Controllable Hydrothermal Conversion from Ni-co-mn Carbonate Nanoparticles to Microspheres

Starting from Ni-Co-Mn carbonate nanoparticles prepared by microreaction technology, uniform spherical particles of Ni 1/3 Co 1/3 Mn 1/3 CO 3 with a size of 3–4 µm were obtained by a controllable hydrothermal conversion with the addition of (NH 4) 2 CO 3. Based on characterizations on the evolution of morphology and composition with hydrothermal treatment time, we clarified the mechanism of this novel method as a dissolution-recrystallization process, as well as the effects of (NH 4) 2 CO 3 concentration on the morphology and composition of particles. By changing concentrations and the ratio of the starting materials for nano-precipitation preparation, we achieved monotonic regulation on the size of the spherical particles, and the synthesis of Ni 0.4 Co 0.2 Mn 0.4 CO 3 and Ni 0.5 Co 0.2 Mn 0.3 CO 3 , respectively. In addition, the spherical particles with a core-shell structure were preliminarily verified to be available by introducing nano-precipitates with different compositions in the hydrothermal treatment in sequence.


Introduction
LiCoO 2 has been widely used as a positive electrode material in commercial lithium-ion batteries because of its high capacity, as well as excellent stability [1]. However, cobalt also causes serious problems, such as high price and environmental concerns. Alternatively, a promising material is Li[Ni x Co y Mn 1−x−y ]O 2 with a layered structure, which is considered to be one of the best replacements for LiCoO 2 for hybrid electric vehicle (HEV) power source systems [2][3][4][5][6][7][8]. Since this material has combined nickel, cobalt, and manganese together, it may show the advantages of these three metals in terms of thermal stability, rate capability, and safety at a proper composition. In detail, introducing Co can increase the stability of the structure and suppress the cation mixing, while too much Co causes capacity loss; increasing the amount of Ni will be benefit the capacity of the material, but too much Ni leads to the cation mixing, which decreases the cycling stability; a proper amount of Mn can improve the safety, but too much Mn can change the structure from layered to spinel [9].
In general, researchers usually prepared NiCoMn (NCM) precursor (hydroxide or carbonate) and then combined it with lithium salt. Many studies have shown that the electrochemical performance of the final product strongly depends on the properties of the precursor, such as morphology [10][11][12][13], size of the primary and secondary particles [13][14][15][16], size distribution [17], composition [6,8,9,[18][19][20][21], as well as structure [22,23]. Therefore, it is vital to control the precursor in order to obtain the materials with excellent performance, for which a simple and controlled preparation method is highly required. Until now, many methods have been developed, including co-precipitation [4,11], the sol-gel method [24,25], the spray-dry method [5,18], the solid-state reaction [26], and others [27][28][29][30]. Among these methods, the most popular one is co-precipitation since it is relatively simple to implement [4,31].  Figure 2 shows the XRD patterns of Ni1/3Co1/3Mn1/3CO3 before and after hydrothermal treatment. Obviously, these two patterns are totally different, indicating that the precipitated sample is amorphous and, upon hydrothermal treatment, a crystalline material is obtained. Additionally, the pattern of the precipitates after hydrothermal is quite consistent with ideal MnCO3 and also shows broad integrated lines which can be attributed to the mix of NiCO3, CoCO3, and MnCO3. Figure 3 is the mapping of a single particle presenting the elemental distribution. The yellow, red, and blue dots represent the distributions of Ni, Co, and Mn, respectively. The brigthness of the color reflects the intensity of the element signal. Uniform brightness distribution reflects the composition distribution of the spherical precipates is uniform. According to these results, we could confirm that it is feasible to prepare spherical carbonate precursor at micron size with the same element ratio as the starting materials by the method combining microreaction technolgy with hydrothermal treatment.   Figure 2 shows the XRD patterns of Ni 1/3 Co 1/3 Mn 1/3 CO 3 before and after hydrothermal treatment. Obviously, these two patterns are totally different, indicating that the precipitated sample is amorphous and, upon hydrothermal treatment, a crystalline material is obtained. Additionally, the pattern of the precipitates after hydrothermal is quite consistent with ideal MnCO 3 and also shows broad integrated lines which can be attributed to the mix of NiCO 3 , CoCO 3 , and MnCO 3 . Figure 3 is the mapping of a single particle presenting the elemental distribution. The yellow, red, and blue dots represent the distributions of Ni, Co, and Mn, respectively. The brigthness of the color reflects the intensity of the element signal. Uniform brightness distribution reflects the composition distribution of the spherical precipates is uniform. According to these results, we could confirm that it is feasible to prepare spherical carbonate precursor at micron size with the same element ratio as the starting materials by the method combining microreaction technolgy with hydrothermal treatment.  Figure 2 shows the XRD patterns of Ni1/3Co1/3Mn1/3CO3 before and after hydrothermal treatment. Obviously, these two patterns are totally different, indicating that the precipitated sample is amorphous and, upon hydrothermal treatment, a crystalline material is obtained. Additionally, the pattern of the precipitates after hydrothermal is quite consistent with ideal MnCO3 and also shows broad integrated lines which can be attributed to the mix of NiCO3, CoCO3, and MnCO3. Figure 3 is the mapping of a single particle presenting the elemental distribution. The yellow, red, and blue dots represent the distributions of Ni, Co, and Mn, respectively. The brigthness of the color reflects the intensity of the element signal. Uniform brightness distribution reflects the composition distribution of the spherical precipates is uniform. According to these results, we could confirm that it is feasible to prepare spherical carbonate precursor at micron size with the same element ratio as the starting materials by the method combining microreaction technolgy with hydrothermal treatment.

Synthesis Mechanism
Hydrothermal treament commonly leads to a complex dissolution-recrystallization process determined by many factors, such as temperature, composition of solution, reaction time, and so on. In order to understand the details of reaction mechanism, we attempted to investigate the effects of these factors separately. Figure 4 shows the SEM images of products after hydrothermal treatment with different time. Evidently, the evolution of the morphology of the products is significant. The products after 24 h hydrothermal treatment ( Figure 4d) have relatively uniform size and very smooth surfaces. Comparatively, for the products after 3 h, 6 h, or 9 h hydrothermal treament, there exist many large particles composed of irregular blocks. The energy dispersive spectroscopic (EDS) (Zeiss, Oberkochen, Germany) characterization shows that the main components of these blocks are Ni and Co. The general tendency is that the size distribution of particles becomes uniform and these large particles dimish gradually with the proceeding of hydrothermal treatment. We also noticed that the solution above the products is always purple, whaterever the hydrothermal treatment time is 3 h, 6 h, or 9 h. However, it becomes almost colorless after 24 h hydrothermal treatment. The color of the solution indicates the existence of metal ions in the solution. These phenomena imply that the evolution of morphology surely carries out via a dissolution process. Since the blocks settled on the surface of spherical particles present a different appearance or size compared with the precipitates and the microblocks composing spherical particles, they might come from the cooling crystallization of metal carbonates from solution. Thus, the small content of Mn in the blocks may be determined the small content of Mn in the solution. Correspondingly, according to the stability contants of different metal ammonia complexes, the order of the metal contents in solution is just Ni > Co > Mn. However, monitoring the evolutions of pH and the ions constributions during hydrothermal treatment could help us understand the mechanism in depth and is worth further investigation. Figure 5 shows the effects of the concentration of (NH4)2CO3 in hydrothermal treatment. As seen, without the addition of (NH4)2CO3, the products are mainly of irregular bulk precipitates with rough surfaces. When 0.05 mol·L −1 (NH4)2CO3 was added, there are some rough spheres surrounded by many blocks (Figure 5b), and the components of these blocks are also mainly Ni and Co. As increasing the concentration of (NH4)2CO3 to 0.10 mol·L −1 , these blocks become smaller, but still larger than 100 nm. Figure 6 shows SEM mapping photographs of Ni, Co, and Mn corresponding to the products in Figure 5c. The signals of Ni and Co are very intensive in the surface layer of the spherical particle, and in the core the distributions of Ni, Co, and Mn are much uniform. It could be reasonably assumed that the environments for generating the core (smooth spheres) and the surface layer (irregular

Synthesis Mechanism
Hydrothermal treament commonly leads to a complex dissolution-recrystallization process determined by many factors, such as temperature, composition of solution, reaction time, and so on. In order to understand the details of reaction mechanism, we attempted to investigate the effects of these factors separately. Figure 4 shows the SEM images of products after hydrothermal treatment with different time. Evidently, the evolution of the morphology of the products is significant. The products after 24 h hydrothermal treatment ( Figure 4d) have relatively uniform size and very smooth surfaces. Comparatively, for the products after 3 h, 6 h, or 9 h hydrothermal treament, there exist many large particles composed of irregular blocks. The energy dispersive spectroscopic (EDS) (Zeiss, Oberkochen, Germany) characterization shows that the main components of these blocks are Ni and Co. The general tendency is that the size distribution of particles becomes uniform and these large particles dimish gradually with the proceeding of hydrothermal treatment. We also noticed that the solution above the products is always purple, whaterever the hydrothermal treatment time is 3 h, 6 h, or 9 h. However, it becomes almost colorless after 24 h hydrothermal treatment. The color of the solution indicates the existence of metal ions in the solution. These phenomena imply that the evolution of morphology surely carries out via a dissolution process. Since the blocks settled on the surface of spherical particles present a different appearance or size compared with the precipitates and the microblocks composing spherical particles, they might come from the cooling crystallization of metal carbonates from solution. Thus, the small content of Mn in the blocks may be determined the small content of Mn in the solution. Correspondingly, according to the stability contants of different metal ammonia complexes, the order of the metal contents in solution is just Ni > Co > Mn. However, monitoring the evolutions of pH and the ions constributions during hydrothermal treatment could help us understand the mechanism in depth and is worth further investigation. Figure 5 shows the effects of the concentration of (NH 4 ) 2 CO 3 in hydrothermal treatment. As seen, without the addition of (NH 4 ) 2 CO 3 , the products are mainly of irregular bulk precipitates with rough surfaces. When 0.05 mol·L −1 (NH 4 ) 2 CO 3 was added, there are some rough spheres surrounded by many blocks (Figure 5b), and the components of these blocks are also mainly Ni and Co. As increasing the concentration of (NH 4 ) 2 CO 3 to 0.10 mol·L −1 , these blocks become smaller, but still larger than 100 nm. Figure 6 shows SEM mapping photographs of Ni, Co, and Mn corresponding to the products in Figure 5c. The signals of Ni and Co are very intensive in the surface layer of the spherical particle, and in the core the distributions of Ni, Co, and Mn are much uniform. It could be reasonably assumed that the environments for generating the core (smooth spheres) and the surface layer (irregular blocks) are quite different. By adding 0.15 mol·L −1 of (NH 4 ) 2 CO 3 , the blocks almost dissapeared and we fortunately obtained uniform and smooth spherical particles with average size of 3-4 µm and the primary particle is less than 100 nm. However, when the concentration of (NH 4 ) 2 CO 3 was increased to 0.25 mol·L −1 and 0.3 mol·L −1 , and the particles are still smooth, but their size distributions become wider. blocks) are quite different. By adding 0.15 mol·L −1 of (NH4)2CO3, the blocks almost dissapeared and we fortunately obtained uniform and smooth spherical particles with average size of 3-4 μm and the primary particle is less than 100 nm. However, when the concentration of (NH4)2CO3 was increased to 0.25 mol·L −1 and 0.3 mol·L −1 , and the particles are still smooth, but their size distributions become wider.   blocks) are quite different. By adding 0.15 mol·L −1 of (NH4)2CO3, the blocks almost dissapeared and we fortunately obtained uniform and smooth spherical particles with average size of 3-4 μm and the primary particle is less than 100 nm. However, when the concentration of (NH4)2CO3 was increased to 0.25 mol·L −1 and 0.3 mol·L −1 , and the particles are still smooth, but their size distributions become wider.    Herein, we proposed a schematic mechanism, as shown in Figure 7, to explain the experimental results based on following three facts: (1) the primary nano-precipitates could partly dissolve in solution; (2) the spherical particles are composed of nanocrystals as the secondary precipitates with crystalline structure different from the primary nano-precipitates; (3) (NH4)2CO3 has remarkable influences on both the dissolution and recrystallization processes. We suppose that the nanocrystals have good thermodynamics stability compared with the primary nano-precipiates and the (NH4)2CO3 in solution could dissociate to release NH3 to accelearte the conversion between them [4,38,44]. The conversion process includes following steps: (1) the primary nano-precipitates dissolve into solution partly to increase the contents of metals in solution to relatively high levels; (2) the nuclei of secondary precipitates generate and grow up to nanocrystals in solution, during which the dissolution of the primary nano-precipitates continue to proceed; (3) the nanocrytals gradually aggregate to lead to the generation and growth of spherical particles in solution; and (4) as all the primary precipitates are consumed, the contents of metals in solution will decrease until the recrystalization process terminates due to the limitation on thermodynamics equilibrium.
The addtion of (NH4)2CO3 could increase the contents of metals in solution, as well as the conversion rate from primary nanoprecipitates to nanocrystals. Due to the difference of Ni, Co, and Mn in complexation ability with NH3, the contents of Ni and Co are much higher than that of Mn. As (NH4)2CO3 is in absence, the conversion is so slow that after 24 h hydrothermal treatment plenty of nanoprecipitates still exist and spherical particles are seldom seen. As (NH4)2CO3 being at low concentration, after 24 h hydrothermal treatment part of conversion could be achieved to generate spherical particles. However, the contents of metals in solution (Ni and Co take the majority) may be still at high levels, which could separated out as MCO3 in the cooling period before sampling. As the concentration of (NH4)2CO3 being high enough, 24 h hydrothermal treatment can deplete the primary precipitates. Meanwhile, with the increase of the concentration of (NH4)2CO3, more nuclei could generate in solution, which will inhibit the growth and aggregation of nanocrystals to increase the size of nanocrystals and the size of their aggregation; the conversion rate will increase to make the growth and aggregation of nanocrystals easy to get out of control, and it will broaden the size distribution of the final products. Nevertheless, the control on the growth and aggregation of nanocrystals, as well as the optimization of (NH4)2CO3 concentration, are worthy of further investigations. Herein, we proposed a schematic mechanism, as shown in Figure 7, to explain the experimental results based on following three facts: (1) the primary nano-precipitates could partly dissolve in solution; (2) the spherical particles are composed of nanocrystals as the secondary precipitates with crystalline structure different from the primary nano-precipitates; (3) (NH 4 ) 2 CO 3 has remarkable influences on both the dissolution and recrystallization processes. We suppose that the nanocrystals have good thermodynamics stability compared with the primary nano-precipiates and the (NH 4 ) 2 CO 3 in solution could dissociate to release NH 3 to accelearte the conversion between them [4,38,44]. The conversion process includes following steps: (1) the primary nano-precipitates dissolve into solution partly to increase the contents of metals in solution to relatively high levels; (2) the nuclei of secondary precipitates generate and grow up to nanocrystals in solution, during which the dissolution of the primary nano-precipitates continue to proceed; (3) the nanocrytals gradually aggregate to lead to the generation and growth of spherical particles in solution; and (4) as all the primary precipitates are consumed, the contents of metals in solution will decrease until the recrystalization process terminates due to the limitation on thermodynamics equilibrium.
The addtion of (NH 4 ) 2 CO 3 could increase the contents of metals in solution, as well as the conversion rate from primary nanoprecipitates to nanocrystals. Due to the difference of Ni, Co, and Mn in complexation ability with NH 3 , the contents of Ni and Co are much higher than that of Mn. As (NH 4 ) 2 CO 3 is in absence, the conversion is so slow that after 24 h hydrothermal treatment plenty of nanoprecipitates still exist and spherical particles are seldom seen. As (NH 4 ) 2 CO 3 being at low concentration, after 24 h hydrothermal treatment part of conversion could be achieved to generate spherical particles. However, the contents of metals in solution (Ni and Co take the majority) may be still at high levels, which could separated out as MCO 3 in the cooling period before sampling. As the concentration of (NH 4 ) 2 CO 3 being high enough, 24 h hydrothermal treatment can deplete the primary precipitates. Meanwhile, with the increase of the concentration of (NH 4 ) 2 CO 3 , more nuclei could generate in solution, which will inhibit the growth and aggregation of nanocrystals to increase the size of nanocrystals and the size of their aggregation; the conversion rate will increase to make the growth and aggregation of nanocrystals easy to get out of control, and it will broaden the size distribution of the final products. Nevertheless, the control on the growth and aggregation of nanocrystals, as well as the optimization of (NH 4 ) 2 CO 3 concentration, are worthy of further investigations.

Regulations on the Size and Composition
According to the synthesis mechanism metioned above, if we added more primary nanoprecipitates into the autoclave, more metals will be provided for generating larger spherical particles. Therefore, we attempted to tune the size of the final products by changing the concentrations of the starting materials used for primary nanoprecipitate preparation. As seen in Figure 8, when the total concentration of metals (CM) is 0.15 mol·L −1 , the average size of final products is 3.39 μm. When we increased CM, the average size becomes 3.95 μm at CM = 0.30 mol·L −1 and 4.76 μm CM = 0.6 mol·L −1 . The size of the particles only increases with the increasing of CM, but it does abide by the proportional relation. A possible reason is that more nuclei may be generated in the initial period due to the acceleration of nanoprecipitate dissolution [45]. It can also explain why the size distribution of final product becomes wider with the increase in concentration.

Regulations on the Size and Composition
According to the synthesis mechanism metioned above, if we added more primary nanoprecipitates into the autoclave, more metals will be provided for generating larger spherical particles. Therefore, we attempted to tune the size of the final products by changing the concentrations of the starting materials used for primary nanoprecipitate preparation. As seen in Figure 8, when the total concentration of metals (C M ) is 0.15 mol·L −1 , the average size of final products is 3.39 µm.
When we increased C M , the average size becomes 3.95 µm at C M = 0.30 mol·L −1 and 4.76 µm C M = 0.6 mol·L −1 . The size of the particles only increases with the increasing of C M , but it does abide by the proportional relation. A possible reason is that more nuclei may be generated in the initial period due to the acceleration of nanoprecipitate dissolution [45]. It can also explain why the size distribution of final product becomes wider with the increase in concentration.

Regulations on the Size and Composition
According to the synthesis mechanism metioned above, if we added more primary nanoprecipitates into the autoclave, more metals will be provided for generating larger spherical particles. Therefore, we attempted to tune the size of the final products by changing the concentrations of the starting materials used for primary nanoprecipitate preparation. As seen in Figure 8, when the total concentration of metals (CM) is 0.15 mol·L −1 , the average size of final products is 3.39 μm. When we increased CM, the average size becomes 3.95 μm at CM = 0.30 mol·L −1 and 4.76 μm CM = 0.6 mol·L −1 . The size of the particles only increases with the increasing of CM, but it does abide by the proportional relation. A possible reason is that more nuclei may be generated in the initial period due to the acceleration of nanoprecipitate dissolution [45]. It can also explain why the size distribution of final product becomes wider with the increase in concentration.  As is well known, dissolution-recrystallization process can only change the morphology and structure of particles. The composition cannot be changed without the addition of other reageants. Inspired by this recognition, we also prepared precursors Ni 0.4 Co 0.2 Mn 0.4 CO 3 and Ni 0.5 Co 0.2 Mn 0.3 CO 3 by changing the ratio of metals in starting materials for primary nanoprecipitates preparation. The morphology of the products was characterized by SEM. As is seen in Figure 9, all three products are of spherical particles. Their compositions were determined by ICP, and are shown in Table 1. The amount of Ni is always slightly lower than the ratio of the starting materials, but, in general, the determined compositions agree well with that of the starting materials. The deviation of Ni content may be attributed to the strong complexation ability between Ni and NH 3 . As a result, the residual amount of Ni in the solution was the greatest, so the amount of Ni in the solid products is the least. As is well known, dissolution-recrystallization process can only change the morphology and structure of particles. The composition cannot be changed without the addition of other reageants. Inspired by this recognition, we also prepared precursors Ni0.4Co0.2Mn0.4CO3 and Ni0.5Co0.2Mn0.3CO3 by changing the ratio of metals in starting materials for primary nanoprecipitates preparation. The morphology of the products was characterized by SEM. As is seen in Figure 9, all three products are of spherical particles. Their compositions were determined by ICP, and are shown in Table 1. The amount of Ni is always slightly lower than the ratio of the starting materials, but, in general, the determined compositions agree well with that of the starting materials. The deviation of Ni content may be attributed to the strong complexation ability between Ni and NH3. As a result, the residual amount of Ni in the solution was the greatest, so the amount of Ni in the solid products is the least.  Reently, the cathode materials with core-shell structures have been drawing more attention [22,46,47], since the core-shell structures may restrict the formation of solid electrolyte interphase and volume expansion. Our method of combining nanoprecipitate dissolution with nanocrystal growth and aggregation may also be applied in preparing spherical particles with core-shell structures, as we introduce the nanoprecipitates with different compositions in hydrothermal treatments in sequence. Figure 10 shows the mapping of the obtained spherical particles with a core-shell structure. The synthesis procedures are illustrated in the context. Herein, the core is Ni1/3Co1/3Mn1/3CO3, and the shell is Co. The green image is the mapping of cobalt. We can see a light circle around the surface of the particle, which means the content of Co in the surface layer is higher than that in the core. Figure 11 shows the energy dispersive spectroscopic (EDS) of the core-shell structure. The line scan was done at the white line in Figure 1a. From the images, we can see that the Co element is in abundance at the surface. Therefore, we can confirm, preliminarily, that the core-shell structure has been obtained.  Reently, the cathode materials with core-shell structures have been drawing more attention [22,46,47], since the core-shell structures may restrict the formation of solid electrolyte interphase and volume expansion. Our method of combining nanoprecipitate dissolution with nanocrystal growth and aggregation may also be applied in preparing spherical particles with core-shell structures, as we introduce the nanoprecipitates with different compositions in hydrothermal treatments in sequence. Figure 10 shows the mapping of the obtained spherical particles with a core-shell structure. The synthesis procedures are illustrated in the context. Herein, the core is Ni 1/3 Co 1/3 Mn 1/3 CO 3 , and the shell is Co. The green image is the mapping of cobalt. We can see a light circle around the surface of the particle, which means the content of Co in the surface layer is higher than that in the core. Figure 11 shows the energy dispersive spectroscopic (EDS) of the core-shell structure. The line scan was done at the white line in Figure 1a. From the images, we can see that the Co element is in abundance at the surface. Therefore, we can confirm, preliminarily, that the core-shell structure has been obtained.
The precursor was synthesized by two steps as shown in Figure 12. The first step is to obtain homogeneous nanoparticles by using a home-made microreactor. The geometric size of the microchannel was 15 mm × 0.5 mm × 0.5 mm (length × width × height). A stainless steel membrane with an average pore diameter of 5 μm was used as the dispersion medium. We chose Ni1/3Co1/3Mn1/3CO3 as an example to describe the experimental procedures: (1) the dispersed fluid
The precursor was synthesized by two steps as shown in Figure 12. The first step is to obtain homogeneous nanoparticles by using a home-made microreactor. The geometric size of the microchannel was 15 mm × 0.5 mm × 0.5 mm (length × width × height). A stainless steel membrane with an average pore diameter of 5 μm was used as the dispersion medium. We chose Ni1/3Co1/3Mn1/3CO3 as an example to describe the experimental procedures: (1) the dispersed fluid Figure 11. The EDS images of the particles with a core-shell structure.  China). All of the chemicals were used without any further purification. Furthermore, the deionized water was used to make up solutions throughout the experiments.

All
The precursor was synthesized by two steps as shown in Figure 12. The first step is to obtain homogeneous nanoparticles by using a home-made microreactor. The geometric size of the microchannel was 15 mm × 0.5 mm × 0.5 mm (length × width × height). A stainless steel membrane with an average pore diameter of 5 µm was used as the dispersion medium. We chose Ni 1/3 Co 1/3 Mn 1/3 CO 3 as an example to describe the experimental procedures: (1) the dispersed fluid containing 0.1M CoSO 4 , 0.1M NiSO 4 , and 0.1M MnSO 4 was mixed with the continuous fluid containing 0.4M Na 2 CO 3 in the microreactor, and both of them were delivered by pumps at the flow rate of 40 mL/min; (2) the slurry containing primary nanoprecipitate Ni 1/3 Co 1/3 Mn 1/3 CO 3 was generated and transferred into a 100 mL Telfon-lined stainless steel autoclave; (3) 0.3 M (NH 4 ) 2 CO 3 at equal volume was added in the autoclave for hydrothermal treatment at 180 • C; (4) after 24 hours, the final precipitates were separated from the solution by centrifugation, washed with deionized water and ethanol several times, and dried in an oven at 120 • C overnight. As for the Ni 0. 5   As for the core-shell structure, the final products without drying were transferred into autoclave as seeds. Then the slurry containing primary nanoprecipitate CoCO3 as the shell material obtained from the microreactor and 0.3 M (NH4)2CO3 at equal volume were added into the autoclave for another hydrothermal treatment. The subsequent steps were the same. The samples used for EDS characterization were calcined at 500 °C in air for five hours.
The morphology of the prepared powders was observed by scanning electron microscope (SEM; JSM-7401, JEOL; HITACHI TM 3000, Tokyo, Japan). Element mapping and line scan were carried out by using a scanning electron microscope with an energy dispersive spectroscope (SEM; Merlin, ZEISS, Oberkochen, Germany). The phase of samples was characterized by X-ray diffraction (XRD; D8-Aduance, BRUKER, Karlsruhe, Germany) using Cu Kα radiation (40 kV and 40 mA) at a scanning rate of 5 o min −1 . The element composition of the synthesized product was determined by an inductively-coupled plasma-optical emission spectroscope (ICP-OES; IRIS Intrepid II XSP SPS, Thermofisher, Walham, MA, USA).

Conclusions
In our work, starting from the preparation of Ni-Co-Mn carbonate nanoparticles by using microreaction technology, we proposed a simple and novel method to realize the controllable hydrothermal conversion from Ni-Co-Mn carbonate nanoparticles to uniform and spherical particles of Ni1/3Co1/3Mn1/3CO3 with the assistance of ammonia carbonate. Based on the systematic characterizations on evolutions of the morphology and composition with hydrothermal treatment time, we clarified the mechanism for this novel method as a dissolution-recrystallization process, as As for the core-shell structure, the final products without drying were transferred into autoclave as seeds. Then the slurry containing primary nanoprecipitate CoCO 3 as the shell material obtained from the microreactor and 0.3 M (NH 4 ) 2 CO 3 at equal volume were added into the autoclave for another hydrothermal treatment. The subsequent steps were the same. The samples used for EDS characterization were calcined at 500 • C in air for five hours.
The morphology of the prepared powders was observed by scanning electron microscope (SEM; JSM-7401, JEOL; HITACHI TM 3000, Tokyo, Japan). Element mapping and line scan were carried out by using a scanning electron microscope with an energy dispersive spectroscope (SEM; Merlin, ZEISS, Oberkochen, Germany). The phase of samples was characterized by X-ray diffraction (XRD; D8-Aduance, BRUKER, Karlsruhe, Germany) using Cu Kα radiation (40 kV and 40 mA) at a scanning rate of 5 o min −1 . The element composition of the synthesized product was determined by an inductively-coupled plasma-optical emission spectroscope (ICP-OES; IRIS Intrepid II XSP SPS, Thermofisher, Walham, MA, USA).

Conclusions
In our work, starting from the preparation of Ni-Co-Mn carbonate nanoparticles by using microreaction technology, we proposed a simple and novel method to realize the controllable hydrothermal conversion from Ni-Co-Mn carbonate nanoparticles to uniform and spherical particles of Ni 1/3 Co 1/3 Mn 1/3 CO 3 with the assistance of ammonia carbonate. Based on the systematic characterizations on evolutions of the morphology and composition with hydrothermal treatment time, we clarified the mechanism for this novel method as a dissolution-recrystallization process, as well as the effects of (NH 4 ) 2 CO 3 concentration on the morphology and composition distribution. Furthermore, by changing the concentrations and the ratio of the starting materials for nano-precipitation preparation, we achieved monotonic regulation on the size of the spherical particles of Ni 1/3 Co 1/3 Mn 1/3 CO 3 , and the synthesis of Ni 0.4 Co 0.2 Mn 0.4 CO 3 and Ni 0.5 Co 0.2 Mn 0.3 CO 3 , respectively. In addition, we preliminarily verified that the spherical particles with core-shell structure were available by introducing nanoprecipitates with different compositions in the hydrothermal treatment in sequence. The potential of this method in applications could be expected for the preparation of spherical particles with specially-designed profiles of composition due to its convenience and adaptability.