Solubilities of resorcinol and pyrocatechol and their mixture in supercritical carbon dioxide

https://doi.org/10.1016/j.tca.2011.04.002Get rights and content

Abstract

The equilibrium solubilities of dihydroxy benzene isomers (resorcinol and pyrocatechol) and its mixture were experimentally determined at different temperatures (308, 318, 328, and 338 K) in the pressure range of 9.8–16.2 MPa. In the ternary system, the solubilities of pyrocatechol increased while the solubilities of resorcinol decreased relative to their binary solubilities. A new association model was developed based on the concept of formation of solvate complex molecules to correlate the solubility of the solid for mixed solids in supercritical carbon dioxide (SCCO2). The model equation relates the solubility of solute in terms of the cosolute composition, temperature, pressure and density of SCCO2. The proposed model correlated the solubilities of sixteen solid systems taken from the literature and current experimental data with an average absolute relative deviation (AARD) of around 4%.

Highlights

► The equilibrium solubilities of dihydroxy benzene isomers were determined. ► The solubilities of mixed solids in supercritical carbon dioxide were determined. ► A new association model was developed to correlate the ternary solubilities. ► The proposed model correlated the solubilities of seventeen systems within 4%.

Introduction

The feasibility of an extraction process depends on the solubility of the solute in the fluid. The extraction of component by using supercritical fluid (SCF) requires the knowledge of its solubility in SCF. A large amount of data is available in the literature on the pure component solubility in supercritical fluids [1], [2], [3], [4]. However, in many cases, the solute does not exist in its pure form, so it has to be extracted from a matrix of compounds. The presence of a solid component along with the other solid may either increase or decrease the solubility of the other solute and this may be effectively utilized to separate the solutes [5]. This demands the determination of solubilities of mixtures in supercritical fluids.

The determination of equilibrium solubilities of solids in SCFs at different temperatures and pressures is expensive; hence, the modeling of the solubilities is essential. Several models have been proposed to correlate and predict the solubility of the component in SCF [6], [7], [8], [9], [10], [11], [12]. These models are based on equation of state (EOS) while some models are either empirical or semi-empirical. EOS models require many physical properties, which are estimated by group contribution methods leading to erroneous correlations. To overcome this limitation, semi-empirical and empirical models have been proposed [8], [9], [10], [11], [12], [13], [14], [15], [16], [17]. Many semi-empirical models have correlated the solubility of the compounds in SCFs better than EOS models. Chrastil [8] assumed that a solvate complex was formed between the solute and the solvent molecules at equilibrium and presented the first density-based model. However, this model is not applicable at high solubilities. Therefore, Adachi and Lu [9] correlated the average association number κ to a second-order polynomial of the solvent density. del Valle and Aguilera [10] modified the Chrastil equation by taking into account the change in the enthalpy of vaporization with temperature. Thus, the Adachi–Lu [9] equation and del Valle and Aguilera [10] equation corrected the effect of density and temperature on solubilities. These effects were combined by Sparks et al. [11]. Sparks et al. [12] also modified the Chrastil equation, which was further modified by Garlapati and Madras [13]. Another density based model was presented by Jouyban et al. [14]. Based on the theory of response surface, Gordillo et al. [15] proposed another model dependent only on temperature and pressure. Ch and Madras [16] have presented a four parameter model based on the association theory. A widely used density-based model based on the theory of dilute solutions was proposed by Méndez Santiago–Teja (MT) [17]. This model is excellent for verifying the data consistency of solubility data across different isotherms. However, none of the above models can directly be applied to correlate the solubilities of solid mixtures because they do not account for the concentration of the second solute.

The presence of one solute may alter the solubility of the other solute at the same temperature and pressure [18]. The phase behaviour of solute mixtures in SCCO2 and the effect of the cosolute have been discussed in the review on solid mixtures in SCCO2 [19]. The semi-empirical model should also involve the composition of the cosolute while correlating the solubility of mixed solids in SCFs. All the semi-empirical models available in literature correlate the solubility of each component separately for mixed solids. To the best of our knowledge, no semi-empirical model is available in literature that correlates the solubility of the mixed solids in SCFs involving the composition of cosolute.

The dihydroxy benzene isomers, which are used as pharmaceutical products, find many applications in the synthesis of organic compounds [20], [21], [22]. Pyrocatechol is used as a model compound in various reactions of phenolic compounds [21]. Resorcinol is used in the preparation of organic aerogels [22]. Therefore, determining the solubilities of these compounds is essential. In this work, we have experimentally determined the pure component solubility of dihydroxy benzene isomers and their mixture solubilities at 308, 318, 328 and 338 K in a pressure range of 9.8–16.2 MPa. Along with the experimental solubility data, we have derived a model based on the concept of association of molecules to correlate the solubility of solid mixture in SCFs. The solubility data of the solid mixtures in SCCO2 was taken from literature and was correlated by MT model along with the association model.

Section snippets

Materials

Carbon dioxide gas (CO2) (CAS number 124-38-9, 99 wt%) was purchased from Vinayaka Gases (India) and was passed through a bed of silica gel for further purification to 99.9 wt%. 1,2-Dihydroxy benzene (catechol, pyrocatechol) (CAS number 120-80-9, 98 wt%), 1,3 dihydroxy benzene (resorcinol, CAS number 108-46-3, 99 wt%) and 1,4-dioxane (CAS number 123-91-1) (HPLC grade) were purchased from Merck (India). Fig. 1 shows the chemical structure of pyrocatechol and resorcinol.

Measurement of solubilities of solids

The flow apparatus, which uses

An association model for mixed solids in SCF

The solubility of the compound differs from its pure component solubility in the presence of the other solute. The presence of a cosolute alters the solubility of the solute in SCF. The solubility of the solid solutes in SCCO2 can be derived in terms of association of solute molecules and formation of solvate complex molecules [16]. A similar kind of approach was used and extended to the multicomponent solutes in SCCO2. It is assumed that each solute molecule is associated with κ molecules of

Results and discussion

The equilibrium solubilities of pure dihydroxy benzene isomers were determined at 308, 318, 328, and 338 K and at 10 and 13.73 MPa (Table 1). Yamini et al. [27] have determined the pure component solubilities of the isomers at the above temperatures and at high pressures i.e. in a pressure range of 12.16–40.53 MPa. Fig. 2 compares the solubility data obtained in this study with that obtained by Yamini et al. [27]. The cross over pressure was between 10 and 12 MPa for resorcinol and was between 11

Conclusions

The equilibrium solubilities of the dihydroxy isomers (resorcinol and pyrocatechol) and their mixture were determined experimentally at 308, 318, 328, and 338 K over the pressure range of 10–16.2 MPa. An association model was developed to correlate the solubility of the solid for a solid mixture. The solubility enhancement of the solute for the solid mixture in SCCO2 was successfully explained in terms of cosolute effect. The model equation involves only five parameters and was able to correlate

Acknowledgment

The authors thank CSIR for financial support.

References (41)

Cited by (0)

View full text