Azeotropic and Mixing H Data for Binaries With Dietoxymethane

Fluid Phase Equilibria 191 (2001) 99–109

Azeotropic and heats of mixing data for variousbinary systems with diethoxymethane

Dana Constantinescua, Roland Wittigb, Jurgen Gmehlingb,∗a Institute of Physical Chemistry, I.G. Murgulescu, Romanian Academy,

Splaiul Independentei 202, Bucharest 77208, Romaniab Universitat Oldenburg, Technische Chemie, Postfach 2503, D-26111 Oldenburg, Germany

Received 11 June 2001; accepted 10 August 2001

Abstract

Reliable azeotropic data have been measured for the binary systems ethanol+ diethoxymethane, diethoxy-methane+ 1-propanol, diethoxymethane+ heptane, and cyclohexane+ diethoxymethane with the help of a wireband column. Additionally molar excess enthalpies (HE) for the binary mixtures of diethoxymethane with hexane,1-octene, methanol and ethanol at 363.15 K have been determined with an isothermal flow calorimeter. The ex-perimental data were compared with the results using the available “ether” main group of the group contributionmethod Modified UNIFAC (Dortmund), which is a well-known model for the prediction of phase equilibria andexcess properties. For the description of theHE data a Redlich–Kister polynomial was also used.

The aim of this work is to supplement the available data base for acetal systems required for the introduction of anew “acetal” group in the Modified UNIFAC (Dortmund) model. © 2001 Elsevier Science B.V. All rights reserved.

Keywords: Excess (properties); Vapor–liquid equilibria; Group contribution method; Modified UNIFAC (Dortmund); Wireband column; Isothermal flow calorimeter

1. Introduction

The knowledge of azeotropic points offers the most important information for the synthesis and designof distillation processes. However, the immense amount of data [1] stored in Dortmund Data Bank (morethan 45,000 entries on azeotropic and zeotropic behavior) cannot only be used for process synthesis, e.g.design of distillation columns, selection of the most suitable solvent for azeotropic distillation, but alsofor the further development of group contribution methods and for fitting reliableGE models parameters.

In the case ofGE models [2], a precondition for a good description of phase equilibria of multicomponentsystems is the determination of reliable binary interaction parameters. In chemical industry, mainly

∗ Corresponding author. Tel.:+49-441-798-3831; fax:+49-411-798-3330.E-mail address: [email protected] (J. Gmehling).

0378-3812/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved.PII: S0378-3812(01)00615-X

100 D. Constantinescu et al. / Fluid Phase Equilibria 191 (2001) 99–109

vapor–liquid equilibria (VLE) data are used to fit the required binary parameters. However, as mentionedin previous papers [3,4], it is recommended to fit the required interaction parameters simultaneously toall available reliable experimental data (VLE, azeotropic data, activity coefficients at infinite dilution(γ ∞), HE, solid–liquid equilibria (SLE) of eutectic systems, etc.), since the parameters obtained by asimultaneous fit allow an improved representation of the real behavior across the whole composition anda large temperature range.

In the Modified UNIFAC method [5,6], different types of phase equilibria and excess properties areused for fitting simultaneously the required temperature-dependent group interaction parameters. Excessenthalpy data are of a particular significance as supporting data at high temperature [7] and providethe most important information about the temperature dependence, because of the direct relationshipbetween the partial molar excess enthalpy (HE

i ) and the observed temperature dependence of the activitycoefficientγ i , by means of the Gibbs–Helmholtz relation:(

∂ ln γi

∂(1/T )

)P,x

= HEi

R(1)

Although a large number of excess enthalpy data around room temperature are available,HE data at hightemperature are scarce. This paper presents excess enthalpies at high temperature.

This paper presents azeotropic data for the binary systems ethanol+ diethoxymethane, diethoxymethane+ 1-propanol, diethoxymethane+ heptane, and cyclohexane+ diethoxymethane. For these systems, alsopublished azeotropic data are available. Additionally, excess enthalpies for binary mixtures of diethoxy-methane with hexane, 1-octene, methanol and ethanol were measured. Since 95% of the publishedHE datawere measured between 283 and 323 K [8], our attention is focused to measure the required supportingHE data at higher temperatures. The measurements were performed using an isothermal flow calorimeterwith a large operational range (273–425 K). Enthalpies of mixing for hexane+ diethoxymethane [9] anddiethoxymethane+1-octene [10] are also reported in literature, but only atT = 298.15 K. The exper-imental azeotropic andHE data were compared graphically with the results predicted by the ModifiedUNIFAC model using the available “ether” main group. For the correlation of theHE data a Redlich–Kisterpolynomial was used.

Diethoxymethane is a compound belonging to the main classes of oxaalkanes. Besides the technicalimportance for these classes of substances there is also an interest to examine, in terms of ModifiedUNIFAC model, vapor–liquid equilibria and related thermodynamic excess functions. The intention ofthis paper is to supplement the available azeotropic andHE data for acetal systems required for theintroduction of a “acetal” group in the Modified UNIFAC model.

2. Experimental section

2.1. Materials

Chemicals with high purity obtained from various suppliers prior to their use were purified by vac-uum distillation. The purity of the compounds was carefully checked by gas chromatography (GC) andKarl–Fischer titration. Besides the supplier, the chemicals together with the pure compound specificationare summarized in Table 1.

D. Constantinescu et al. / Fluid Phase Equilibria 191 (2001) 99–109 101

Table 1Suppliers, purities, water contents of the chemicals used

Compound Supplier Purity (% GC) Water content(mass ppm)

Hexane Fluka >99.9 40Heptane Merck >99.5 50Cyclohexane Scharlau >99.9 101-Octene Acros >99.8 70Methanol Scharlau >99.9 30Ethanol Merck >99.9 301-Propanol Riedel de Haen >99.8 50Diethoxymethane Acros >99.5 10

2.2. Measurement of the azeotropic points

A commercially available wire band column with an electronically controlled reflux ratio (supplier:Normag, Hofheim, Germany) was used to distil the mixtures, as described earlier [1]. The reflux ratiois realized on the basis of a vapor dividing principle. For pressures below atmospheric pressure, thedesired pressure is kept constant with the help of a vacuum pump. The temperature is determined usinga resistance thermometer with an accuracy of±0.1 K and the pressure by means of a sensor (DruckLimited, type PDCR) with an accuracy of±0.05 kPa. The accuracy of the azeotropic compositionyaz

determined is approximately±0.2 mol%.Approximately 30 cm3 of the binary mixture with estimated azeotropic composition are distilled at

constant pressure and a small pressure drop at nearly total reflux for approximately 60 min. Then thecomposition of the distillate is determined analytically by GC. For all systems investigated homogeneouspressure maximum azeotropes were obtained. To verify whether the system shows azeotropic (separationfactor α12 = 1) and not quasi-azeotropic (α12 ≈ 1) behavior, the experiments were always repeatedstarting with a different feed composition.

2.3. HE measurements

A commercially available isothermal flow calorimeter from Hart Scientific (model 7501) was usedfor the excess enthalpy measurements [11]. The calorimeter consists of two solvents pumps (ISCO, LC2600, 260 cm3), a thermostated flow cell, a back pressure regulator, which prevents evaporation (it isknown [12] that vaporization effects may cause considerable errors at higher temperatures) and providesthe opportunity to measureHE data also at higher pressures (up to 15 MPa). The flow cell, containing apulsed heater, a calibration heater, a Peltier cooler and a mixing tube wound around a copper cylinder[13], is located in a stainless steel cylinder which is immersed in a thermostat.

The power per pulse can be varied between 0.05 and 20�J. The exact energy per pulse is obtainedby calibration using the energy dissipated from a precise resistor fixed at the cylinder of the flow cell.The temperature of the flow cell is maintained constant by adjusting the frequency of the pulsed heaterto balance the cooling from the Peltier cooler and the heat effect. The temperature of the liquid pumpsand the thermostat are monitored with Hart Scientific platinum resistance thermometers (model 1006Micro-Therm) with an accuracy of±0.005 K.


From the recorded frequency change of the pulsed heater, the flow rate, the pure component densitiesand the molar mass of the components the molar excess enthalpies can be calculated using the energyevolved per pulse. The uncertainty inHE is estimated to be less than±1%, as shown previously [11].Studies of test mixtures indicated that the precision of the results was better than±1% over the wholecomposition range.

3. Experimental results

The experimental azeotropic data (T, P, composition) for the investigated systems ethanol+ diethoxy-methane, diethoxymethane+ 1-propanol, diethoxymethane+ heptane, and cyclohexane+ diethoxy-methane are given in Table 2 together with the data from literature [14–20]. In all cases homogeneouspressure maximum azeotropes were observed.

The experimental data for the acetal systems were compared with the results of the group contributionmethod Modified UNIFAC (Dortmund) using the “ether” group. The results are shown in Figs. 1–4 in theform of T−yaz-diagrams. As can be seen, there is a good agreement between experiment and prediction

Table 2Experimental azeotropic data for different acetal systems

System Experimental data T (K) Published data Reference

P (kPa) y1 P (kPa) y1

Ethanol (1)+ diethoxymethane (2) 347.35 101.32 0.6081 [14]347.35 101.32 0.6208 [15]

102.35 0.6497 348.3095.99 0.6350 346.8552.51 0.5765 331.0027.48 0.5123 315.9520.02 0.4700 309.25

Diethoxymethane (1)+ 1-propanol (2) 359.85 101.32 0.7800 [16]359.30 99.67 0.8236 [17]

99.43 0.7680 359.0156.27 0.8367 342.2916.82 0.9356 311.65

Diethoxymethane (1)+ heptane (2) 102.30 0.9101 361.27360.95 101.32 [18]343.15 56.95 0.9100 [19]

35.66 0.8830 333.03323.15 27.10 0.8950 [19]

11.21 0.8476 302.53

Cyclohexane (1)+ diethoxymethane (2) 353.25 101.32 0.8580 [16]101.39 0.8226 353.2144.33 0.8366 328.0714.97 0.8623 300.83

298.15 0.9000 [20]


Fig. 1. Experimental and predicted azeotropic data for the binary system ethanol(1)+ diethoxymethane (2): (�) from [14]; (�)from [15]; (�) our data; (—) Modified UNIFAC (Dortmund).

Fig. 2. Experimental and predicted azeotropic data for the binary system diethoxymethane(1) + 1-propanol (2): (�) from [16];(�) from [17]; (�) our data; (—) Modified UNIFAC (Dortmund).

Fig. 3. Experimental and predicted azeotropic data for the binary system diethoxymethane(1)+heptane (2): (�) from [19]; (�)our data; (—) Modified UNIFAC (Dortmund).


Fig. 4. Experimental and predicted azeotropic data for the binary system cyclohexane(1)+diethoxymethane (2): (�) from [16];(�) from [20]; (�) our data; (—) Modified UNIFAC (Dortmund).

for ethanol+ diethoxymethane, reasonable agreement for diethoxymethane+ heptane, diethoxymethane+ 1-propanol and poor agreement for cyclohexane+ diethoxymethane.

The Clausius–Clapeyron and Gibbs–Helmholtz equation give the alteration of the azeotropic compo-sition with temperature. In most cases, a correct temperature dependence is predicted, which in majorityof cases depend mainly on the slope of the vapor pressure data and only to a smaller extend on thetemperature dependence of the activity coefficient (partial molar excess enthalpies) [1].

However, to improve the results of Modified UNIFAC (Dortmund) additional measurements for thedetermination of parameters for a new “acetal” group fitted simultaneously to VLE,HE, γ ∞, SLE ofeutectic systems, etc. are required.

Therefore, also heat of mixing measurements for systems with diethoxymethane were performed. Themeasured excess enthalpies for the binary systems of diethoxymethane with hexane, 1-octene, methanoland ethanol at 363.15 K are listed in Table 3. The experimental data have been compared with the resultsof the group contribution method Modified UNIFAC (Dortmund) using the “ether” group. Additionally,the Redlich–Kister [21] expansion was used to correlate the results:

HE

x1x2=

m∑j

Aj (2xj − 1)j−1 (2)

using the following objective function [22]:

F =n∑i

[(HE

x1x2

)exp

−(

HE

x1x2

)calc

]2

(3)

The fitted Redlich–Kister parametersAj together with the root mean square deviations (RMSD) are givenin Table 4. The number of parameters used to represent the experimental data depends on the molecularcomplexity of theHE behavior, the quality of the data, and the number of data points available. The reliableand extensive data justify the use of five parameters. For all studied systems, the obtained RMSDs arewithin the estimated experimental error of±1% [12].


Table 3Excess molar enthalpies for mixtures containing diethoxymethane at 363.15 K

x1 HE (J mol−1) x1 HE (J mol−1) x1 HE (J mol−1)

Hexane (1)+ diethoxymethane (2),P = 1.31 MPa0.0479 91.8 0.3891 490.3 0.7413 393.10.0960 172.6 0.4387 506.5 0.7926 337.30.1443 249.9 0.4886 514.9 0.8441 269.90.1928 313.1 0.5387 511.6 0.8958 187.50.2415 370.1 0.5890 500.5 0.9478 95.50.2905 415.4 0.6395 470.30.3397 458.6 0.6900 437.4

Diethoxymethane (1)+ 1-octene (2),P = 1.34 MPa0.0620 54.3 0.4557 313.3 0.7903 216.90.1225 113.9 0.5068 319.8 0.8340 182.30.1814 167.7 0.5567 320.2 0.8768 141.90.2390 209.3 0.6055 312.8 0.9187 98.50.2951 246.9 0.6533 299.2 0.9598 51.40.3499 279.4 0.6999 279.70.4034 302.2 0.7456 254.3

Methanol (1)+ diethoxymethane (2),P = 1.34 MPa0.1397 828.1 0.6729 1171.0 0.9250 426.00.2553 1233.7 0.7163 1063.0 0.9459 240.90.3525 1409.8 0.7552 958.2 0.9652 153.30.4354 1451.9 0.7904 845.2 0.9832 69.90.5070 1423.4 0.8223 737.00.5694 1356.3 0.8514 629.70.6242 1271.9 0.9025 524.7

Ethanol (1)+ diethoxymethane (2),P = 1.34 MPa0.1013 710.9 0.5881 1563.3 0.8653 692.40.1922 1166.9 0.6367 1471.5 0.8955 550.70.2743 1445.7 0.6817 1367.8 0.9239 407.40.3487 1598.8 0.7236 1244.0 0.9507 266.10.4166 1662.6 0.7626 1113.4 0.9760 123.00.4786 1672.8 0.7991 980.00.5356 1633.6 0.8333 837.0

Table 4Redlich–Kister-parameters

Component 1 Component 2 T (K) A1 A2 A3 A4 A5 RMSDa

(J mol−1)

Hexane Diethoxymethane 363.15 2055.34 −56.24 −36.96 105.02 −79.33 0.734Diethoxymethane 1-Octene 363.15 1283.52−124.78 −127.82 −113.62 −83.99 1.078Methanol Diethoxymethane 363.15 5686.34 1376.37 1101.35−107.55 −1258.72 1.155Ethanol Diethoxymethane 363.15 6615.09 1102.18 963.79 199.58−1049.38 0.998

a RMSD =√

1n

∑n(H

Eexp − H E

calc)2.


Fig. 5. Experimental and correlated, respectively predictedHE data for hexane(1) + diethoxymethane (2) at 363.15 K (�) ourdata, and at 298.15 K (�) from [9]; (—) Redlich–Kister; (- - -) Modified UNIFAC (Dortmund).

In Figs. 5–7, the Modified UNIFAC (Dortmund) and Redlich–Kister results are compared with theexperimental results. As can be seen all systems show relatively symmetric endothermal behavior.

In Fig. 5, theHE data for the system hexane+diethoxymethane at 363.15 K are shown together withpreviously published data [9] at 298.15 K. The prediction of theHE data at 298.15 K with Modified

Fig. 6. Experimental and correlated, respectively predictedHE data for diethoxymethane(1) + 1-octene (2) at 363.15 K (�) ourdata, and at 298.15 K (�) from [10]; (—) Redlich–Kister; (- - -) Modified UNIFAC (Dortmund).


Fig. 7. Experimental and correlated, respectively predictedHE data for methanol (1) (�) and ethanol (1) (�) + diethoxymethane(2) at 363.15 K; (—) Redlich–Kister; (- - -) Modified UNIFAC (Dortmund).

UNIFAC (Dortmund) using the “ether” main group is reasonably good, but for theHE data at 363.15 Kpoor agreement is obtained.

In Fig. 6, theHE data for the system diethoxymethane+1-octene at 363.15 K are shown together withpublished data [10] at 298.15 K. With Modified UNIFAC (Dortmund) the excess enthalpies are predictedtoo low at 298.15 K, respectively too high at 363.15 K.

Fig. 7 illustrates that the excess enthalpies of the two lowest alcohols, methanol and ethanol withdiethoxymethane differ only slightly. With Modified UNIFAC (Dortmund) lowerHE values are predictedfor both data sets.

In conclusion, it can be recognized that the predicted excess molar enthalpies with Modified UNIFAC(Dortmund) using the existing group interaction parameters for the “ether” group provide poor resultsfor the investigated mixtures especially at high temperature. To overcome this problem, a new groupshould be introduced so that the proximity effect, due to the two O atoms in acetals, can be taken intoaccount.

4. Conclusions

Azeotropic data for the binary systems ethanol+ diethoxymethane, diethoxymethane+ 1-propanol,diethoxymethane+ heptane, and cyclohexane+ diethoxymethane have been measured and comparedwith the results of the group contribution method Modified UNIFAC (Dortmund).

Additionally, excess enthalpies were measured for the binary mixtures of diethoxymethane with hexane,1-octene, methanol and ethanol at 363 K. TheHE data have been compared with data using ModifiedUNIFAC (Dortmund) and correlated using the Redlich–Kister expansion.


Because of the proximity effect the predictions using the available “ether” group of the ModifiedUNIFAC model provide only poor results for the acetal systems investigated. Preliminary extensivestudies have already demonstrated this effect for oxaalkanes, including linear and cyclic acetals [23–29].To overcome this problem, a new structural group for acetals will be introduced in Modified UNIFAC(Dortmund).

The main objective within this work was to provide azeotropic andHE data for the introduction ofthe new “acetal” main group. Additional phase equilibrium measurements (VLE,γ ∞, SLE of eutecticsystems) for fitting temperature-dependent parameters will be carried out in the future.

List of symbolsA Redlich–Kister parameterF objective functionG molar Gibbs enthalpy (J mol−1)GC gas chromatographyH molar enthalpy (J mol−1)P pressure (kPa)R general gas constant (R = 8.3144 J mol−1 K−1)SLE solid–liquid equilibriaT temperature (K)VLE vapor–liquid equilibriax liquid phase compositiony vapor phase composition

Greek lettersα12 separation factorγ activity coefficient

SuperscriptsE excess property∞ at infinite dilution

Subscripts1, 2,i components 1, 2,iaz at the azeotropic pointcalc calculated valueexp experimental valueP at constant pressurex at constant mole fraction

Acknowledgements

The authors thank Mr. R. Bolts for his technical assistance in carrying out the calorimetricmeasurements.


References

[1] J. Gmehling, J. Menke, J. Krafczyk, K. Fischer, Fluid Phase Equilib. 103 (1995) 51–76.[2] J.M. Prausnitz, R.N. Lichtenthaler, E.G. de Azevedo, Molecular Thermodynamics of Fluid Phase Equilibria, Prentice Hall,

Englewood Cliffs, NJ, 1999.[3] J.R. Rarey-Nies, D. Tiltmann, J. Gmehling, Chem. Eng. Technol. 61 (1989) 407–410.[4] J. Gmehling, Pure Appl. Chem. 71 (6) (1999) 939–949.[5] U. Weidlich, J. Gmehling, Ind. Eng. Chem. Res. 26 (1987) 1372–1381.[6] J. Gmehling, J. Lohmann, A. Jakob, J. Li, R. Joh, Ind. Eng. Chem. Res. 37 (1998) 4876–4882.[7] J. Lohmann, J. Gmehling, Chem. Technik 51 (1999) 184–190.[8] J. Gmehling, Fluid Phase Equilib. 144 (1998) 37–47.[9] R.J. Meyer, International Data Series, Sel. Data Mixtures Ser. A (1977) 139–149.

[10] R.J. Meyer, G. Giusti, Thermochim. Acta 25 (1978) 81–89.[11] J. Gmehling, J. Chem. Eng. Data 38 (1993) 143–146.[12] J. Krug, Experimentelle Untersuchungen von Exzessenthalpien und Exzessvolumen binarer Flussigkeitsgemische, Ph.D.

Dissertation, TU Berlin, 1985.[13] J.B. Ott, C.E. Stouffer, J.V. Cornett, B.F. Woodfield, R.C. Withlin, J.J. Christensen, U.K. Deiters, J. Chem. Thermodyn. 18

(1986) 1–12.[14] M. Ghysels, Bull. Soc. Chim. Belg. 33 (1924) 57–78.[15] H. Wuyts, Bull. Soc. Chim. Belg. 33 (1924) 167–192.[16] M. Lecat, Ann. Chim. 2 (1947) 158–202.[17] H. Wuyts, P. Docquier, Bull. Soc. Chim. Belg. 44 (1935) 297–306.[18] M. Lecat, Compt. Rend. 223 (1946) 286–289.[19] T. Treszczanowicz, B.C.-Y. Lu, J. Chem. Thermodyn. 18 (1986) 213–220.[20] L. Lepori, E. Matteoli, B. Marongiu, Fluid Phase Equilib. 42 (1988) 229–240.[21] O. Redlich, A.T. Kister, Ind. Eng. Chem. 40 (1948) 345–348.[22] J. Gmehling, C. Christensen, P. Rasmussen, U. Weidlich, Th. Holderbaum, Heats of Mixing Data Collection, DECHEMA

Chemistry Data Series, Vol. III, Frankfurt, 1984 (4 parts).[23] R.J. Meyer, J.V. Metzger, C. Kehiaian, H.V. Kehiaian, Thermochim. Acta 38 (1980) 197–209.[24] M.R. Tine, H.V. Kehiaian, Fluid Phase Equilib. 32 (1987) 211–248.[25] H.V. Kehiaian, M.R. Tine, Fluid Phase Equilib. 59 (1990) 233–245.[26] H.V. Kehiaian, M.R. Tine, L. Lepori, E. Matteoli, B. Marongiu, Fluid Phase Equilib. 46 (1989) 131–177.[27] H.S. Wu, S.I. Sandler, AIChE J. 35 (1989) 168–172.[28] H.S. Wu, S.I. Sandler, Ind. Eng. Chem. Res. 30 (1991) 881–889.[29] H.S. Wu, S.I. Sandler, Ind. Eng. Chem. Res. 30 (1991) 889–897.

Documents

Azeotropic and Mixing H Data for Binaries With Dietoxymethane