of K 2 CO 3 and MgO in the NaCl-KCl molten salt, K + and Mg 2+ ions diffuse at different rates in the liquid phase, approaching the dispersed TiO 2 particles. When α = 0.7, KMTO particles are directly formed and then gradually evolve to crystalline platelets as the temperature reaches 1050 • C. When α ≥ 1.5, with abundant Na + in the system, low-melting intermediate phase Na 8 Ti 5 O 14 (melting point 965~985 • C) is formed first and then it interacts with Mg 2+ and K + in the melt to form more stable NMTO bars (melting point 1100 • C). Based on the thermodynamic calculation, the ion exchange from Na + to K + will spontaneously occur when the system temperature is above 675 • C. So as the temperature continues to increase, Na + in NMTO exchanges with K + from the molten salt, resulting in a more stable high melting KMTO phase (melting point 1300 • C) which retains a long strip shape. The morphology and crystalline phase of the intermediate NMTO phase were confirmed by characterizing the samples rapidly annealed at 750, 850, and 950 • C by using SEM and XRD ( Figures S2 and S4). °C). Based on the thermodynamic calculation, the ion exchange from Na + to K + will spontaneously occur when the system temperature is above 675 °C. So as the temperature continues to increase, Na + in NMTO exchanges with K + from the molten salt, resulting in a more stable high melting KMTO phase (melting point 1300 °C) which retains a long strip shape. The morphology and crystalline phase of the intermediate NMTO phase were confirmed by characterizing the samples rapidly annealed at 750, 850, and 950 °C by using SEM and XRD ( Figures S2 and S4). This kind of morphology control cannot be obtained by using KCl (β = 1) or NaCl (β = 0) alone as the molten salt. Figures 4 and 5 show the XRD patterns and SEM images of the samples prepared at different mole ratios of the molten salt to the raw materials (α =0 .5, 2, and 6) after calcination at 1050 °C for 4 h. When using KCl as the molten salt (β = 1), all three samples are pure KMTO phase (Figure 4a1-a3). As the molar ratio α increases from 0.5 to 2 and 6, the relative peak intensity changes. However, all three samples have the platelet morphology of several micrometers (Figure 5a1-a3), indicating that single KCl molten salt cannot cause the platelet-board-bar morphology evolution as KCl-NaCl. When using single NaCl as molten salt (β = 0), a new phase of sodium magnesium titanate (Na0.9Mg0.45Ti1.55O4) appears and the product becomes slender as the amount of NaCl in the molten salt increases. At α = 0.5, the product is a mixture of KMTO particles and NMTO bars (Figures 4b1 This kind of morphology control cannot be obtained by using KCl (β = 1) or NaCl (β = 0) alone as the molten salt. Figures 4 and 5 show the XRD patterns and SEM images of the samples prepared at different mole ratios of the molten salt to the raw materials (α = 0.5, 2, and 6) after calcination at 1050 • C for 4 h. When using KCl as the molten salt (β = 1), all three samples are pure KMTO phase (Figure 4a1-a3). As the molar ratio α increases from 0.5 to 2 and 6, the relative peak intensity changes. However, all three samples have the platelet morphology of several micrometers (Figure 5a1-a3), indicating that single KCl molten salt cannot cause the platelet-board-bar morphology evolution as KCl-NaCl. When using single NaCl as molten salt (β = 0), a new phase of sodium magnesium titanate (Na 0.9 Mg 0.45 Ti 1.55 O 4 ) appears and the product becomes slender as the amount of NaCl in the molten salt increases. At α = 0.5, the product is a mixture of KMTO particles and NMTO bars (Figures 4b1 and 5b1). When α increases to 2, the KMTO phase disappears and the product becomes pure NMTO rods (Figures 4b2 and 5b2), indicating that NaCl in the molten salt provided Na + to participate in the crystal growth reaction. When α increases to six, the product is long NMTO whiskers (Figures 4b3 and 5b3). Although the equilibrium constant for NaCl(l) + K + (s) → Na + (s) + KCl(l) is 1 [16], the K + ions in the layer structured KMTO can still be displaced by Na + via ion exchange when the concentration of surrounding Na + ions is sufficiently large. The above experimental results indicate that the molten salt does not only provide a liquid phase environment for reactions, however the cations may also participate in reactions and strongly affect the growth of crystalline products. crystal growth reaction. When α increases to six, the product is long NMTO whiskers (Figures 4b3  and 5b3). Although the equilibrium constant for NaCl(l) + K + (s) → Na + (s) + KCl(l) is 1 [16], the K + ions in the layer structured KMTO can still be displaced by Na + via ion exchange when the concentration of surrounding Na + ions is sufficiently large. The above experimental results indicate that the molten salt does not only provide a liquid phase environment for reactions, however the cations may also participate in reactions and strongly affect the growth of crystalline products.   crystal growth reaction. When α increases to six, the product is long NMTO whiskers (Figures 4b3  and 5b3). Although the equilibrium constant for NaCl(l) + K + (s) → Na + (s) + KCl(l) is 1 [16], the K + ions in the layer structured KMTO can still be displaced by Na + via ion exchange when the concentration of surrounding Na + ions is sufficiently large. The above experimental results indicate that the molten salt does not only provide a liquid phase environment for reactions, however the cations may also participate in reactions and strongly affect the growth of crystalline products.    transformed to products at appropriate annealing temperatures (NaFeTiO 4 , 900 • C; Na 0.75 Fe 0.75 Ti 0.25 O 2 , 1000 • C). The chemical composition of products is also confirmed by EDS analyses ( Figure S6). As shown in Figure 7, NaFeTiO 4 is in the shape of needles with the length of 20-50 µm and diameter of 0.5-2 µm, while Na 0.75 Fe 0.75 Ti 0.25 O 2 has the platelet shape with the size range of 5-20 µm. The products show the best morphology when the molar ratio of NaCl to the raw materials is α = 4 ( Figure 8). The influence of the ratio of reactants on the product was also investigated and the condition of Na:Fe:Ti = 3.3:2.2:1 shows the pure phase Na 0.75 Fe 0.75 Ti 0.25 O 2 with relatively uniform morphology ( Figure S7). Figures S8 and S9 show the products prepared at different reaction temperatures (600, 700, 800, 900, and 1000 • C) and different reaction durations (2, 4, and 6 h). NaFeTiO 4 needles with best morphology were obtained at 900 • C for 4 h and Na 0.75 Fe 0.75 Ti 0.25 O 2 platelets were obtained at 1000 • C for 4 h. Figures 6 and 7 show the XRD pattern and SEM images of the NFTO samples synthesized while using NaCl as the molten salt. At the condition of Na:Fe:Ti = 1.3:1:1 and Na:Fe:Ti = 3.3:2.2:1, the XRD patterns of the products can be assigned to NaFeTiO4 (PDF#33-1255, Figure 6a) and Na0.75Fe0.75Ti0.25O2 (PDF#25-0877, Figure 6b), respectively. None of the noticeable peaks belong to the unreacted reactants (Na2CO3) or intermediate phase (Fe2O3), indicating that the starting materials have been completely transformed to products at appropriate annealing temperatures (NaFeTiO4, 900 °C; Na0.75Fe0.75Ti0.25O2, 1000 °C). The chemical composition of products is also confirmed by EDS analyses ( Figure S6). As shown in Figure 7, NaFeTiO4 is in the shape of needles with the length of 20-50 μm and diameter of 0.5-2 μm, while Na0.75Fe0.75Ti0.25O2 has the platelet shape with the size range of 5-20 μm. The products show the best morphology when the molar ratio of NaCl to the raw materials is α = 4 (Figure 8). The influence of the ratio of reactants on the product was also investigated and the condition of Na:Fe:Ti = 3.3:2.2:1 shows the pure phase Na0.75Fe0.75Ti0.25O2 with relatively uniform morphology ( Figure S7). Figures S8, and S9 show the products prepared at different reaction temperatures (600, 700, 800, 900, and 1000 °C) and different reaction durations (2, 4, and 6 h). NaFeTiO4 needles with best morphology were obtained at 900 °C for 4 h and Na0.75Fe0.75Ti0.25O2 platelets were obtained at 1000 °C for 4 h.    Figures 6 and 7 show the XRD pattern and SEM images of the NFTO samples synthesized while using NaCl as the molten salt. At the condition of Na:Fe:Ti = 1.3:1:1 and Na:Fe:Ti = 3.3:2.2:1, the XRD patterns of the products can be assigned to NaFeTiO4 (PDF#33-1255, Figure 6a) and Na0.75Fe0.75Ti0.25O2 (PDF#25-0877, Figure 6b), respectively. None of the noticeable peaks belong to the unreacted reactants (Na2CO3) or intermediate phase (Fe2O3), indicating that the starting materials have been completely transformed to products at appropriate annealing temperatures (NaFeTiO4, 900 °C; Na0.75Fe0.75Ti0.25O2, 1000 °C). The chemical composition of products is also confirmed by EDS analyses ( Figure S6). As shown in Figure 7, NaFeTiO4 is in the shape of needles with the length of 20-50 μm and diameter of 0.5-2 μm, while Na0.75Fe0.75Ti0.25O2 has the platelet shape with the size range of 5-20 μm. The products show the best morphology when the molar ratio of NaCl to the raw materials is α = 4 (Figure 8). The influence of the ratio of reactants on the product was also investigated and the condition of Na:Fe:Ti = 3.3:2.2:1 shows the pure phase Na0.75Fe0.75Ti0.25O2 with relatively uniform morphology ( Figure S7). Figures S8, and S9 show the products prepared at different reaction temperatures (600, 700, 800, 900, and 1000 °C) and different reaction durations (2, 4, and 6 h). NaFeTiO4 needles with best morphology were obtained at 900 °C for 4 h and Na0.75Fe0.75Ti0.25O2 platelets were obtained at 1000 °C for 4 h.   . SEM images of (a1-a3) NaFeTiO4 needles and (b1-b3) Na0.75Fe0.75Ti0.25O2 platelets while using NaCl alone as the molten salt. (a1, b1) α = 2, (a2, b2) α = 4, and (a3, b3) α = 6; β = 0.

Synthesis of NFTO with Different Morphologies
The growth mechanism of NFTO is proposed as depicted in Figure 9, based on the analyses using SEM ( Figure S8), XRD (Figure S10), and TG-DSC ( Figure S11). As the calcination temperature increases, free water in the raw materials gets released and Na2CO3 decomposes to Na2O and CO2 below 700 °C. After FeTiO3 is completely converted to Fe2O3 and Fe2Ti3O9 around 620 °C, the reaction system changes from Na2CO3-FeTiO3 to Na2O-Fe2O3-Fe2Ti3O9. The NaFeTiO4 phase starts to appear after 700 °C. The Fe2O3 and Fe2Ti3O9 were completely consumed at 900 °C. The product obtained at 900 °C exhibits the best needle-like morphology and a relatively narrow size dispersion. The average diameter and length of the as-prepared NaFeTiO4 needles are in the range of 0.5-2 μm and 20-50 μm, respectively. At 1000 °C, part of NaFeTiO4 starts to break into small pieces. Fe2O3 was added in the starting materials to increase the Fe content. NaFeTiO4 with low crystallinity forms at 700 °C. With sufficient Fe source, Na0.75Fe0.75Ti0.25O2 platelets start to appear at 800 °C. Thus, Na0.75Fe0.75Ti0.25O2 platelets and a small amount of NaFeTiO4 rods are both present in products from 800-900 °C. At 1000 °C, NaFeTiO4 phase is disappeared and pure phase Na0.75Fe0.75Ti0.25O2 platelets are obtained.
To investigate the influence of molten salt type on the growth of NFTO, KCl-NaCl composite molten salt with different ratios were used for reactions. Figures 10 and 11 show the XRD patterns and SEM images while using KCl-NaCl as the composite molten salt and α is fixed at 4. When β is below 0.5, only the NaFeTiO4 phase is detectable. At β = 0.25, the product has both rod-like and platelet shapes. At β = 0.5, the rods are apparently larger in size, accompanied with randomly shaped particles. While β is 0.75, the product is a mixture of NaFeTiO4 and K2.3Fe2.3Ti5.7O16 and the product contains both big rods and random particles. When the KCl content reaches 100% (β = 1), pure phase K2.3Fe2.3Ti5.7O16 is observed. The morphology changes to a mixture of large plates and small particles. Hence, the results indicate that KCl in the molten salt can participate in the crystal growth of NFTO and should be avoided for obtaining pure phase NFTO. The growth mechanism of NFTO is proposed as depicted in Figure 9, based on the analyses using SEM ( Figure S8), XRD (Figure S10), and TG-DSC ( Figure S11). As the calcination temperature increases, free water in the raw materials gets released and Na 2 CO 3 decomposes to Na 2 O and CO 2 below 700 • C. After FeTiO 3 is completely converted to Fe 2 O 3 and Fe 2 Ti 3 O 9 around 620 • C, the reaction system changes from Na 2 CO 3 -FeTiO 3 to Na 2 O-Fe 2 O 3 -Fe 2 Ti 3 O 9 . The NaFeTiO 4 phase starts to appear after 700 • C. The Fe 2 O 3 and Fe 2 Ti 3 O 9 were completely consumed at 900 • C. The product obtained at 900 • C exhibits the best needle-like morphology and a relatively narrow size dispersion. The average diameter and length of the as-prepared NaFeTiO 4 needles are in the range of 0.5-2 µm and 20-50 µm, respectively. At 1000 • C, part of NaFeTiO 4 starts to break into small pieces. To investigate the influence of molten salt type on the growth of NFTO, KCl-NaCl composite molten salt with different ratios were used for reactions. Figures 10 and 11 show the XRD patterns and SEM images while using KCl-NaCl as the composite molten salt and α is fixed at 4. When β is below 0.5, only the NaFeTiO 4 phase is detectable. At β = 0.25, the product has both rod-like and platelet shapes. At β = 0.5, the rods are apparently larger in size, accompanied with randomly shaped particles. While β is 0.75, the product is a mixture of NaFeTiO 4 and K 2.