Measurement and correlation of solid-liquid equilibrium data for nitroguanidine

CHEN Li-zhen, ZHAO Chong-yang, ZHANG Le, LIU Yuan-yuan, WANG Jian-long, CAO Duan-lin

(School of Chemical Engineering and Technology, North University of China, Taiyuan 030051, China)

Abstract： Nitroguanidine(NQ) is a high-energy and low-sensitivity explosive and solid-liquid equilibrium data are significant for study on crystallization of NQ. The solubilities of NQ in water, dimethyl sulfoxide, N, N-dimethylformamide, 1,4-butyrolactone and dimethyl sulfoxide+water, N, N-dimethylformamide+water were measured by dynamic laser monitoring within a temperature range from 298.15 K to 338.15 K. The experimental data were correlated by modified Apelblat equation, λh equation, CNIBS/R-K model, and Jouyban-Acree model. The results show that the four thermodynamic models can all be used to predict solubility with high accuracy. Accrding to the Akaike’s information criterion (AIC), the better models for correlating the solubility of NQ are judged. Additionally, the dissolution enthalpy, entropy and Gibbs free energy were calculated by the van’t Hoff equation.

Key words： nitroguanidine (NQ); solubility; correlation models; thermodynamic properties

0　Introduction

Nitroguanidine(NQ) is a high-energy and low-sensitivity explosive, and it is widely used as a propellant and composite explosives charge[1]. However, NQ crystal from the production is in the form of long, thin needles, which leads to a very low bulk density. Thus, it is necessary to improve crystal technology to obtain spherulitic NQ particles.

The solid-liquid phase equilibrium data are a key element in the crystallization process in respect of the choice of the best solvent medium, the initially dosage of solute and the terminal holdings of solute in the solvent, and the operation curve of process. In addition, the mathematical expression of phase equilibrium data is the essential basis for the study on the crystallization dynamics and process simulation. Though several research reports on the solubility of NQ have been published[2-5], the solid-liquid equilibrium data of NQ in water and organic solvents have not been systematically reported in the literatures.

In recent years, the measurement and prediction of solid-liquid equilibrium data have received more attention. Some analysis or dynamic methods have also been established[6-7]. Many prediction models suitable for single solvent and cosolvent have been proposed, too[8-9]. In present study, a laser dynamic method was used for the measurement of solubility of NQ in water, N,N-dimethylformamide (DMF), dimethyl sulfoxide (DMSO), 1,4-butyrolactone (GBL), binary solvent of (DMSO+water) and (DMF+water), within a temperature range from 298.15 K to 338.15 K at atmospheric pressure. The solubility data were correlated by the modified Apelblat model,λhmodel, CNIBS/R-K model and Jouyban-Acree model. The predicted value were compared with the experiment values. Furthermore, the thermodynamic parameters of NQ were calculated according to van’t Hoff equation and Gibbs equation.

1　Experiment

1.1　Materials and apparatus

The crude NQ was supplied by Liaoning Qingyang Chemical (Group) Co. Ltd., China, subsequently recrystallized in water, and the purity was analyzed by HPLC, 99%. All solvents were analytic reagent grade, supplied by Guangdong Xilong Chemical Co. Ltd, China, and the distilled water was self-made.

The solubility measurement device consists of a crystallizer, a temperature controlling device, and a laser monitoring system. The laser monitoring system was manufactured by Department of Physics, Peking University, China.

1.2　Solubility measurement

The solubility was measured by the laser dynamic method[10]. All experiments were done three replications, and the average values were used to calculate the solubility.

The mole fraction solubility (x1) of NQ in a solvent is calculated by

(1)

wherex1represents the mole fraction solubility of NQ in solvents;m1andM1mean the mass and molar mass of NQ, respectively;miandMi(i=2,3) mean the mass and molar mass of solvents, respectively. Whenn=2, the solvent is single; andn=3, the solvent is binary mixture.

2　Correlated models

In the last 20-30 years, many very useful semi-theoretical or empirical mathematic models describing the solute solubility in single or mixed solvent were studied[11-14]. In this paper, the solubilities of NQ in four pure solvents are correlated with temperature and composition by modified Apelblat equation andλhequation, of which the solubilities of NQ in two binary solvents are correlated by modified Apelblat equation, CNIBS/R-K equation, and Jouyban-Acree model, respectively.

2.1　Modified Apelblat equation

The modified Apelblat equation[15-16]is derived from the Clausius-Clapeyron equation, which can be used to correlate the solubility with temperature. The equation is expressed as

lnx1=A+B/T+ClnT,(2)

(2)

whereTrepresents the absolute temperature;A,BandCare model parameters.

2.2　λh equation

Theλhequation was proposed by Buchowski[17-18], which used to express the relationship between the mole fraction solubility and the temperature. The equation is described as

(3)

whereTandTmare the system temperature and the melting temperature of the solute, respectively;λandhare model parameters.

2.3　CNIBS/R-K equation

The CNIBS/R-K equation proposed by Acree[19]is mainly used to correlate the solubility with the composition of the binary solvent when the temperature is constant[20]. The expression is

(4)

wherexAis the mole fraction of solvent A in the binary mixtures;B0,B1,B2,B3andB4are the model parameters.

2.4　Jouyban-Acree model

The Jouyban-Acree(J-A) model[21]is widely used for describing the effect of both temperature and solvent initial composition on solute solubility in a binary solvent system. The equation can be expressed as

(5)

whereA1-A9are model parameters.

3　Estimation of thermodynamic parameters of solution

According to the van’t Hoff equation, the dissolution enthalpy, entropy and Gibbs energy can be estimated when the solubility of solute was measured[22-23]. The relationship can be expressed by

(6)

4　Results and discussion

4.1　Measured solubility

The solubility data of NQ in water, DMF, DMSO, and GBL are listed in Table 1 and Fig.1, respectively.

As shown in Table 1, the solubilities of NQ in above solvents all increase with the temperature. The solubility order is DMSO>DMF>GBL>water. A weaker temperature dependence is shown in DMF and DMSO. As the temperature increases from 298.15 K to 338.15 K, the solubility values in DMF and DMSO increase by 55% and 21%, respectively; while increase by 696% and 597% in water and GBL, respectively.

Table 1　Experimental and calculated solubility values of NQ in pure solvents(×10-2 mol/mol)

As for crystallization, too high or too low solubility is not appropriate. Therefore, the binary solvent is considered.

The solubility values of NQ in (DMSO+water) and (DMF+ water) are listed in Tables 2-3 and Fig.1, respectively. The moderate solubility is shown in them, all of which increases with temperature and the mole fraction of DMSO or DMF.

Table 2　Experimental and calculated solubility values of NQ in DMSO+water (×10-2 mol/mol)

Table 3　Experimental and calculated solubility values of NQ in DMF+water (×10-2 mol/mol)

Fig.1　Solubilities for NQ in different solvents

4.2　Solubility correlation

The solubility of NQ in pure solvents is correlated by the modified Apelblat equation andλhequation, of which that in (DMSO+water), and (DMF+water) is correlated by the modified Apelblat model, CNIBS/R-K model, and the Jouyban-Acree model. The fitted model parameters are listed in Tables 4-9. The calculated solubility values are shown in Tables 1-3.

The root-mean-square deviation (δ) is calculated to evaluate the accuracy of predicted models as

(7)

The δ values and the square of the correlation coefficient (R2) of these models are shown in Tables 4-7.

Table 4　Parameters, R2 and δ of Apelblat equation in solvents

Table 7　Parameters, R2 and δ of Jouyban-Acree model in binary solvents

4.3　Evaluation of thermodynamic models

In order to find the optimal model for NQ, the Akaike Information Criterion (AIC)[24-25]is employed to evaluate the relative applicability of these models. The values of AIC for all models are listed in Table 8.

The result shows that the AIC of modified Apelblat equation is lower in pure solvents, which has better correlation effect, and the Jouyban-Acree model is more suitable for correlating the solubility of NQ in binary solvents.

Table 8　AIC values of two models in pure solvents and three models in binary solvents

4.4　Thermodynamic parameters of solutions

According to Eq.(6), the calculated thermodynamic parameters of NQ in the above mentioned solvents are listed in Table 9. The dissolution Gibbs free energy is calculated by

ΔdisG=ΔdisH-TmeanΔdisS,

(8)

whereTmeanis mean temperature, 317.63 K.

The dissolution enthalpy is positive in all solvents, which means that all the dissolution processes are an endothermic process.

Table 9　Thermodynamic properties of NQ in solvents at 317.63 K

5　Conclusion

The solubilities of NQ in pure solvents (water, DMSO, DMF, GBL) and binary solvents (DMSO+water, DMF+water) are measured. The solubility order is DMSO>DMF>GBL>water. The solubilities increase with the temperature and mole fraction of DMSO or DMF. The solubilities of NQ in DMF and DMSO have a weaker temperature dependence. The solubilities of NQ in pure solvents are correlated by modified Apelblat andλhequation. The modified Apelblat equation gives the better correlate precision. The solubilities in (DMSO+water) and (DMF+water) are correlated by the modified Apelblat, CNIBS/R-K, and the Jouyban-Acree model, of which the Jouyban-Acree model is more adaptable. At 317.63 K, the dissolution enthalpy values of NQ in water, DMSO, DMF and GBL are 39.95, 3.98, 9.48 and 37.65 kJ·mol-1, respectively; and the dissolution Gibbs free energy values are 17.40, 3.89, 6.04 and 13.13 kJ·mol-1, respectively.