Abstract
The phase equilibriums of the ternary system KCl-(NH2)2CO-H2O at 303.15 K, 323.15 K, 333.15 K, and 343.15 K were studied using the isothermal dissolution equilibrium method, in which the composition of equilibrium solid phase was determined by Schreinemaker’s wet residue method and X-ray diffraction (XRD) method. It was found that the ternary system is a simple cosaturated system, without the formation of neither double salt nor solid solution. Wilson and NRTL models were employed to correlate in the solubility data of the system at experimental temperatures. The maximum values of RAD and RMSD of the Wilson model were 2.18 × 10−2 and 0.83 and those of the NRTL model were 1.69 × 10−2 and 0.40, respectively. The two models were utilized to forecast solubility data at various temperatures, and the obtained outcomes were in line with the literature data. Based on the experimental solubility data at 343.15 K, the cooling crystallization process of the system was monitored online by focused beam reflectance measurement (FBRM) and particle video microscope (PVM). The crystal products were characterized by XRD, scanning electron microscope (SEM), and energy dispersive spectrometry (EDS). The results showed that the precipitation of (NH2)2CO occurred during the crystallization process, and this was followed by KCl. KCl was formed on the surface of (NH2)2CO crystal. The crystal was a simple mixture containing KCl and (NH2)2CO.
1. Introduction
Water-soluble fertilizer (WSF) has been intensively investigated in the field of agriculture because it can effectively reduce soil acidification and improve soil nutrient status and crop growth [1, 2]. Compared with traditional fertilizers, the WSF has three main advantages: (1) Free Collocation of Nutrients. The WSF can be used singly or in combination depending on the variety of crops and growth stages. (2) High Utilization Rate by Crops. The WSF can be absorbed more quickly by the leaves and roots of crops with water as the medium and does not cause any harm to the crops or land. (3) Convenient and Economical Application. The WSF was utilized in conjunction with both sprinkler and drip irrigation, allowing for precise and uniform distribution of nutrients across crop growth and development zones. This enhances nutrient utilization efficiency while reducing water and fertilizer consumption [3–5].
Typically, the methods used to produce WSF include physical mixing and chemical synthesis methods. The physical mixing method mainly mixes raw materials directly using a specific formula through mechanical methods such as crushing and stirring. The chemical synthesis method is to produce stable products through a series of specific chemical reactions such as dissolution, impurity removal, synthesis, and concentration and then separate final products through crystallization under a certain temperature, pH, and other control conditions. To form high-quality WSF, chemical synthesis is the mostly commonly used method, in which cocrystallization is the key step [6–8].
The commonly used raw materials for the production of WSF are N, P, and K including urea ((NH2)2CO), ammonium dihydrogen phosphate (NH4H2PO4), potassium dihydrogen phosphate (KH2PO4), potassium chloride (KCl), and ammonium chloride (NH4Cl) [9]. When KCl and (NH2)2CO are used as raw materials to produce WSF-containing N and K, the cocrystallization of the ternary system KCl-(NH2)2CO-H2O needs to be studied. The phase equilibrium and crystallization process of related systems are the necessary basis for the study of cocrystallization. Currently, several studies have reported the phase equilibrium relationship of the ternary system KCl-(NH2)2CO-H2O at 283.15 K [10], 313.15 K, and 353.15 K [11], but these phase equilibrium data are not enough to support the study of the cocrystallization process of the system from 303.15 K to 353.15 K. Therefore, the isothermal solution equilibrium method was applied to measure the solubility data of the ternary system KCl-(NH2)2CO-H2O at 303.15 K, 323.15 K, 333.15 K, and 343.15 K in the present study.
The calculation and prediction of solubility provide a theoretical basis for the study of the crystallization process. Currently, the thermodynamic models of solid-liquid equilibrium mainly include Wilson, NRTL, and Pitzer models, which are semiempirical models with varying accuracy. Deng et al. [12] calculated the solubility data of the system NH4Cl-(NH2)2CO-H2O at 283.15 K, 313.15 K, and 353.15 K using the Wilson and NRTL models, demonstrating that the two thermodynamic models can efficiently fit the experimental data. Sun et al. [13] used the Pitzer model to explore the solid-liquid equilibrium of the system HCl-NaCl-MgCl2-H3BO3-H2O at 298.15 K, as the predicted solubility data of the system. The evaporation crystallization route and the order of salt precipitation were determined using the model of the quinary system at 298.15 K.
The cooling crystallization is a method mainly for the crystal. The method has the advantages of simple operation and has low requirements for equipment and operators. Li [14] studied the growth mechanism of high purity (NH2)2CO crystals by using intermittent sampling method and analyzed the particle size under cooling mode. Dang et al. [15] explored the effects of four typical impurities (Fe3+, Al3+, F−, and SO42−) on the morphology of H3PO4·0.5H2O by using the solution cooling crystallization method and the online focusing beam reflection monitoring device.
The cocrystallization process of the system was studied by the cooling crystallization method combining with FBRM and PVM online monitoring technology based on the solubility data of the system, and the crystal composition and morphology were analyzed by XRD, SEM, and EDS characterization methods.
2. Experimental Section
2.1. Reagents and Apparatus
All of the chemical reagents used in the experiment are of analytical grade and are presented in Table 1. The water used in the experiment was self-made ultrapure water with a resistivity of 18.25 MΩ·cm−1. All instruments and equipment used in this study are listed in Table 2.
2.2. Experimental Methods
2.2.1. Phase Equilibrium Experiment
The isothermal solution equilibrium method was applied to study the phase equilibrium of the system [16]. First, a saturated solution of one component was prepared in the conical bottle, into which different amounts of another component were added. Second, the conical bottle was sealed with a rubber plug, placed in a constant temperature water bath, and continuously oscillated at the set temperature to ensure complete mixing of the materials in the conical bottle. The pre-experimental study revealed that the equilibrium time was 5.5 hours at 303.15 K, 323.15 K, and 333.15 K and 6.0 hours at 343.15 K. Therefore, 6.0 hours was selected as the equilibrium time for the experiment at each temperature. After 6.0 hours of constant temperature oscillation, the oscillation was terminated and the static temperature was maintained for 10 minutes. Once the solution had been clarified, the supernatant was extracted using a pipette equipped with a filter head. Similarly, the moist residue was transferred into the designated weighed volumetric bottles using a medication spoon. The electronic balance was used to weigh, and the solution was diluted to a constant volume with ultrapure water immediately after weighing, and the composition was analyzed. The remaining solid phase was filtered directly, heated, and finally dried. It was ground and characterized by XRD. The feasibility of the phase equilibrium research method has been described in the literature [11, 17].
2.2.2. Crystallization Process Experiment
The experimental device of cooling crystallization method is shown in Figure 1. First, the saturated solution was prepared from the solubility data of the ternary system KCl-(NH2)2CO-H2O. The substance was completely dissolved at 5 K higher than the saturation temperature of the solution; then, the temperature of the solution was reduced to the set temperature at a constant cooling rate of 0.2 K/min. During the process, the crystallization process was monitored using the FBRM and PVM. Finally, the crystal slurry was filtered and dried to obtain the crystal product.

2.3. Analysis Methods
The content of (NH2)2CO was analyzed with the p-dimethylaminobenzaldehyde colorimetric method [18], the content of KCl was analyzed by the indicator method of ferroammonium alum [19], the content of H2O was calculated through the subtractive method, and the equilibrium solid phase composition was identified by Schreinemaker’s method of wet residues [20] and XRD method.
The XRD test was conducted with the following parameters: the starting angle was set to 5°, the ending angle to 90°, and the scanning step to 0.013. The tube voltage and tube current were set to 45 kV and 40 mA, respectively, and the Cu target Kα radiation source was used (with KA1 = 1.540598 nm and KA2 = 1.544426 nm). SEM test parameter setting: acceleration voltage was 5 kV and working distance was from 9.0 mm to 12.4 mm. EDS test parameter setting: acceleration voltage was 5 kV and working distance was 8.4 mm.
3. Results and Discussion
3.1. Phase Equilibrium of the Ternary System KCl-(NH2)2CO-H2O
The phase equilibrium data of the ternary system KCl-(NH2)2CO-H2O at 303.15 K, 323.15 K, 333.15 K, and 343.15 K are presented in Table 3.
The equilibrium phase diagram and stereoscopic phase diagram of the ternary system KCl-(NH2)2CO-H2O at 303.15 K, 323.15 K, 333.15 K, and 343.15 K are presented in Figures 2 and 3, respectively. In Figure 2, the curves M1E1, M2E2, M3E3, and M4E4 indicate the solubility curve of (NH2)2CO, whereas the N1E1, N2E2, N3E3, and N4E4 represent the solubility curve of KCl. Region I illustrates the unsaturated solution region, region II represents the crystallization region of pure crystal (NH2)2CO, region III represents the eutectic region of crystal KCl and (NH2)2CO, and region IV represents the crystallization region of pure crystal KCl. The composition of the equilibrium solid phases in points E1, E2, E3, and E4 analyzed by the X-ray diffraction method is shown in Figure 4.



As Figure 2 indicates, the equilibrium phase diagram of the ternary system is composed of one invariant point, two isotherm solubility curves that are univariate, one unsaturated region, and three regions of crystallization, namely, crystallization zone of (NH2)2CO and KCl and eutectic zone of KCl and (NH2)2CO at 303.15 K, 323.15 K, 333.15 K, and 343.15 K. In Table 3, we observe that when the temperature drops, the position of the invariant point moves down, and the area of regions I and IV of the system decreases, whereas that of regions II and III increases, indicating that the decrease in temperature promotes precipitation of (NH2)2CO. At 303.15 K, the compositions of the eutectic saturated solution of the invariant point E1 are 13.96% of KCl and 51.14% of (NH2)2CO. At 323.15 K, E2 are 12.49% of KCl and 60.15% of (NH2)2CO. At 333.15 K, E3 are 11.85% of KCl and 63.39% of (NH2)2CO. At 343.15 K, E4 are 10.29% of KCl and 69.27% of (NH2)2CO. It can be seen that unlike KCl, the content of (NH2)2CO in the eutectic saturated solution increases with the rise in temperature. Figure 3 demonstrates that the phase diagram trend of the system is consistent at 303.15 K, 323.15 K, 333.15 K, and 343.15 K. Therefore, the change of temperature only affects the solubility data but does not affect the change law of the system. Figure 4 shows that the strength peak of the equilibrium solid phase at each temperature corresponds to the characteristic peak of KCl and (NH2)2CO standard cards, which imply that E1, E2, E3, and E4 are the invariant points, and the equilibrium solid phase is the simple mixture of KCl and (NH2)2CO, without the formation of solid solution, double salt, or hydrate. Therefore, the ternary system is a simple cosaturation system, which is consistent with previous reports [10, 21].
3.2. Correlation Calculation of the Thermodynamic Model
Based on the traditional solid-liquid equilibrium theory, the solubility equation of solute in solvent can be simplified as illustrated by (1) under fixed temperature and pressure [22, 23]:where xi is the mole fraction of substance i in the liquid phase, γi represents the activity coefficient of substance i in the liquid phase, and ΔfusH is the enthalpy of melting of pure substance i. R refers to the gas constant, and its value is 8.314 J/(mol·K). Tm is the melting temperature of substance i, and T is the experimental temperature.
3.2.1. Wilson Model
The Wilson model was proposed in 1964. Based on the concept of local composition and molecular considerations, Wilson first suggested the expression of the excess Gibbs energy of a binary solution and derived the equation of the activity coefficients. Its activity coefficient can be calculated using the following equations [22]:where γi represents the activity coefficient, Λij stands for the Wilson model parameter, Δλij(=λij − λii) is the binary interaction energy parameter, and Vi and Vj refer to the molar volumes of pure matter i and j, respectively.
To increase the accuracy of the binary parameters, parameters aij and bij are introduced and Λij can be calculated as follows [24, 25]:where aij and bij are the parameters used to associate solubility data with the Wilson model.
3.2.2. NRTL Model
The NRTL model, which is based on the concept of local composition, has been widely employed in fluid-phase equilibrium modeling. The activity coefficient can be calculated using the following equations [26]:where γi refers to the activity coefficient, Gij is the NRTL model parameter, Δ is the binary interaction energy parameter, tij is the bivariate correlation parameter, and α is an adjustable parameter that ranges from 0.2 to 0.47; in this paper, α = 0.3.
Similarly, to make the binary parameters more accurate, aij and bij were also introduced into the NRTL model and Gij can be calculated as follows [21, 25]:where aij and bij are the parameters used to associate solubility data with the NRTL model.
3.2.3. Error Analysis Method
The objective function of Wilson and NRTL models is described in the following equation:where N refers to the number of experimental points, represents the experimental value of mass fraction of component , and denotes the calculated value of the mass fraction of component i.
Relative mean deviation (RAD) and root mean square difference (RMSD) are introduced to determine the correlation degree of the model [27]:
To begin with, the binary correlation parameters for the Wilson and NRTL models of KCl-H2O and (NH2)2CO-H2O can be obtained by considering the pertinent physical parameters and the solubility data of KCl and (NH2)2CO at various temperatures documented in the literature [22, 28, 29]. Second, these data were put into the Wilson and NRTL equations and the model parameters of KCl-(NH2)2CO were calculated through the regression method. Finally, the solubility data and RAD and RMSD with two models were determined. The results of two models’ parameters are presented in Table 4, and the calculated values of solubility and RAD and RMSD with two models are shown in Table 3. Comparison of experimental values of solubility with the calculated values of model association is illustrated in Figure 5.

(a)

(b)
The obtained Wilson and NRTL models’ parameters are used to predict the solubility data at 283.15 K, 313.15 K, and 353.15 K. The calculated values of RAD and RMSD of the predicted values are listed in Table 5. The comparison between the predicted values and the literature values [10, 11] is shown in Figure 6.

(a)

(b)
Table 3 demonstrates that the maximum values of RAD and RMSD of the Wilson model were 2.18 × 10−2 and 0.83, respectively. The maximum value of RAD and RMSD of the NRTL model was 1.69 × 10−2 and 0.40, respectively. Figure 5 illustrates that two models’ calculation value and experimental value show good correlation, suggesting that these two models can be used to calculate the solubility data of the system.
Table 5 shows that the maximum values of RAD and RMSD of prediction and the literature value of the Wilson model was 4.81 × 10−2 and 1.05, respectively, and that of the NRTL model was 3.51 × 10−2 and 0.93, respectively. Figure 6 shows that the predicted values of two models match with the literature values at 283.15 K to 353.15 K, and the NRTL model is slightly better than the Wilson model.
3.3. Crystallization Process Analysis of the System KCl-(NH2)2CO-H2O
Based on the composition of KCl-(NH2)2CO-H2O at the invariant point at 343.15 K ( ((NH2)2CO) = 10.29%, (KCl) = 69.27%, and (H2O) = 20.44%), the temperature of the solution was decreased from 348.15 K to 333.15 K at 0.2 K/min. The changes in temperature, total particle number, and average chord length in the crystallizer during the cooling crystallization process were monitored using FBRM, and the results are presented in Figure 7. Crystal growth images of the crystallization process were collected using PVM, and the results are shown in Figure 8. XRD and SEM characterization analyses’ results are presented in Figures 9 and 10, respectively.




Figure 7 illustrates how the temperature fluctuation in the system can be categorized into three distinct processes: isothermal dissolution, cooling crystallization, and isothermal crystal growth.
Isothermal dissolution process contains Stage A. Above the saturation temperature of the solution of 5 K, when the total number of particles in the mold and the average chord length are close to 0, KCl and (NH2)2CO in the solution are considered to be completely dissolved at this stage.
Cooling crystallization process contains Stages B–D. In stage B, at a constant cooling rate (0.2 K/min), the total number of particles in the crystallizer and the average chord length are stably close to 0, indicating that no crystal nucleus has been precipitated. In Stage C, crystal nuclei appear when the temperature drops to 338.55 K, as shown in Figure 8(a). During this stage, a substantial amount of heat is released during nucleation, causing the temperature inside the mold to increase to 339.05 K. This increase in temperature is reflected in a minor peak in the temperature curve, resulting in a temperature difference of 0.27 K. The results indicated that the data should be recorded quickly during the measurement of the width of the metastable zone through monitoring the temperature in the mold. When optimizing the crystallization process, the seed should be added in advance to induce nucleation and avoid nucleation outbreak. In Stage D, as the cooling process continued, the total number of particles gradually increased and the average chord length did not change significantly, indicating that the cooling crystallization process of the system was controlled by nucleation. In Stages C and D, PVM monitoring crystal growth processes are shown in Figures 8(a)–8(c) and suggest that the crystal nucleus longitudinally grew rapidly, forming crystals with fine needle-like shapes. Then, as the longitudinal growth rate of the crystal gradually decreased, the lateral growth rate is relatively growing stably from 151 μm to 630 μm. According to the findings, the longitudinal growth of crystals is highly responsive to variations in temperature, whereas the transverse growth rate remains relatively steady. These results can serve as a useful guide for the subsequent crystallization process and the regulation of the length-to-diameter ratio of crystal products.
Isothermal crystal growth process contains Stage E. The variation range of total particle number is small and the average chord length appears unchanged, indicating the basic equilibrium at this stage. The crystal image monitored by PVM at this stage is shown in Figure 8(d), and the transverse width of the crystal is 643 μm in this stage.
Figure 9 indicates that the strength peaks of the crystal at 2θ = 22.2°, and 2θ = 28.3° corresponded to the characteristic peaks of KCl and (NH2)2CO standard cards, respectively, and the peak strength of (NH2)2CO is stronger. The chemical analysis shows that the crystal compositions are (KCl) = 5.58% and ((NH2)2CO) = 94.42%, indicating that the content of (NH2)2CO in the crystal product exceeds that of KCl. Combined with the analysis of the cooling crystallization process, due to the initial feeding (NH2)2CO was more, the cooling promoted the precipitation of (NH2)2CO and the nucleation of pure (NH2)2CO occurred at the initial stage of cooling crystallization [14], which corresponded to the existence of explosive nucleation in Stage C of the cooling crystallization process in this paper. Therefore, we speculate that (NH2)2CO was precipitated first, and this was followed by KCl precipitation during the whole cooling process. Finally, the crystal is composed of KCl and (NH2)2CO, and (NH2)2CO is in majority.
In Figure 10(a), the crystal size distribution is not uniform and had a rod-like and irregular geometric shape. There may be crystal nucleus precipitation in the later stage of the crystallization process, resulting in a wide crystal size distribution. Therefore, when optimizing the crystallization process, the cooling rate should be reduced or the holding time should be extended to obtain more uniform crystal products. In Figure 10(b), the end face of the crystal product appears hollow and uneven, which is caused by the rapid longitudinal growth of the crystal. Combined with Figure 10(d), we infer that the crystal belongs to the layer growth mechanism. As shown in Figure 10(c), there is a bulge on the crystal surface which forms the basis of nucleation and crystallization in Figure 10(e). To further verify the composition of the crystal, energy dispersion spectroscopy (EDS) was conducted with 6500 magnification times, as shown in Figures 11 and 12. The results show that the composition of this microregion mainly contains five elements: C, N, O, K, and Cl. Based on the distribution of elements C, N, and O in Figures 11(c)–11(e), the surface of the crystal body contains the following key elements: C, N, and O, indicating that the crystal body is (NH2)2CO. From the distribution of K and Cl elements in Figures 12(f)–12(g), the component of the microcrystal protruding from the crystal body surface mainly contains KCl. Therefore, during the cooling crystallization process, it is possible that KCl grows on the surface of (NH2)2CO crystal and finally forms simple mixed crystal.


4. Conclusions
(1)The study on the phase equilibrium of the ternary system KCl-(NH2)2CO-H2O indicates that the system is a simple cosaturated system at 303.15 K, 323.15 K, 333.15 K, and 343.15 K, and no double salt and solid solutions are formed. The trend of the phase diagram is consistent, and there are three crystallization zones in all of them, namely, crystallization zones of (NH2)2CO and KCl and eutectic zone of KCl and (NH2)2CO.(2)The solubility data of the ternary system at 303.15 K, 323.15 K, 333.15 K, and 343.15 K were calculated using the Wilson and NRTL models, and the correlation parameters of models were obtained. The maximum value of RAD and RMSD of the Wilson model was 2.18 × 10−2 and 0.83, respectively, and that of the NRTL model was 1.69 × 10−2 and 0.40, respectively. This demonstrates that the calculation value and experimental value showed good correlation. Therefore, the two models can be applied to calculate the solubility data of the ternary system. These two models can also be applied to predict solubility data at other temperatures, and the NRTL model exceeded that of the Wilson model.(3)The crystallization process of the system was studied by the cooling crystallization method combined with FBRM and PVM online monitoring technology. It was found that the cooling crystallization process was roughly divided into isothermal dissolution, cooling crystallization, and isothermal crystal growth processes. According to the characteristic analysis of crystal products, it can be speculated that (NH2)2CO precipitates first during the crystallization process, followed by KCl. Therefore, it is possible that KCl grows on the surface of (NH2)2CO crystal, so a simple mixed crystal can be gained.Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
This work was financially supported by the National Natural Science Foundation of China (21868010) and the Science and Technology Programs of Guizhou Province ((2018) 5781).