- Email: [email protected]
Calorimetric study of solvation of low polar solutes in propylene glycol and methyl cellosolve at 298 K
Thermochimica Acta 589 (2014) 247–251
Contents lists available at ScienceDirect
Thermochimica Acta journal homepage: www.elsevier.com/locate/tca
Calorimetric study of solvation of low polar solutes in propylene glycol and methyl cellosolve at 298 K I.A. Sedov *, M.A. Stolov, B.N. Solomonov Chemical Institute, Kazan Federal University, Kremlevskaya 18, Kazan 420008, Russia
A R T I C L E I N F O
A B S T R A C T
Article history: Received 3 February 2014 Received in revised form 31 May 2014 Accepted 3 June 2014 Available online 5 June 2014
Enthalpies of solution of low-polar substances: aliphatic and aromatic hydrocarbons, including alkanes, cycloalkanes, alkylbenzenes, and halobenzenes in two solvents, propylene glycol and methyl cellosolve, were measured at temperature T = 298.15 K using titration calorimetry. In addition, the enthalpies of solution of poorly soluble solid solutes, naphthalene and biphenyl, in these solvents were determined using indirect method by measuring the enthalpies of solution of liquid mixtures containing them. The enthalpies of solvation from the gas phase were calculated from these data. The enthalpies of solvation in propylene glycol and methyl cellosolve and previously reported enthalpies of solvation in ethylene glycol were correlated with each other and with molecular parameters of solutes. ã 2014 Elsevier B.V. All rights reserved.
Keywords: Propylene glycol Methyl cellosolve Ethylene glycol Enthalpy of solution Solvation Solution calorimetry
1. Introduction Methyl cellosolve (2-methoxyethanol, CH3OCH2CH2OH, further MC) and propylene glycol (1,2-propanediol, CH3CH(OH)CH2OH, further PG) are two structural isomers and two important industrial solvents [1]. Methyl cellosolve is used as a solvent in dye and rubber industries, production of cellulose acetate films and laboratory organic synthesis. MC is a toxic compound which can be converted by alcohol dehydrogenase into methoxyacetic acid which can enter the Krebs cycle [2]. In contrast, PG has a very low toxicity for humans.[3] Thus, it is used as a solvent for pharmaceuticals and in food industry, as well as in rubber and polymer production, and as a component of automotive antifreeze. Both MC and PG molecules differ by a methyl group from a molecule of another common industrial solvent, ethylene glycol (1,2-ethanediol, HOCH2CH2OH, further EG). The solvent effect on equilibria in chemical reactions, solubility, interfacial distribution properties of dissolved compounds is governed by the Gibbs free energy change. The enthalpy of solvation (transfer of solute from the gas phase into solvent) determines the dependence of the Gibbs energy of solvation on temperature according to Van't Hoff equation. Despite a broad practical use of PG and MC, thermodynamic functions of solvation or solution in these solvents are almost not studied.
* Corresponding author. Tel.: +7 9600503916; fax: +7 8432315346. E-mail address: [email protected] (I.A. Sedov). http://dx.doi.org/10.1016/j.tca.2014.06.003 0040-6031/ ã 2014 Elsevier B.V. All rights reserved.
Our special interest is the solvophobic effect in self-associated solvents and its influence on the solvation properties. Both MC and PG are self-associated due to intermolecular hydrogen bonds. In our previous papers [4–6] we developed a methodology to describe the solvophobic effect in various solvents on the basis of the values of Gibbs free energy and enthalpy of solvation for lowpolar solutes. Thus, in this work we have selected a number of such solutes to determine their enthalpies of solution in PG and MC. Calorimetric measurement of the enthalpies of solution for non-polar solutes in PG is complicated by their low solubility, low speed of dissolution, and a high viscosity of solvent. No values of the enthalpies of solution of hydrocarbons and their halogenated derivatives in PG were found in literature. The solubility of hydrocarbons in MC is higher, but the calorimetric data for their solutions are also very limited. In the present work, novel thermodynamic data for solutions of hydrocarbons and their halogenated derivatives in PG and MC at infinite dilution and T = 298 K are obtained using titration calorimetry technique. 2. Experimental 2.1. Materials and methods Propylene glycol and methyl cellosolve with purity >0.99 were purchased from Acros Organics. All the solutes were at least 0.99 pure grade from Sigma–Aldrich, Acros and Fluka. They have been used without further purification. The absence of significant
248
I.A. Sedov et al. / Thermochimica Acta 589 (2014) 247–251
amounts of impurities has been confirmed by gas chromatography. 2.2. Enthalpy of solution measurements Enthalpies of solution at infinite dilution at temperature T = 298.15 K and pressure p = 0.1 MPa have been determined using TAM III precision solution calorimeter in 100 ml glass calorimetric vessel equipped with a gold stirrer, a Joule heater and a thermistor. The calorimeter operates under semi-adiabatic condition. The vessel is placed inside a thermostatted metal cylinder, and the temperature change inside the vessel is measured. The vessel exchanges heat with the thermostat in agreement with Newton's law of cooling, which leads to the exponential dependence of the vessel's temperature from time, if no heat is evolved or absorbed due to any physical or chemical processes. In experiments, the heat effects were calculated from temperature vs time curves using Regnault–Pfaundler method [7], which takes into account exponential character of the baseline. This method is integrated into SolCal software used to operate the calorimeter. For liquid solutes, calorimetric titration method was used. The vessel was initially filled with 100 ml of pure solvent. After thermostatting and heater calibrations, 50–100 ml of solute was added from an electronically operated microsyringe in 5–20 ml portions. The heat effect of each addition was determined from a calorimetric curve. For every solute, the experiments were repeated 2–3 times with a fresh portion of solvent. An average value of enthalpy of solution from all dissolution curves was taken. If a solute is volatile and poorly soluble, we should make a correction in order to account for the heat loss due to evaporation of a part of the solute molecules to a free volume of the calorimetric vessel. The standard molar enthalpy of solution obtained from calorimetric curves DHobs is equal to the sum of heat effects of dissolution and evaporation of some part of a solute:
DHobs ¼ ð1 ’ÞDsoln HA=S þ ’Dvap HA
(1),
is the molar enthalpy of solution of solute A in where DsolnH solvent S, DvapHA is the enthalpy of evaporation of A, ’ is the molar fraction of solute in the vapor phase that can be calculated by equation: A/S
1 A=S 1 þ RTnS = pAsat g 1 V free
’¼
(2)
Here pAsat is the saturated vapor pressure of pure substance A,
g A=S 1 is the limiting activity coefficient of A dissolved in S, Vfree is the volume of the free space in the calorimetric vessel, vS – number of
moles of solvent in the vessel. However, in all the cases except for saturated hydrocarbons dissolved in PG, the contribution of evaporation process to the measured enthalpies of solution is very small. For example, even for volatile hexane in MC the estimated value of ’DvapHA is only 0.05 kJmol1, which is less than the uncertainty of the experiment. Thus, we assumed that DHobs=DsolnHA/S. For alkanes and cyclohexane in PG, which have a poor solubility and low liquid–gas distribution coefficient, we calculated the corrected values of DsolnHA/S by Eq. (1) using literature values of DvapHA [8], pAsat [9], and the values of g A=S 1 estimated from the solubilities reported in literature [10]. Solid solutes were placed in sealed glass ampoules, which were then broken inside a calorimetric vessel with the solvent. The heat effect of dissolution following ampoule break was measured. However, the process of complete dissolution took up to several hours. The results were very sensitive to the method of calculation since the baseline is not absolutely stable on such a long timescale. The values of DsolnHA/S can differ by up to 3–4 kJmol1 depending
on which moment is thought to be the end of dissolution and which part of the baseline after dissolution is chosen. To obtain more accurate results, we developed another experimental method. Given that we want to find the molar enthalpy of solution DsolnHA/S of a solid compound A in a solvent S, but A has a poor solubility in S. We choose one of the solvents (let us denote it by B) in which compound A has the best solubility and prepare a concentrated (but not saturated) solution (A + B) in this solvent with the molar fraction of A xA. It is important for B to be relatively well soluble in S, otherwise the results can be inaccurate. Then we measure the molar enthalpy of solution of the mixture (A + B) by means of titration calorimetry in solvent S (DsolnHA+B/S) and in another (reference) solvent R (DsolnHA+B/R). The molar enthalpies of solution for compound B must be measured separately in both solvents S and R (DsolnHB/S and DsolnHB/R). If all solutions are very diluted and we can neglect the interactions between the molecules of A and B, then the following relationship is correct (due to the fact that enthalpy is a function of state): 0
Dsoln HA=S ¼
1
B C C ðAþBÞ=S Dsoln HA=R þ B Dsoln HðAþBÞ=R BDsoln H C @ A B=S B=R ð1 xA Þ Dsoln H Dsoln H xA (3).
Here the enthalpy of solution of solid A in solvent R, DsolnHA/R must be known or measured. If we choose a solvent in which A has a good solubility, then the value of DsolnHA/R can be measured with a good precision. It is also important to use a mixture with a high value of xA, because the uncertainties of four measured enthalpies of solution are divided by xA making the error of DsolnHA/S inversely proportional to it. This approach resembles that used by Desnoyers et al. [11] for a liquid solute (benzene) to determine its enthalpy of solution in water-rich binary aqueous-organic solvent. In our experiments, we prepared the mixtures of naphthalene and biphenyl with tetrahydrofuran with the molar fraction of a solute around 0.2. The enthalpies of solution of tetrahydrofuran in PG, MC, and cyclohexane were measured and found to be 1.57, 1.20, and 3.31 kJmol1 respectively. Solid aromatic hydrocarbons rapidly dissolve in other hydrocarbons, and cyclohexane has been chosen as (reference) solvent R. For both solid naphthalene and biphenyl, the enthalpies of solution in cyclohexane were measured (22.7 and 23.8 kJmol1 respectively). The enthalpies of solution of the liquid mixtures in PG, MC, and cyclohexane were also measured, and the enthalpies of solution of naphthalene and biphenyl in PG and MC have been calculated by Eq. (3). To verify our method, two additional measurements were conducted for the mixture of biphenyl with tetrahydrofuran dissolved in acetone and pure tetrahydrofuran dissolved in acetone. From these experiments and Eq. (3), the enthalpy of solution of biphenyl in acetone was found to be 19.1 kJmol1, while a direct measurement gives 19.3 kJmol1. The results of measurements along with the standard deviations u(DsolnHA/S) are given in Tables 1–3. The standard state for liquid (at T = 298 K) solutes is the pure liquid and for solids at T = 298 K it is the pure solid substance. 3. Discussion The existing literature calorimetric data on the considered systems are limited to the excess enthalpies of mixtures of MC with cyclohexane [12–14] at several temperatures including T = 298.15 K, and the enthalpies of mixing of MC with n-hexane, n-heptane, n-octane, and n-decane at T = 323.15, 348.15, and 373.15 K [15]. An estimated value of the enthalpy of solution of
I.A. Sedov et al. / Thermochimica Acta 589 (2014) 247–251
249
Table 1 Experimental values of enthalpies of solution at infinite dilution in PG, vaporization and solvation in PG at T = 298.15 K and atmospheric pressure. Solute (A)
DsolnHA/PG/(kJmol1)
u(DsolnHA/PG)/(kJmol1)
DvapHA/(kJmol1) [8]
DsolvHA/PG/(kJmol1)
n-Hexane n-Heptane n-Octane n-Nonane n-Decane Cyclohexane Benzene Toluene Ethylbenzene o-Xylene m-Xylene p-Xylene Fluorobenzene Chlorobenzene Bromobenzene Naphthalene Biphenyl
5.8 6.5 7.3 8.0 8.4 4.6 3.0 3.51 3.90 4.3 4.15 4.01 1.6 2.04 2.68 21.4 23.2
0.7 0.6 0.8 1 0.8 0.3 0.15 0.05 0.05 0.3 0.1 0.1 0.15 0.05 0.1 1.0 1.0
31.4 36.6 41.6 46.5 51.4 33.1 33.9 38.0 42.4 42.9 42.7 42.3 34.5 40.3 43.8 72.3 82.9
25.6 30.1 34.3 38.5 43.0 28.5 30.9 34.5 38.5 38.6 38.5 38.3 32.9 38.3 41.1 50.9 59.7
cyclohexane in MC at infinite dilution is 5.9 kJmol1 using Ref. [13] and 7.0 kJmol1 using Ref. [12] against our value of 6.5 kJmol1. The obtained enthalpies of solution can be compared with the recently reported [4] enthalpies of solvation of low polar solutes in ethylene glycol (EG), which differs from both molecules of PG and MC by a single methyl group. In Figs.1 and 2, the values of DsolvHA/PG and DsolvHA/MC are plotted against the values ofDsolvHA/EG. Aliphatic solutes are denoted with triangles and aromatic solutes are circles. A straight line corresponds to equation y = x. From the graphs, it is obvious that the enthalpies of solvation in MC and EG or in PG and EG are, in general, intercorrelated. When we compare solutions in MC with solutions in EG, we can write a single good correlation for both aliphatic and aromatic solutes:
Dsolv HA=MC ¼ 0:96Dsolv HA=EG 2:95 1
n ¼ 17; s ¼ 0:72 kJmol
(4),
; R2 ¼ 0:9945:
Here and below, n corresponds to the number of data points (solutes), s is the root mean square deviation, and R2 is the squared correlation coefficient. On the other hand, the relationship between the enthalpies of solution in PG and EG is better to describe in terms of two separate correlations for aliphatic (Eq. (5)) and aromatic (including
alkylaromatic, Eq. (6)) compounds:
Dsolv HA=PG ¼ 1:10Dsolv HA=EG 0:27 n ¼ 5; s ¼ 0:09 kJmol
1
(5),
; R2 ¼ 0:9996:
Dsolv HA=PG ¼ 0:84Dsolv HA=EG 5:14 n ¼ 11; s ¼ 0:65 kJmol
1
(6),
; R2 ¼ 0:9932:
This fact may seem unexpected because PG is “more similar” to EG in the sense that both solvents are 1,2-diols and MC is not. In addition, many physical properties of EG such as dielectric constant, enthalpy of vaporization, Hildebrand solubility parameter or numerous other solubility parameters [16] have closer values to those of PG than of MC. However, a similar comparison of the enthalpies of solvation for linear alkanes and aromatic hydrocarbons in methanol and ethanol (Fig. 3), which are homologues like PG and EG, shows that there are also two separate correlations for aliphatic and aromatic compounds [17]. (A straight line corresponds to equation y = x.) In general, the enthalpy of solution of an alkane grows up with the polarity of a solvent. This is not due to polar solute-solvent interactions but to solvent–solvent interactions responsible for
Table 2 Experimental values of enthalpies of solution at infinite dilution in MC, vaporization and solvation in MC at T = 298.15 K and atmospheric pressure. Solute (A)
DsolnHA/MC/(kJmol1)
u(DsolnHA/MC)/(kJmol1)
DvapHA/(kJmol1) [8]
DsolvHA/MC/(kJmol1)
n-Hexane n-Heptane n-Octane n-Nonane n-Decane n-Undecane n-Dodecane Cyclohexane Methylcyclohexane Benzene Toluene Ethylbenzene o-Xylene m-Xylene p-Xylene Fluorobenzene Chlorobenzene Bromobenzene Naphthalene Biphenyl
6.8 7.9 8.8 9.8 10.8 11.8 12.9 6.5 6.75 0.67 1.22 1.88 2.05 2.27 2.03 –1.36 0.99 0.70 16.7 18.7
0.5 0.2 0.15 0.1 0.2 0.2 0.2 0.3 0.1 0.03 0.05 0.05 0.05 0.05 0.05 0.03 0.1 0.03 1.0 1.0
31.4 36.6 41.6 46.5 51.4 56.4 62.1 33.1 35.1 33.9 38.0 42.4 42.9 42.7 42.3 34.5 40.3 43.8 72.3 82.9
24.6 28.7 32.8 36.7 40.6 44.6 49.2 26.6 28.4 33.2 36.8 40.5 40.9 40.4 40.3 35.9 41.3 44.5 55.6 64.2
250
I.A. Sedov et al. / Thermochimica Acta 589 (2014) 247–251 Table 3 Experimental values of enthalpies of solution at infinite dilution at T = 298.15 K and atmospheric pressure used to make calculations and comparisons. Solute (A)
Solvent (S)
DsolnHA/S/(kJmol1)
u(DsolnHA/S)/(kJmol1)
Tetrahydrofuran Tetrahydrofuran Tetrahydrofuran Tetrahydrofuran Naphthalene Biphenyl Biphenyl
Propylene glycol Methyl cellosolve Cyclohexane Acetone Cyclohexane Cyclohexane Acetone
1.57 1.20 3.31 0.62 22.7 23.8 19.3
0.03 0.02 0.05 0.05 0.2 0.2 0.3
the increase of the cost of formation of a cavity in a bulk solvent to accommodate the solute molecule. Addition of methyl groups to the solvent molecule usually decreases the enthalpies of solution of alkanes in this solvent. This fact is clearly seen when one considers the enthalpies of solvation of alkanes in many different solvents [18,19]. In our case, the enthalpies of solution of alkanes in PG and MC are also lower than in EG and lie below y = x in Figs. 1 and 2. In contrast with alkanes, aromatic molecules are polarizable, and their favorable interactions with polar solvents decrease the enthalpy of solvation, while the contribution due to the cavity formation still increases with growing solvent polarity. Thus, there is no single tendency connecting the enthalpy of solvation of aromatic solvents with the solvent polarity [20]. As seen from the Figs. 1 and 2, the intermolecular interactions of aromatics with PG turn out to be more favorable than with MC and EG. The above can be translated into the language of correlations of the enthalpies of solvation in PG, MC, and EG with solute parameters reflecting molecular size and polarizability. An example of such parameters are the characteristic molecular volume Vx reflecting molecular size and polarity/polarizability parameter pH 2 [21]. The following correlations are obtained:
Dsolv HA=PG ¼ 28:53V x 22:04pH2 þ 0:75
(7),
Dsolv HA=EG ¼ 28:76V x 29:44pH2 þ 4:22 1
n ¼ 17; s ¼ 1:16 kJmol
Dsolv HA=PG ¼ 29:65V x 22:44ph2 þ 2:11 1
n ¼ 17; s ¼ 0:99 kJmol
1
n ¼ 20; s ¼ 0:71 kJmol
(10),
:
Dsolv HA=MC ¼ 28:32V x 28:87pH2 þ 2:11
(11),
:
Dsolv HA=EG ¼ 26:88V x 29:04pH2 þ 2:11 n ¼ 17; s ¼ 1:21 kJmol
; R2 ¼ 0:9855:
; R2 ¼ 0:9869:
The equations for MC and EG have close coefficients before Vx H and pH 2 , while for PG the coefficient before p2 is lower by absolute value leading to less negative enthalpies of solvation of aromatic solutes in PG (for alkanes, pH 2 = 0). To compare the contributions of cavity formation in different solvents, it is more correct to fix the intercept for all three solvents at same value, e.g., the average from Eqs. (7–9) (2.11), and obtain three new correlations:
1
1
n ¼ 17; s ¼ 0:96 kJmol
(9),
(12)
:
(8),
The coefficient before pH 2 is still significantly different for PG, while the magnitude of the term proportional to Vx and reflecting the enthalpy of cavity formation, now decreases in the order EG > MC > PG, causing the enthalpies of solvation of alkanes to decrease in the same order.
Fig. 1. Comparison of enthalpies of solvation in PG for low polar aliphatic (triangles) and aromatic (circles) solutes with the enthalpies of solvation for the same compounds in EG at T = 298 K.
Fig. 2. Comparison of enthalpies of solvation in MC for low polar aliphatic (triangles) and aromatic (circles) solutes with the enthalpies of solvation for the same compounds in EG at T = 298 K.
Dsolv HA=MC ¼ 27:78V x 28:56pH2 þ 1:37 1
n ¼ 20; s ¼ 0:70 kJmol
; R2 ¼ 0:9845:
I.A. Sedov et al. / Thermochimica Acta 589 (2014) 247–251
251
these solvents on the basis of comparison with the values of the Gibbs free energies of solvation. These effects do not significantly affect the enthalpies of solvation in PG, MC or EG as well as in monohydric alcohols, but can significantly decrease the entropies and increase the Gibbs free energies. Acknowledgments This work is performed according to the Russian Government Program of Competitive Growth of Kazan Federal University and supported by RFBR (project No. 14-03-31990). References
Fig. 3. Comparison of enthalpies of solvation in ethanol vs methanol for low polar aliphatic(triangles)andaromatic(circles)solutesatT = 298 K.DatatakenfromRef.[17].
4. Conclusion Calorimetric measurement of enthalpies of solution for low polar compounds in self-associating liquids is complicated by their low solubility and low speed of dissolution. Alkanes have the worst solubility and their enthalpies of solution are measured with the largest uncertainty, while for liquid aromatic solutes the reproducibilityisverygood.Solidcompoundshaveaverylongtimeofcomplete dissolution making direct determination of the enthalpy problematic. Thus, we used a method of indirect measurement by dissolving a liquid mixture of a studied compound with some other solvent. Thermodynamic functions of solution and solvation are useful for the development of predictive models to describe thermodynamic properties of mixtures. The measured enthalpies of solution in PG and MC will also be used to study the solvophobic effects in
[1] Noyes Data Corp, Industrial Solvents Handbook, 5th ed., Noyes Data Corp, Westwood, N.J, 1998. [2] F. Welsch, Toxicol. Lett. 156 (2005) 13. [3] J.R. Fowles, M.I. Banton, L.H. Pottenger, Crit. Rev. Toxicol. 43 (2013) 363. [4] I.A. Sedov, M.A. Stolov, B.N. Solomonov, Fluid Phase Equilib. 354 (2013) 95. [5] I.A. Sedov, M.A. Stolov, B.N. Solomonov, J. Phys. Org. Chem. 24 (2011) 1088. [6] I.A. Sedov, M.A. Stolov, B.N. Solomonov, J. Chem. Thermodyn. 64 (2013) 120. [7] L.M.N.B.F. Santos, M.T. Silva, B. Schröder, L. Gomes, J. Therm. Anal. Calorim. 89 (2007) 175. [8] W. Acree, J.S. Chickos, J. Phys. Chem. Ref. Data 39 (2010) 043101. [9] USEPA, Estimation Programs Interface SuiteTM for Microsoft1 Windows, United States Environmental Protection Agency, Washington, D.C., USA, 2011. [10] F. Oprea, Bul. Univ. Pet. -Gaze Ser. Teh. (2008) 18. [11] J. Lara, L. Avdikian, G. Perron, J.E. Desnoyers, J. Solut. Chem. 10 (1981) 301. [12] G. Conti, J. Chem. Thermodyn. 30 (1998) 855. [13] M. Nishimoto, S. Tabata, K. Tamura, S. Murakami, Fluid Phase Equilib. 136 (1997) 235. [14] J. Valero, M. Gracia, C. Gutierrez Losa, J. Chem. Thermodyn. 11 (1979) 1101. [15] W. Schulz, R.N. Lichtenthaler, Thermochim. Acta 151 (1989) 261. [16] A.F.M. Barton, CRC Handbook of Solubility Parameters and Other Cohesion Parameters, 2nd ed., CRC Press, Boca Raton, 1991. [17] C. Mintz, T. Ladlie, K. Burton, M. Clark, W.E. Acree, M.H. Abraham, Qsar Comb. Sci. 27 (2008) 627. [18] R. Fuchs, W.K. Stephenson, Can. J. Chem. 63 (1985) 349. [19] B.N. Solomonov, V.B. Novikov, M.A. Varfolomeev, N.M. Mileshko, J. Phys. Org. Chem. 18 (2005) 49. [20] W.K. Stephenson, R. Fuchs, Can. J. Chem. 63 (1985) 2529. [21] M.H. Abraham, J. Andonian-Haftvan, G.S. Whiting, A. Leo, R.S. Taft, J. Chem. Soc. Perkin Trans. 2 (1994) 1777.
Contents lists available at ScienceDirect
Thermochimica Acta journal homepage: www.elsevier.com/locate/tca
Calorimetric study of solvation of low polar solutes in propylene glycol and methyl cellosolve at 298 K I.A. Sedov *, M.A. Stolov, B.N. Solomonov Chemical Institute, Kazan Federal University, Kremlevskaya 18, Kazan 420008, Russia
A R T I C L E I N F O
A B S T R A C T
Article history: Received 3 February 2014 Received in revised form 31 May 2014 Accepted 3 June 2014 Available online 5 June 2014
Enthalpies of solution of low-polar substances: aliphatic and aromatic hydrocarbons, including alkanes, cycloalkanes, alkylbenzenes, and halobenzenes in two solvents, propylene glycol and methyl cellosolve, were measured at temperature T = 298.15 K using titration calorimetry. In addition, the enthalpies of solution of poorly soluble solid solutes, naphthalene and biphenyl, in these solvents were determined using indirect method by measuring the enthalpies of solution of liquid mixtures containing them. The enthalpies of solvation from the gas phase were calculated from these data. The enthalpies of solvation in propylene glycol and methyl cellosolve and previously reported enthalpies of solvation in ethylene glycol were correlated with each other and with molecular parameters of solutes. ã 2014 Elsevier B.V. All rights reserved.
Keywords: Propylene glycol Methyl cellosolve Ethylene glycol Enthalpy of solution Solvation Solution calorimetry
1. Introduction Methyl cellosolve (2-methoxyethanol, CH3OCH2CH2OH, further MC) and propylene glycol (1,2-propanediol, CH3CH(OH)CH2OH, further PG) are two structural isomers and two important industrial solvents [1]. Methyl cellosolve is used as a solvent in dye and rubber industries, production of cellulose acetate films and laboratory organic synthesis. MC is a toxic compound which can be converted by alcohol dehydrogenase into methoxyacetic acid which can enter the Krebs cycle [2]. In contrast, PG has a very low toxicity for humans.[3] Thus, it is used as a solvent for pharmaceuticals and in food industry, as well as in rubber and polymer production, and as a component of automotive antifreeze. Both MC and PG molecules differ by a methyl group from a molecule of another common industrial solvent, ethylene glycol (1,2-ethanediol, HOCH2CH2OH, further EG). The solvent effect on equilibria in chemical reactions, solubility, interfacial distribution properties of dissolved compounds is governed by the Gibbs free energy change. The enthalpy of solvation (transfer of solute from the gas phase into solvent) determines the dependence of the Gibbs energy of solvation on temperature according to Van't Hoff equation. Despite a broad practical use of PG and MC, thermodynamic functions of solvation or solution in these solvents are almost not studied.
* Corresponding author. Tel.: +7 9600503916; fax: +7 8432315346. E-mail address: [email protected] (I.A. Sedov). http://dx.doi.org/10.1016/j.tca.2014.06.003 0040-6031/ ã 2014 Elsevier B.V. All rights reserved.
Our special interest is the solvophobic effect in self-associated solvents and its influence on the solvation properties. Both MC and PG are self-associated due to intermolecular hydrogen bonds. In our previous papers [4–6] we developed a methodology to describe the solvophobic effect in various solvents on the basis of the values of Gibbs free energy and enthalpy of solvation for lowpolar solutes. Thus, in this work we have selected a number of such solutes to determine their enthalpies of solution in PG and MC. Calorimetric measurement of the enthalpies of solution for non-polar solutes in PG is complicated by their low solubility, low speed of dissolution, and a high viscosity of solvent. No values of the enthalpies of solution of hydrocarbons and their halogenated derivatives in PG were found in literature. The solubility of hydrocarbons in MC is higher, but the calorimetric data for their solutions are also very limited. In the present work, novel thermodynamic data for solutions of hydrocarbons and their halogenated derivatives in PG and MC at infinite dilution and T = 298 K are obtained using titration calorimetry technique. 2. Experimental 2.1. Materials and methods Propylene glycol and methyl cellosolve with purity >0.99 were purchased from Acros Organics. All the solutes were at least 0.99 pure grade from Sigma–Aldrich, Acros and Fluka. They have been used without further purification. The absence of significant
248
I.A. Sedov et al. / Thermochimica Acta 589 (2014) 247–251
amounts of impurities has been confirmed by gas chromatography. 2.2. Enthalpy of solution measurements Enthalpies of solution at infinite dilution at temperature T = 298.15 K and pressure p = 0.1 MPa have been determined using TAM III precision solution calorimeter in 100 ml glass calorimetric vessel equipped with a gold stirrer, a Joule heater and a thermistor. The calorimeter operates under semi-adiabatic condition. The vessel is placed inside a thermostatted metal cylinder, and the temperature change inside the vessel is measured. The vessel exchanges heat with the thermostat in agreement with Newton's law of cooling, which leads to the exponential dependence of the vessel's temperature from time, if no heat is evolved or absorbed due to any physical or chemical processes. In experiments, the heat effects were calculated from temperature vs time curves using Regnault–Pfaundler method [7], which takes into account exponential character of the baseline. This method is integrated into SolCal software used to operate the calorimeter. For liquid solutes, calorimetric titration method was used. The vessel was initially filled with 100 ml of pure solvent. After thermostatting and heater calibrations, 50–100 ml of solute was added from an electronically operated microsyringe in 5–20 ml portions. The heat effect of each addition was determined from a calorimetric curve. For every solute, the experiments were repeated 2–3 times with a fresh portion of solvent. An average value of enthalpy of solution from all dissolution curves was taken. If a solute is volatile and poorly soluble, we should make a correction in order to account for the heat loss due to evaporation of a part of the solute molecules to a free volume of the calorimetric vessel. The standard molar enthalpy of solution obtained from calorimetric curves DHobs is equal to the sum of heat effects of dissolution and evaporation of some part of a solute:
DHobs ¼ ð1 ’ÞDsoln HA=S þ ’Dvap HA
(1),
is the molar enthalpy of solution of solute A in where DsolnH solvent S, DvapHA is the enthalpy of evaporation of A, ’ is the molar fraction of solute in the vapor phase that can be calculated by equation: A/S
1 A=S 1 þ RTnS = pAsat g 1 V free
’¼
(2)
Here pAsat is the saturated vapor pressure of pure substance A,
g A=S 1 is the limiting activity coefficient of A dissolved in S, Vfree is the volume of the free space in the calorimetric vessel, vS – number of
moles of solvent in the vessel. However, in all the cases except for saturated hydrocarbons dissolved in PG, the contribution of evaporation process to the measured enthalpies of solution is very small. For example, even for volatile hexane in MC the estimated value of ’DvapHA is only 0.05 kJmol1, which is less than the uncertainty of the experiment. Thus, we assumed that DHobs=DsolnHA/S. For alkanes and cyclohexane in PG, which have a poor solubility and low liquid–gas distribution coefficient, we calculated the corrected values of DsolnHA/S by Eq. (1) using literature values of DvapHA [8], pAsat [9], and the values of g A=S 1 estimated from the solubilities reported in literature [10]. Solid solutes were placed in sealed glass ampoules, which were then broken inside a calorimetric vessel with the solvent. The heat effect of dissolution following ampoule break was measured. However, the process of complete dissolution took up to several hours. The results were very sensitive to the method of calculation since the baseline is not absolutely stable on such a long timescale. The values of DsolnHA/S can differ by up to 3–4 kJmol1 depending
on which moment is thought to be the end of dissolution and which part of the baseline after dissolution is chosen. To obtain more accurate results, we developed another experimental method. Given that we want to find the molar enthalpy of solution DsolnHA/S of a solid compound A in a solvent S, but A has a poor solubility in S. We choose one of the solvents (let us denote it by B) in which compound A has the best solubility and prepare a concentrated (but not saturated) solution (A + B) in this solvent with the molar fraction of A xA. It is important for B to be relatively well soluble in S, otherwise the results can be inaccurate. Then we measure the molar enthalpy of solution of the mixture (A + B) by means of titration calorimetry in solvent S (DsolnHA+B/S) and in another (reference) solvent R (DsolnHA+B/R). The molar enthalpies of solution for compound B must be measured separately in both solvents S and R (DsolnHB/S and DsolnHB/R). If all solutions are very diluted and we can neglect the interactions between the molecules of A and B, then the following relationship is correct (due to the fact that enthalpy is a function of state): 0
Dsoln HA=S ¼
1
B C C ðAþBÞ=S Dsoln HA=R þ B Dsoln HðAþBÞ=R BDsoln H C @ A B=S B=R ð1 xA Þ Dsoln H Dsoln H xA (3).
Here the enthalpy of solution of solid A in solvent R, DsolnHA/R must be known or measured. If we choose a solvent in which A has a good solubility, then the value of DsolnHA/R can be measured with a good precision. It is also important to use a mixture with a high value of xA, because the uncertainties of four measured enthalpies of solution are divided by xA making the error of DsolnHA/S inversely proportional to it. This approach resembles that used by Desnoyers et al. [11] for a liquid solute (benzene) to determine its enthalpy of solution in water-rich binary aqueous-organic solvent. In our experiments, we prepared the mixtures of naphthalene and biphenyl with tetrahydrofuran with the molar fraction of a solute around 0.2. The enthalpies of solution of tetrahydrofuran in PG, MC, and cyclohexane were measured and found to be 1.57, 1.20, and 3.31 kJmol1 respectively. Solid aromatic hydrocarbons rapidly dissolve in other hydrocarbons, and cyclohexane has been chosen as (reference) solvent R. For both solid naphthalene and biphenyl, the enthalpies of solution in cyclohexane were measured (22.7 and 23.8 kJmol1 respectively). The enthalpies of solution of the liquid mixtures in PG, MC, and cyclohexane were also measured, and the enthalpies of solution of naphthalene and biphenyl in PG and MC have been calculated by Eq. (3). To verify our method, two additional measurements were conducted for the mixture of biphenyl with tetrahydrofuran dissolved in acetone and pure tetrahydrofuran dissolved in acetone. From these experiments and Eq. (3), the enthalpy of solution of biphenyl in acetone was found to be 19.1 kJmol1, while a direct measurement gives 19.3 kJmol1. The results of measurements along with the standard deviations u(DsolnHA/S) are given in Tables 1–3. The standard state for liquid (at T = 298 K) solutes is the pure liquid and for solids at T = 298 K it is the pure solid substance. 3. Discussion The existing literature calorimetric data on the considered systems are limited to the excess enthalpies of mixtures of MC with cyclohexane [12–14] at several temperatures including T = 298.15 K, and the enthalpies of mixing of MC with n-hexane, n-heptane, n-octane, and n-decane at T = 323.15, 348.15, and 373.15 K [15]. An estimated value of the enthalpy of solution of
I.A. Sedov et al. / Thermochimica Acta 589 (2014) 247–251
249
Table 1 Experimental values of enthalpies of solution at infinite dilution in PG, vaporization and solvation in PG at T = 298.15 K and atmospheric pressure. Solute (A)
DsolnHA/PG/(kJmol1)
u(DsolnHA/PG)/(kJmol1)
DvapHA/(kJmol1) [8]
DsolvHA/PG/(kJmol1)
n-Hexane n-Heptane n-Octane n-Nonane n-Decane Cyclohexane Benzene Toluene Ethylbenzene o-Xylene m-Xylene p-Xylene Fluorobenzene Chlorobenzene Bromobenzene Naphthalene Biphenyl
5.8 6.5 7.3 8.0 8.4 4.6 3.0 3.51 3.90 4.3 4.15 4.01 1.6 2.04 2.68 21.4 23.2
0.7 0.6 0.8 1 0.8 0.3 0.15 0.05 0.05 0.3 0.1 0.1 0.15 0.05 0.1 1.0 1.0
31.4 36.6 41.6 46.5 51.4 33.1 33.9 38.0 42.4 42.9 42.7 42.3 34.5 40.3 43.8 72.3 82.9
25.6 30.1 34.3 38.5 43.0 28.5 30.9 34.5 38.5 38.6 38.5 38.3 32.9 38.3 41.1 50.9 59.7
cyclohexane in MC at infinite dilution is 5.9 kJmol1 using Ref. [13] and 7.0 kJmol1 using Ref. [12] against our value of 6.5 kJmol1. The obtained enthalpies of solution can be compared with the recently reported [4] enthalpies of solvation of low polar solutes in ethylene glycol (EG), which differs from both molecules of PG and MC by a single methyl group. In Figs.1 and 2, the values of DsolvHA/PG and DsolvHA/MC are plotted against the values ofDsolvHA/EG. Aliphatic solutes are denoted with triangles and aromatic solutes are circles. A straight line corresponds to equation y = x. From the graphs, it is obvious that the enthalpies of solvation in MC and EG or in PG and EG are, in general, intercorrelated. When we compare solutions in MC with solutions in EG, we can write a single good correlation for both aliphatic and aromatic solutes:
Dsolv HA=MC ¼ 0:96Dsolv HA=EG 2:95 1
n ¼ 17; s ¼ 0:72 kJmol
(4),
; R2 ¼ 0:9945:
Here and below, n corresponds to the number of data points (solutes), s is the root mean square deviation, and R2 is the squared correlation coefficient. On the other hand, the relationship between the enthalpies of solution in PG and EG is better to describe in terms of two separate correlations for aliphatic (Eq. (5)) and aromatic (including
alkylaromatic, Eq. (6)) compounds:
Dsolv HA=PG ¼ 1:10Dsolv HA=EG 0:27 n ¼ 5; s ¼ 0:09 kJmol
1
(5),
; R2 ¼ 0:9996:
Dsolv HA=PG ¼ 0:84Dsolv HA=EG 5:14 n ¼ 11; s ¼ 0:65 kJmol
1
(6),
; R2 ¼ 0:9932:
This fact may seem unexpected because PG is “more similar” to EG in the sense that both solvents are 1,2-diols and MC is not. In addition, many physical properties of EG such as dielectric constant, enthalpy of vaporization, Hildebrand solubility parameter or numerous other solubility parameters [16] have closer values to those of PG than of MC. However, a similar comparison of the enthalpies of solvation for linear alkanes and aromatic hydrocarbons in methanol and ethanol (Fig. 3), which are homologues like PG and EG, shows that there are also two separate correlations for aliphatic and aromatic compounds [17]. (A straight line corresponds to equation y = x.) In general, the enthalpy of solution of an alkane grows up with the polarity of a solvent. This is not due to polar solute-solvent interactions but to solvent–solvent interactions responsible for
Table 2 Experimental values of enthalpies of solution at infinite dilution in MC, vaporization and solvation in MC at T = 298.15 K and atmospheric pressure. Solute (A)
DsolnHA/MC/(kJmol1)
u(DsolnHA/MC)/(kJmol1)
DvapHA/(kJmol1) [8]
DsolvHA/MC/(kJmol1)
n-Hexane n-Heptane n-Octane n-Nonane n-Decane n-Undecane n-Dodecane Cyclohexane Methylcyclohexane Benzene Toluene Ethylbenzene o-Xylene m-Xylene p-Xylene Fluorobenzene Chlorobenzene Bromobenzene Naphthalene Biphenyl
6.8 7.9 8.8 9.8 10.8 11.8 12.9 6.5 6.75 0.67 1.22 1.88 2.05 2.27 2.03 –1.36 0.99 0.70 16.7 18.7
0.5 0.2 0.15 0.1 0.2 0.2 0.2 0.3 0.1 0.03 0.05 0.05 0.05 0.05 0.05 0.03 0.1 0.03 1.0 1.0
31.4 36.6 41.6 46.5 51.4 56.4 62.1 33.1 35.1 33.9 38.0 42.4 42.9 42.7 42.3 34.5 40.3 43.8 72.3 82.9
24.6 28.7 32.8 36.7 40.6 44.6 49.2 26.6 28.4 33.2 36.8 40.5 40.9 40.4 40.3 35.9 41.3 44.5 55.6 64.2
250
I.A. Sedov et al. / Thermochimica Acta 589 (2014) 247–251 Table 3 Experimental values of enthalpies of solution at infinite dilution at T = 298.15 K and atmospheric pressure used to make calculations and comparisons. Solute (A)
Solvent (S)
DsolnHA/S/(kJmol1)
u(DsolnHA/S)/(kJmol1)
Tetrahydrofuran Tetrahydrofuran Tetrahydrofuran Tetrahydrofuran Naphthalene Biphenyl Biphenyl
Propylene glycol Methyl cellosolve Cyclohexane Acetone Cyclohexane Cyclohexane Acetone
1.57 1.20 3.31 0.62 22.7 23.8 19.3
0.03 0.02 0.05 0.05 0.2 0.2 0.3
the increase of the cost of formation of a cavity in a bulk solvent to accommodate the solute molecule. Addition of methyl groups to the solvent molecule usually decreases the enthalpies of solution of alkanes in this solvent. This fact is clearly seen when one considers the enthalpies of solvation of alkanes in many different solvents [18,19]. In our case, the enthalpies of solution of alkanes in PG and MC are also lower than in EG and lie below y = x in Figs. 1 and 2. In contrast with alkanes, aromatic molecules are polarizable, and their favorable interactions with polar solvents decrease the enthalpy of solvation, while the contribution due to the cavity formation still increases with growing solvent polarity. Thus, there is no single tendency connecting the enthalpy of solvation of aromatic solvents with the solvent polarity [20]. As seen from the Figs. 1 and 2, the intermolecular interactions of aromatics with PG turn out to be more favorable than with MC and EG. The above can be translated into the language of correlations of the enthalpies of solvation in PG, MC, and EG with solute parameters reflecting molecular size and polarizability. An example of such parameters are the characteristic molecular volume Vx reflecting molecular size and polarity/polarizability parameter pH 2 [21]. The following correlations are obtained:
Dsolv HA=PG ¼ 28:53V x 22:04pH2 þ 0:75
(7),
Dsolv HA=EG ¼ 28:76V x 29:44pH2 þ 4:22 1
n ¼ 17; s ¼ 1:16 kJmol
Dsolv HA=PG ¼ 29:65V x 22:44ph2 þ 2:11 1
n ¼ 17; s ¼ 0:99 kJmol
1
n ¼ 20; s ¼ 0:71 kJmol
(10),
:
Dsolv HA=MC ¼ 28:32V x 28:87pH2 þ 2:11
(11),
:
Dsolv HA=EG ¼ 26:88V x 29:04pH2 þ 2:11 n ¼ 17; s ¼ 1:21 kJmol
; R2 ¼ 0:9855:
; R2 ¼ 0:9869:
The equations for MC and EG have close coefficients before Vx H and pH 2 , while for PG the coefficient before p2 is lower by absolute value leading to less negative enthalpies of solvation of aromatic solutes in PG (for alkanes, pH 2 = 0). To compare the contributions of cavity formation in different solvents, it is more correct to fix the intercept for all three solvents at same value, e.g., the average from Eqs. (7–9) (2.11), and obtain three new correlations:
1
1
n ¼ 17; s ¼ 0:96 kJmol
(9),
(12)
:
(8),
The coefficient before pH 2 is still significantly different for PG, while the magnitude of the term proportional to Vx and reflecting the enthalpy of cavity formation, now decreases in the order EG > MC > PG, causing the enthalpies of solvation of alkanes to decrease in the same order.
Fig. 1. Comparison of enthalpies of solvation in PG for low polar aliphatic (triangles) and aromatic (circles) solutes with the enthalpies of solvation for the same compounds in EG at T = 298 K.
Fig. 2. Comparison of enthalpies of solvation in MC for low polar aliphatic (triangles) and aromatic (circles) solutes with the enthalpies of solvation for the same compounds in EG at T = 298 K.
Dsolv HA=MC ¼ 27:78V x 28:56pH2 þ 1:37 1
n ¼ 20; s ¼ 0:70 kJmol
; R2 ¼ 0:9845:
I.A. Sedov et al. / Thermochimica Acta 589 (2014) 247–251
251
these solvents on the basis of comparison with the values of the Gibbs free energies of solvation. These effects do not significantly affect the enthalpies of solvation in PG, MC or EG as well as in monohydric alcohols, but can significantly decrease the entropies and increase the Gibbs free energies. Acknowledgments This work is performed according to the Russian Government Program of Competitive Growth of Kazan Federal University and supported by RFBR (project No. 14-03-31990). References
Fig. 3. Comparison of enthalpies of solvation in ethanol vs methanol for low polar aliphatic(triangles)andaromatic(circles)solutesatT = 298 K.DatatakenfromRef.[17].
4. Conclusion Calorimetric measurement of enthalpies of solution for low polar compounds in self-associating liquids is complicated by their low solubility and low speed of dissolution. Alkanes have the worst solubility and their enthalpies of solution are measured with the largest uncertainty, while for liquid aromatic solutes the reproducibilityisverygood.Solidcompoundshaveaverylongtimeofcomplete dissolution making direct determination of the enthalpy problematic. Thus, we used a method of indirect measurement by dissolving a liquid mixture of a studied compound with some other solvent. Thermodynamic functions of solution and solvation are useful for the development of predictive models to describe thermodynamic properties of mixtures. The measured enthalpies of solution in PG and MC will also be used to study the solvophobic effects in
[1] Noyes Data Corp, Industrial Solvents Handbook, 5th ed., Noyes Data Corp, Westwood, N.J, 1998. [2] F. Welsch, Toxicol. Lett. 156 (2005) 13. [3] J.R. Fowles, M.I. Banton, L.H. Pottenger, Crit. Rev. Toxicol. 43 (2013) 363. [4] I.A. Sedov, M.A. Stolov, B.N. Solomonov, Fluid Phase Equilib. 354 (2013) 95. [5] I.A. Sedov, M.A. Stolov, B.N. Solomonov, J. Phys. Org. Chem. 24 (2011) 1088. [6] I.A. Sedov, M.A. Stolov, B.N. Solomonov, J. Chem. Thermodyn. 64 (2013) 120. [7] L.M.N.B.F. Santos, M.T. Silva, B. Schröder, L. Gomes, J. Therm. Anal. Calorim. 89 (2007) 175. [8] W. Acree, J.S. Chickos, J. Phys. Chem. Ref. Data 39 (2010) 043101. [9] USEPA, Estimation Programs Interface SuiteTM for Microsoft1 Windows, United States Environmental Protection Agency, Washington, D.C., USA, 2011. [10] F. Oprea, Bul. Univ. Pet. -Gaze Ser. Teh. (2008) 18. [11] J. Lara, L. Avdikian, G. Perron, J.E. Desnoyers, J. Solut. Chem. 10 (1981) 301. [12] G. Conti, J. Chem. Thermodyn. 30 (1998) 855. [13] M. Nishimoto, S. Tabata, K. Tamura, S. Murakami, Fluid Phase Equilib. 136 (1997) 235. [14] J. Valero, M. Gracia, C. Gutierrez Losa, J. Chem. Thermodyn. 11 (1979) 1101. [15] W. Schulz, R.N. Lichtenthaler, Thermochim. Acta 151 (1989) 261. [16] A.F.M. Barton, CRC Handbook of Solubility Parameters and Other Cohesion Parameters, 2nd ed., CRC Press, Boca Raton, 1991. [17] C. Mintz, T. Ladlie, K. Burton, M. Clark, W.E. Acree, M.H. Abraham, Qsar Comb. Sci. 27 (2008) 627. [18] R. Fuchs, W.K. Stephenson, Can. J. Chem. 63 (1985) 349. [19] B.N. Solomonov, V.B. Novikov, M.A. Varfolomeev, N.M. Mileshko, J. Phys. Org. Chem. 18 (2005) 49. [20] W.K. Stephenson, R. Fuchs, Can. J. Chem. 63 (1985) 2529. [21] M.H. Abraham, J. Andonian-Haftvan, G.S. Whiting, A. Leo, R.S. Taft, J. Chem. Soc. Perkin Trans. 2 (1994) 1777.