Metastable Dissolution Regularity of Nd3+ in Na2CO3 Solution and Mechanism.

The carbonate solution-dissolved rare earth showed some metastable chemical characteristics. In this paper, the systematic investigation of metastable dissolution regularity of Nd3+ in Na2CO3 solution was carried out. The results showed that Nd3+ has an instantaneous saturated solubility in Na2CO3 solution. When the amount of the dissolution Nd3+ did not reach the instantaneous saturated solubility, the solution was in a stable-state period. Once the concentration of Nd3+ exceeded the instantaneous saturated solubility, the solution was no longer in the metastable state and generated the neodymium double salt of carbonate precipitates rapidly. The molecular dynamics simulation of the solution in the metastable state was carried out. In high concentration Na2CO3 solution, dissolved Nd3+ had a coordination reaction with the CO3 2-. Also, there was a stronger interaction between Na+ and CO3 2-, which caused the effective concentration of free CO3 2- which could react with Nd3+ to become lower. Thus, these reasons make the solution exhibit a metastable state. In that metastable period, the dissolved Nd3+ becomes steady and hard to generate the neodymium double salt of carbonate precipitates.

Introduction

A large amount of ammonia-nitrogen or high salinity wastewater is generated by using the traditional rare-earth separation methods, such as solvent extraction and ion exchange,[1,2] which has been causing trouble to the healthy development of the rare-earth industry. Some new rare-earth separation methods, such as the nonsaponification, saponified, coordination extraction, and ionic liquid extraction,[3−6] have some effects, but they are costly and difficult to resolve wastewater problems. Hence, it is necessary to develop a new, green, higher efficiency, and low-cost separation method for rare earth.

Rare-earth carbonate, the solubility of which is lower than 10–5 to 10–7 mol·L–1 in water, is a substance difficult to dissolve.[7,8] However, when the rare-earth salt solution was added at the high concentrated carbonate solution, the rare earth could be dissolved in the carbonate solution. The solubility of the rare earth did not decrease with the increase of the concentration of carbonate but increased. During the precipitation process of rare-earth carbonate in industry, the dissolution of rare earth caused the higher rare-earth residual concentration and brought about the unnecessary waste of rare-earth resource. However, on the other hand, the concentration of rare earth in carbonate solution increased regularly with the increase of the rare-earth atomic number. This may provide an idea for a new rare-earth separation method.

Since 1963, the ability of rare earth to dissolve into potassium carbonate solution was noted and investigated first by Taketatsu.[9,10] The rare-earth cations reacted with carbonate anions and generated rare-earth carbonate precipitates at first, but as the vibration and reaction time prolonged, it was dissolved again. In addition, the dissolution concentration of rare earth increased with the increase of potassium carbonate concentration and atomic number of the rare earth. Afterward, a series of studies confirmed the dissolution of rare earth obtained by Taketatsu, not only in potassium carbonate solution. For instance, De Vasconcellos et al.[11] chose ammonium carbonate solution to dissolve rare earth and obtained the same fruitage with Taketatsu. The dissolution amount of rare earth also increased with the increase of the concentration of ammonium carbonate and rare-earth atomic number. Nevertheless, from the viewpoint of the precipitation yield of rare earth, it could be clearly known that the yield of rare earth decreased with the increase of the dosage of the carbonate precipitator.[12−14] Thus, it was another piece of evidence that the dissolution amount of rare earth increased with the increase of the concentration of carbonate in solution.

Moreover, the ionic strength of the carbonate solution could also affect the dissolution process of the rare earth. Rao et al.[15] investigated the chemical balance behaviors of Nd3+ in the solution system of Na+–Cl––CO32––HCO3– and found that at the same concentration of sodium carbonate, the dissolution amount of rare earth was enhanced with the existence of impurity salt sodium chloride, compared to without sodium chloride. The higher the concentration of sodium chloride, the more is the amount of rare earth dissolved. De Vasconcellos et al.[16] also noted that the higher concentration of impurity ammonium anions increased the dissolution amount of rare earth. In addition, according to Tang et al.’s result from the research on rare-earth adsorption in water-bearing sand layer,[17] it could be found that increased ionic strength of the solution could cause weakening of the adsorption behavior of rare earth on the sand layer because of the increase of the dissolution amount of rare earth. Therefore, the ionic strength of the solution was another important determinant in dissolving rare earth.[18]

Nevertheless, we discovered that the sodium carbonate solution which dissolved rare earth exhibited some unique metastable properties. It had a metastable state period, in which rare earth could steadily be dissolved into a carbonate solution, and did not generate the precipitates of rare-earth carbonate. However, over the period, the precipitate of rare-earth carbonate still formed. Nevertheless, the period could be artificially controlled by changing the ionic strength of carbonate solution via adjusting the condition of the circumstance of the solution or enhancing the concentration of carbonate, which could affect the dissolution process of rare earth.

Inspired by the differential solubility in carbonate solution of different rare-earth elements previously introduced, the artificial domination of the metastable state period may have the potential application and was feasibly utilized in rare-earth separation.

However, existing reports of the dissolution of rare earth in carbonate solution are mostly aimed at the chemical balance and rarely involved in rare-earth separation. As for the metastable state, carbonate solution which dissolved rare earth was basically lacking. To investigate the regularity of the metastable state and provide a reference for the development of new rare-earth separation methods, in this paper, the neodymium was selected as the representative for rare-earth elements and metastable dissolution regularity of Nd3+ in Na2CO3 solution was systematically investigated. The research on the interaction between various ions in solution was performed using molecular dynamics (MD) simulation.

Results and Discussions

Metastable Dissolution Regularity of Nd3+ in Na2CO3 Solution

Determination of Instantaneous Saturated Solubility of Nd3+

The results showed that when the NdCl3 solution was added into Na2CO3 solution, flocculent precipitates were formed first and were then dissolved again with vigorous shaking. At this time, the solution was in a clean state. It means that the Nd3+ completely dissolved in Na2CO3 solution. From Figure , we can see that there existed the maximum solubility point of Nd3+ in Na2CO3 solution.

Figure 1

Instantaneous saturated solubility of Nd3+

However, when the additional amount of Nd3+ exceeded the maximum solubility point, the solution was no longer clear, and a part of Nd3+ dissolved in the solution began to precipitate. Moreover, we also observed that the solubility of Nd3+ in Na2CO3 solution was not stable, and it would decrease within a few minutes. As shown in Figure , the maximum solubilities of Nd3+ in Na2CO3 solutions of 1, 1.5, and 2 mol·L–1 are 1.186, 2.394, and 3.566 g·L–1, respectively.

The generated precipitate was confirmed to be NaNd(CO3)2 (see Section and the FTIR result in Figure ). According to the solubility product of NaNd(CO3)2 (log Ksp was −21.39),[15] for the reactionunder the experimental condition, the concentration of sodium carbonate was 1, 1.5, and 2 mol·L–1 and the activity coefficient of sodium carbonate was calculated to be 0.236, 0.205, and 0.188 via Pitzer theory,[19] respectively. Then, the corresponding concentration of Nd3+ in equilibrium was evaluated, and it should be 1.563 × 10–19, 1.911 × 10–19, and 2.450 × 10–19 g·L–1. Compared to the solubility we had measured, the deviation was quite huge. Obviously, it was not the equilibrium concentration from the dissolution process of NaNd(CO3)2. Hence, the maximum solubility point was defined as instantaneous saturated solubility.

Figure 6

FTIR results of solid precipitates of neodymium (a) saturated and (b) unsaturated.

Furthermore, from the fitted result of the instantaneous saturated solubility of Nd3+ with Na2CO3 concentration in Figure , we could see that the solubility of Nd3+ increased with the increase of Na2CO3 concentration in the solution, and it approached a significant linear relationship y = −1.18877 + 2.38043x because the R2 was 0.99992.

Figure 2

Instantaneous saturated solubility variance with the concentration of CO32–.

Dissolution of Nd3+ with Time

In order to evaluate the dissolution of Nd3+ with time in Na2CO3 solution, two experiments were carried out. In the first one, the initial dissolution concentration of Nd3+ in Na2CO3 solution was controlled to be equal to the instantaneous saturated solubility of Nd3+, which was 1.186, 2.394, and 3.566 g·L–1 in 1, 1.5, and 2 mol·L–1 Na2CO3 solution, respectively (the experiments are named “Nd3+ saturated”). In the other one, the initial dissolution concentration of Nd3+ was controlled to be lower than the solubility, which was 0.734, 1.362, and 2.621 g·L–1 in 1, 1.5, and 2 mol·L–1 Na2CO3 solution, respectively (the experiments are named “Nd3+ unsaturated”).

As for Nd3+ saturated dissolved in Na2CO3 solution, in the beginning, all the Na2CO3 solution dissolved Nd3+ was in a clarification state, but the clarification period was truly short. The solution precipitated gradually at the bottom of the conical bottle as the standing time went on. From Figure , all the results illustrate that the Nd3+ concentration of Na2CO3 solution of various concentrations showed a downward trend in the former 0–10 min stage. In addition, the decrease of the concentration of Nd3+ was slightly slower in the first 0–120 min but became faster later. The reason may be that the precipitation and crystallization need a crystal core, and the core was under the forming and growing process initially. Later, the precipitation became faster because abundant cores had been generated in the static process.

Figure 3

Time dependence of the concentration of Nd3+ in the saturated solution.

The consequence concluded that the Nd3+ saturated dissolved in Na2CO3 solution was not stable. However, it could also be admitted that the solution was basically not in the metastable state. Hence, the situation under this condition ought not to be investigated in the following studies.

As for the experiment of Nd3+ unsaturated dissolved in Na2CO3 solution, the solution exhibited some metastable characteristics. The solution was clean, and the Nd3+ was stable dissolved and did not precipitate with a period. As shown in Figure , in the first 90 min, the solution was in the period of the metastable state and the dissolution concentration of Nd3+ was almost unchanged.

Figure 4

Time dependence of the concentration of Nd3+ in the unsaturated solution.

However, this metastable state still had a time limit that is called the metastable period. The solution began to precipitate slowly after the metastable period. For example, when the aging time was 120 min, the concentration of Nd3+ in 1, 1.5, and 2 mol·L–1 Na2CO3 solution was 0.708, 1.329, and 2.610 g·L–1, respectively. Compared to the initial concentration, it was decreased slightly by only 3.47, 2.43, and 0.42%, but it was still a reduction, which indicates that in the solution over the metastable period, the self-precipitation occurred.

In addition, with the prolongation of the settling time, the concentration of Nd3+ in the solution reduced increasingly. For instance, when the aging time reached 360 min, the corresponding concentration of Nd3+ in 1, 1.5, and 2 mol·L–1 Na2CO3 solution was 0.580, 1.220, and 2.501 g·L–1, respectively. The decrease of the concentration of Nd3+ was significant, and the reduction ratios were about 20.92, 10.43, and 4.58% approximately than the initial.

Compared with the previous instantaneous saturated data (see in Figure ), it can be viewed in Figure that when the dissolution Nd3+ did not reach the corresponding instantaneous saturated solubility, the solutions with various concentrations of Na2CO3 all possessed a metastable period. At the same time, with the higher concentration of Na2CO3 solution, a longer metastable period was obtained. The solution metastable period in 1 mol·L–1 Na2CO3 solution was 90 min, but in 1.5 and 2 mol·L–1, the metastable period was 120 min. Moreover, the results could suggest that a higher concentration of Na2CO3 solution makes the dissolution concentration of Nd3+ become higher. As shown in Figure , the dissolution concentration of Nd3+ in 2 mol·L–1 Na2CO3 solution could reach 2.621 g·L–1 in the metastable period, but in 1.5 mol·L–1, the dissolution concentration of Nd3+ in Na2CO3 solution was only 2.394 g·L–1.

Coordination Behavior of Nd3+ in the Metastable Solution

The UV–vis full-wavelength scanning for the solution in a better metastable state, with 2.621 g·L–1 Nd3+ in 2 mol·L–1 Na2CO3 solution, which did not reach instantaneous saturated solubility, was carried out. To provide experimental contrast, the solution not in the metastable state, with 3.566 g·L–1 Nd3+ in 2 mol·L–1 Na2CO3 solution, which was equal to instantaneous saturated solubility, was also scanned. The results are presented in Figure .

Figure 5

UV–vis spectrum of the solution (a) saturated and (b) unsaturated.

From Figure , we could know whether the dissolution concentration of Nd3+ in Na2CO3 solution was equal to or did not reach instantaneous saturated solubility; there were two characteristic peaks of neodymium[20] found in the spectra of samples at 340–370 nm, which were located at 349 and 357 nm. It is worth noting that the characteristic peaks of neodymium, which was obtained from those Na2CO3 solutions that dissolved Nd3+, were slightly red-shifted from 347 and 354 to 349 and 357 nm than the blank NdCl3 solution. The reason was that the alkalinity of the solution dissolved Nd3+ was more than that of the blank NdCl3 solution.

Also, there were two characteristic peaks of neodymium[20] at the wavelengths 524 and 575 nm in all the samples. It should be noticed that the spectra of the samples of solution dissolved Nd3+ showed a new stronger characteristic peak at 583 nm, but it could not be found in the spectra of the blank sample. Thus, the dissolution Nd3+ in Na2CO3 solution has a significant coordination reaction with CO32–.

By referring to the results of Vercouter and Vitorge’s[21] studies on the equilibrium steady-state dissolution of rare earth in carbonate solution, we can know that Nd3+ could dissolve in high concentration Na2CO3 solution because the coordination reaction occurred between neodymium and CO32– forms of the complex ions like Nd(CO3)3. Therefore, it can be further guessed that the existence of a metastable state in Na2CO3 solution was probably due to the complex coordination between neodymium and CO32–, which causes the dissolution Nd3+ to not precipitate immediately.

Characterization of Precipitates

During the typical aging period, the precipitates formed from a metastable state, in which 2.621 g·L–1 Nd3+ (not reached instantaneously saturated) dissolved in 2 mol·L–1 Na2CO3 solution, were collected as samples and detected by Fourier transform infrared (FTIR). To provide experimental contrast, the precipitates, that were generated from the solution with 3.566 g·L–1 Nd3+ (was equaled instantaneously saturated) dissolved in 2 mol·L–1 Na2CO3 solution, were also collected and detected. The results are presented in Figure .

From Figure , the FTIR spectra of precipitates show that the characteristic infrared peaks were consistent with the solid phase of blank sample NaNd(CO3)2, regardless of whether the initial dissolution concentration of Nd3+ in the solution was equal to or did not reach instantaneous saturated solubility.

The results were also in agreement with the study of the equilibrium steady-state solubility of rare earth in carbonate solution by Rao[15,22] et al; that is, under the high concentration of CO32– solution environment, the stable solid phase in the solution was only the complex salt NaNd(CO3)2, and Nd2(CO3)3 almost did not exist.

Simulation of the Mechanism of the Metastable State

Establishment and Optimization of Metastable Solution Models

Although the UV–vis full-wavelength scanning results of the above solution samples can qualitatively explain the dissolution behavior of Nd3+ in Na2CO3 solution, and combined with the previous literature, it can also prove that Nd3+ in metastable solution did coordinate with CO32–, the results cannot fully explain the essence of the existence of a metastable state.

To explore the mechanism of the metastable state, the simulation processes of the solution in a better metastable state, with 2.621 g·L–1 Nd3+ in 2 mol·L–1 Na2CO3 solution, which did not reach instantaneous saturated solubility, were carried out by MD calculation using the software Materials Studio. The blank Na2CO3 solution only added water, and the solution that is not in the metastable state, with 3.566 g·L–1 Nd3+ in 2 mol·L–1 Na2CO3 solution, which was equal to instantaneous saturated solubility, was also simulated. All the solution establishment parameters are listed in Table .

Table 1

Solution Establishment Parameters

	the solution, dissolution Nd3+ saturated				the solution, dissolution Nd3+ unsaturated
	2 mol L–1 Na2CO3 ρ: 1.137 g L–1		the blank solution ρ: 1.116 g L–1		2 mol L–1 Na2CO3 ρ: 1.148 g L–1		the blank solution ρ: 1.132 g L–1
components	number	mass fraction (%)	number	mass fraction (%)	number	mass fraction (%)	number	mass fraction (%)
H2O	10 590	87.7	10 590	88.1	10 590	86.2	10 590	86.5
Na+	488	5.2	488	5.2	560	5.8	560	5.8
CO32–	244	6.7	244	6.8	280	7.6	280	7.6
Nd3+	4	0.3	0	0.0	3	0.2	0	0.0
Cl+	12	0.2	0	0.0	9	0.1	0	0.0

The established solution model after geometry optimization is shown in Figure .

Figure 7

Constructed solution model after geometry optimization (a) saturated, (b) blank solution relates to saturated solution (c) unsaturated, and (d) blank solution relates to unsaturated solution.

The energy changing of the solution model is shown in Figure . According to that, we could know that in the optimization process, the overall energy of each model decreased gradually with the increase of the optimization steps, there was no large-scale energy disturbance, and the energy tended to be stable at a low level in the end. This means that the optimized solution model was at the lowest local energy level, and the model could be considered to be in a relatively stable state.

Figure 8

Energy change of the modes during the geometry optimization (a) saturated and (b) unsaturated.

MD Calculation Process

The temperature changes of each solution model during the MD calculation are shown in Figure . The temperature of each solution model was raised first and then decreased steadily, but the temperature of each solution model was all stable at 298 K (±10%) at the end of calculation, and there was no significant disturbance. It should be focused that the credibility of the solution model dynamic is closely related to the corresponding termination temperature. The result is credible while the temperature of the model is within ±10% range of the initial at the end of the calculation. Hence, the results were credible.

Figure 9

Temperature changes of each solution model during the MD calculation (a) saturated and (b) unsaturated.

The solution models after the MD calculation are shown in Figure .

Figure 10

Solution models after the dynamic calculation (a) saturated, (b) blank solution relates to saturated solution, (c) unsaturated, and (d) blank solution relates to unsaturated solution.

As shown in Figure , the solution was homogeneous in general after dynamic calculation. However, in the local region, each component was not randomly and evenly distributed in the solution. The interaction between ions (molecules) result in varying degrees of agglomeration phenomenon at the local level.

Among them, no matter what the dissolution level (saturated or unsaturated) of Nd3+ in Na2CO3 solution is, the Nd3+ was all surrounded by CO32–, and it was coordinated by about 2–4 CO32–; the specific coordination situation is shown in Figure . This was consistent with the results obtained from the full-wavelength UV–vis scan of the solution sample before.

Figure 11

Coordination between neodymium and CO32– in the models (a) Nd(CO3)2–, (b) Nd(CO3)33–, and (c) Nd(CO3)45–.

It was also confirmed with Vercouter and Vitorge’s conclusion,[21] which was that rare-earth elements and CO32– coordination in high concentration CO32– solution and all kinds of complex ions in the form of Nd(CO3)3 (m ≥ 2) existed but in different proportions.

In addition, all the solution models, including corresponding blank solutions, showed local agglomeration of CO32– distribution and a large number of Na+ distributed around them. It could surmise that Na+ and CO32– did not dissociate completely in the circumstance of the high concentration Na2CO3 solution, which means that the free CO32– concentration in the solution was at a low level.

Radial Distribution Function and Coordination Behaviors

The coordination behavior of neodymium with CO32– in high concentration Na2CO3 solution could be visually expressed by the solution model after the MD calculation. To further quantify the interaction between components in the solution model at the micro level, for the main ion pairs in the solution, such as Nd3+–CO32–, Na+–CO32–, and Nd3+–Cl–, the analyses of radial distribution function (RDF) were conducted using the Forcite.

As shown in Figure , the peak positions of RDF of Nd3+–CO32– in the solution were almost the same, regardless of whether the dissolution concentration of Nd3+ was equal to the instantaneous saturated solubility or unsaturated, but there are still some strong differences. The first RDF peak of Nd3+–CO32– in saturated situation was stronger than that in the unsaturated situation. This phenomenon indicated that Nd3+ could bind more closely with CO32– at a closer microscopic distance.

Figure 12

RDF of Nd3+–CO32– ion pairs and coordination situation (a) saturated and (b) unsaturated.

However, in the overall chemical bond range (r < 2.6 Å),[23] the average coordination number of Nd3+ with CO32– at the solution with Nd3+ saturated was about 2.37. It was lower than the number 2.50, in which, the dissolution of Nd3+ in the solution was unsaturated. However, when Nd3+ saturated dissolved in the solution, some of the dissolution Nd3+ precipitated and form NaNd(CO3)2 in a short time, and the average coordination number of Nd3+ and CO32– in NaNd(CO3)2 is 2.0. Thus, the average coordination number of Nd3+ and CO32– was lower. Correspondingly, when the dissolution Nd3+ was unsaturated, the solution was in a better metastable state, and the dissolution Nd3+ can be stable in the solution for a long time; thus, the coordination number was slightly higher.

From Figure , it could be found that there was no interaction between Nd3+ and Cl– because the RDF strength was zero. In this case, no matter whether dissolution Nd3+ was saturated or not, there was no coordination behavior between Nd3+ and Cl–.

Figure 13

RDF of Nd3+–Cl– ion pairs and coordination situation (a) saturated and (b) unsaturated.

Previous solution models had been able to visualize the local agglomeration of CO32–. By analyzing the RDF and the average coordination of Na+–CO32– ion pairs, the essence of the metastable state could be further elaborated. As shown in Figure , there was an interaction between Na+ and CO32– in all solutions, and the corresponding intensity peaks that appear in the RDF spectra were basically the same, but the strength was also different. When the dissolution concentration Nd3+ in the solution was equal to instantaneous saturated solubility, the number of CO32– around Na+ was about 1.27, which is the same with the corresponding blank solution with only added water. However, when the dissolution concentration Nd3+ did not reach the instantaneous saturated solubility, the number of CO32– around Na+ increased to 1.30 and that corresponding to the blank solution with water increased to 1.33.

Figure 14

RDF of Na+–CO32– ion pairs and coordination situation (a) saturated and the corresponding blank solution and (b) unsaturated and the corresponding blank solution.

This phenomenon indicated that the dissociation of CO32– in the solution was strongly related to the excess water added to the solution. When dissolution concentration Nd3+ reaches instantaneous saturated solubility (3.566 g·L–1), the required volume of 10 g·L–1 NdCl3 was 14 mL, but the solution when unsaturated (2.621 g·L–1) was only 9 mL. The excess water diluted the high concentration Na2CO3 solution, resulting in more CO32– being dissociated into the solution. That is to say, the concentration of free CO32–, which could react with Nd3+ to generate precipitates in the solution, was further increased. The interaction between Na+ and CO32– in the corresponding blank solution we had calculated could also provide an evidence.

In other words, the total concentration of CO32– in the solution was slightly lower when the dissolution Nd3+ was saturated than when the dissolution Nd3+ was unsaturated, but the corresponding concentration of free CO32– that could react with Nd3+ to generate precipitates was higher and causes the solution to become unstable. Therefore, the higher concentration of CO32–, but the lower concentration of free CO32– in the Na2CO3 solution, was also one of the important reasons for the existence of the solution metastable period.

Conclusions

In this work, the instantaneous saturated solubility of Nd3+ in 1–2 mol·L–1 Na2CO3 solution, metastable dissolution regularity of Nd3+, and metastable mechanism were investigated in detail. The main conclusions are reproduced below.

First, instantaneous saturation solubility of Nd3+ was positively correlated with the concentration of Na2CO3 solution. In 1.0, 1.5, and 2.0 mol·L–1 Na2CO3 solution, the solubility was 1.186, 2.394, and 3.566 g·L–1, respectively.

Second, when the initial dissolution concentration of Nd3+ in the solution was equal to the instantaneous saturated solubility, the solution was not in the metastable period and the dissolved Nd3+ precipitated quickly. However, when the initial dissolution concentration of Nd3+ in the solution did not reach the instantaneous saturated solubility, the solution had a metastable period, and in that period, the dissolution of Nd3+ was stable, but it still had the precipitation behavior, while beyond the period. Moreover, the precipitants obtained were all in the form of NaNd(CO3)2.

And the last, the coordination reaction between Nd3+ and CO32– occurred, and all kinds of complex ions in the form Nd(CO3)3 (m ≥ 2) existed, and the interaction between Nd3+ and Cl– was not found in the solution. The higher concentration of CO32– in the solution, but the lower concentration of free CO32– that could react with neodymium to generate precipitate, was an important reason for the existence of the solution metastable state.

Experiments

Materials and Equipment

In the experiment, the rare-earth raw material was 1.3568 mol·L–1 high purity NdCl3 solution and is produced by rare-earth separating factory in Longnan, Jiangxi Province. The composition content of the high purity NdCl3 solution is shown in Table .

Table 2

Composition Content of the High Purity NdCl3 Solution

concentration of Nd3+ (mol L–1)	concentration of H+ (mol L–1)	specific gravity (g mL–1)	rare-earth impurities/REO (μg mL–1)
1.3568	<0.10	1.326	La2O3	CeO2	Pr6O11	Sm2O3	Eu2O3	Gd2O3	Tb2O3
non-rare-earth impurities (μg mL–1)			<100	<100	500	<100	<100	<100	<100
Fe2O3	SiO2	CaO	Dy2O3	Ho2O3	Er2O3	Tm2O3	Yb2O3	Lu2o3	Y2O3
<0.50	2.49	7.3	<100	<100	<100	<100	<100	<100	<100

NdCl3 feed solution (10 g·L–1 ) was prepared by diluting the high purity solution with deionized water. The analytical pure (AR) Na2CO3 was used for the preparation of 1, 1.5, and 2 mol·L–1 solution by using deionized water. The purity of other chemical reagents, such as HCl, ethylenediaminetetraacetic acid (EDTA), and so on, was also of AR grade. The information on experimental equipment is given in Table .

Table 3

Information on Equipment

equipment	type specification	manufactures
high-speed centrifuge	TGL16MS	Yancheng Anxin Experimental Instrument Co., Ltd.
computer server	IBM System X3850	International Business Machines Corporation
UV–visible spectrophotometer (UV–vis)	UV-55000PC	Shanghai yoke instrument Co., Ltd.
Fourier transform infrared spectrometer (FTIR)	ALPHA	Bruker Corporation
inductively coupled plasma–optical emission spectroscopy (ICP–OES)	ULTIMA2	HORIBA Jobin Yvon
Kang’s oscillator	KS	Changzhou Putian instrument manufacturing company

Na2CO3 solutions (1, 1.5, and 2 mol·L–1 ) were selected as a basal solution for dissolving Nd3+. Na2CO3 solution (25 mL) was put into a conical flask, and then, 10 g·L–1 NdCl3 solution was dropwise added into it by shaking the flask via an oscillator at 20–25 °C. First, the white precipitates of basic rare-earth carbonate were formed and then dissolved with vigorous shaking. However, the precipitates were generated again while the dissolution concentration of Nd3+ in Na2CO3 solution attained the instantaneous saturated solubility. At this time, the solution was muddy and the addition of NdCl3 solution was stopped, and then, a known amount of muddy solution was centrifuged with 6000 rpm for 5 min. After that, 5 mL of liquor was transferred and the carbonate system was destroyed completely by using diluted HCl. Subsequently, the dissolution concentration of Nd3+ in Na2CO3 solution was measured by EDTA titration or inductively coupled plasma–optical emission spectroscopy (ICP–OES).

In order to control whether the dissolution concentration of Nd3+ was equal to or lower than the instantaneous saturated solubility, a known volume of 10 g·L–1 NdCl3 solutions was dropwise added into Na2CO3 solution with shaking, at 20–25 °C. After adding dropwise, the static aging process was carried out and the time was set between 0 and 480 min. At the end of each aging time, a known amount of the solution was transferred and centrifuged to get the liquor; subsequently, 5 mL of liquor was acidized completely by diluted HCl. The dissolution concentration of Nd3+ in Na2CO3 solution was titrated with EDTA or analyzed by ICP–OES.

To monitor the coordination behavior of Nd3+ quickly in the metastable solution, aqueous samples in different aging time were collected and full-wavelength scanned by using UV–vis. The blank solution, such as single Na2CO3 and NdCl3 solution, was also scanned.

Characterization of Precipitates

To investigate the formation of precipitates of basic rare-earth carbonate generated from the metastable solution, precipitate samples in different aging times were collected and detected by FTIR. It should be noticed that the precipitates samples were prestored in deionized water to prevent them from decomposition. Before analyzing, the water was completely removed by sucking filtration, and the samples were analyzed immediately.

Simulation of the Mechanism of the Metastable State

Materials Studio 8.0 is powerful software for material simulation and modeling and is developed by Accelrys.[24] In this section, all the components of the metastable solution, such as H2O, CO32–, Nd3+, Na+, and Cl–, were established by the Visualizer and optimized under the universal force field by Forcite in software Materials Studio 8.0. After the establishment and optimization process of the components, the solution box model was built, and the well-optimized components were added into the box via an amorphous cell. The values of density and the concentration of each component were settled by following the actual metastable solution. Subsequently, the geometry optimization of the solution model was performed. Also, the blank solutions, which were the Na2CO3 solution with only added water, were also established and optimized. During the optimization step, it was necessary to check whether the energy of the system was converted smoothly to a lower energy state.

The MD calculation was conducted by Forcite after the optimization of the solution models. The calculation ensemble was NVT. The simulation temperature was 298 K, and the time was set at 20 ps and calculated once per 1 fs. The parameters of the initial heating rate and temperature control sets were selected at Random and Andersen. To save experimental time, the accuracy of calculation was chosen as a middle. At the end of the calculation, it was necessary to verify whether the result of the dynamic calculation is credible. When the temperature is stable at the range of 298 ± 29.8 K in the end and there is no large temperature disturbance, the calculation result is credible; otherwise, it needs to be recalculated.

RDF and Coordination Behaviors

The schematic diagram of the RDF[25] is presented in Figure . After the MD calculation, the RDF analysis was carried out. Utilizing the formula 1 to establish the relationship between the results of RDF analysis and the coordination behaviors, the average coordination number between each component was obtained. The entire simulation flowchart is shown in Figure .N(L) refers to the number of coordination atoms (molecules) in the 0–L spherical shell around the target atom. ρ refers to the number density of coordination atoms (molecules), and the value is the ratio of the number of atoms (molecules) to the volume of space. g(r) refers to the RDF value, and it indicates the probability of the occurrence of coordination atoms (molecules) within a certain distance. r refers to the cutoff radius.

Figure 15

Schematic diagram of the RDF.

Figure 16

Simulation flowchart.