Azeotropic and solid–liquid equilibria data for several binary organic systems containing one acetal compound

Fluid Phase Equilibria 204 (2003) 267–280

Azeotropic and solid–liquid equilibria data for several binaryorganic systems containing one acetal compound

Mariana Teodorescua, Michael Wilkenb, Roland Wittigb, Jürgen Gmehlingb,∗a Institute of Physical Chemistry “I. G. Murgulescu”, Romanian Academy,

Splaiul Independentei 202, 77208 Bucharest, Romaniab Lehrstul für Technische Chemie (FB9TC), Carl von Ossietzky Universität Oldenburg,

Postfach 2503, D-26111 Oldenburg, Germany

Received 13 May 2002; accepted 7 August 2002

Abstract

Reliable azeotropic data have been measured for the following four binary systems: methanol+diethoxymethane,2-propanol+ diethoxymethane, diethoxymethane+ dimethyl carbonate and 2,2-dimethoxybutane+ toluene bymeans of a wire band column. Additionally, solid–liquid equilibria (SLE) for the six binary systems of benzene andcyclohexane with diethoxymethane, 2,2-dimethoxybutane, and 1,1-diethoxyethane have been measured by a visualtechnique. Due to correlation requirements of the SLE experimental data, the heat of fusion for 2,2-dimethoxybutaneand 1,1-diethoxyethane have been measured using a Tian–Calvet batch calorimeter. The azeotropic data werecompared with literature values, when available, or with predicted data by Modified UNIFAC (Dortmund) when allrequired parameters were at disposal, too. For the description of the SLE data the NRTL model has been used withgood results. Also, the results of the Modified UNIFAC (Dortmund) model have been checked using the alreadyavailable “ether” group parameters. The work was carried out in order to supplement the available data base foracetal systems required for the introduction of a new “acetal” group in the Modified UNIFAC (Dortmund) model.© 2002 Elsevier Science B.V. All rights reserved.

Keywords: Acetals; Azeotropic data; Solid–liquid equilibria; Experiments; Modified UNIFAC (Dortmund)

1. Introduction

Acetals are oxygenated organic compounds used as starting materials for perfumes, agricultural chem-icals, pharmaceuticals, and fragrances, for flavoring alcoholic drinks and more recently, as oxygenatedadditives to diesel or environmental fuels since they can reduce drastically the emission of not desiredcompounds as NOx or to decrease the auto ignition temperature[1]. Acetal compounds belong to the mainclasses of oxaalkanes. Preliminary extensive studies on proximity effects in terms of group contribution

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

0378-3812/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved.PII: S0378-3812(02)00260-1

268 M. Teodorescu et al. / Fluid Phase Equilibria 204 (2003) 267–280

models have already demonstrated this effect for oxaalkanes, including linear and cyclic acetals[2–8].UNIFAC accounts for large proximity effects just by defining new structural groups. It seems to be thecase for the “acetal” group (O–C–O), especially since preliminary results already showed that by usingan “ether” group the description of the thermodynamic properties of the investigated acetal systems werenot satisfying[9]. To overcome this problem, a new structural group for acetals has to be introduced inModified UNIFAC (Dortmund), too.

For the design of synthesis and separation columns as well as for the selection of the most suitable solventfor azeotropic distillation, the knowledge of azeotropic points is most important. Also, the solid–liquidequilibria (SLE) data are of great technical interest in the design and optimization of crystallizationprocesses when thermolabile components have to be separated or when unfortunate separation factors(α12 ≈ 1) cannot be changed by selective solvents. Besides their practical significance, azeotropic andSLE data are of great theoretical importance in supporting the development of group contribution models.In order to obtain the best representation of the real behavior of the systems across the whole compositionand a large temperature range, it is recommended to fit group interaction parameters simultaneously to allavailable reliable experimental data (VLE, azeotropic, SLE, activity coefficients at infinite dilution,γ∞,excess enthalpies,HE, etc.)[10,11]. The Modified UNIFAC (Dortmund) method[12,13]is one of the fewapplied models which follows this procedure in fitting temperature-dependent binary group interactionparameters. It is well known that the determination of reliable binary parameters is a precondition for agood description of phase equilibria for multicomponent systems. For this purpose, the immense amount ofdata[14] stored in Dortmund Data Bank (more than 45,700 entries on azeotropic and zeotropic behavior)is used.

The present work represents a continuation of the project concerning phase behavior of systemscontaining acetal compounds. The data are required for the introduction of an “acetal” group in theModified UNIFAC (Dortmund) model. Azeotropic and heat of mixing data for various binary sys-tems containing diethoxymethane have been presented previously[9]. In this work, azeotropic data formethanol+ diethoxymethane, 2-propanol+ diethoxymethane, diethoxymethane+ dimethyl carbonateand 2,2-dimethoxybutane+ toluene are presented from low to moderate pressure. For methanol, dimethylcarbonate[15] and 2-propanol[16] with diethoxymethane, azeotropic data at atmospheric pressure wereavailable for comparison. In addition, SLE data for benzene and cyclohexane with three different acetalcompounds, namely diethoxymethane, 2,2-dimethoxybutane, and 1,1-diethoxyethane have been mea-sured from 193 to 280 K. When possible, the measurements were carried for the whole compositionrange. No experimental data for the studied systems have been found in the literature for comparison.Due to the correlation requirements for the SLE data the heat of fusion�fusH as well as the melting pointfor 2,2-dimethoxybutane and 1,1-diethoxyethane have been measured too.

2. Experimental

2.1. Materials

Chemicals of high purity obtained from different suppliers were purified by vacuum distillation beforetheir use. The purity was carefully checked by gas chromatography (GC) and Karl–Fischer titration.Besides the supplier, the chemicals together with the pure component specification are summarized inTable 1.

M.Teodorescu

etal./Fluid

Phase

Equilibria

204(2003)

267–280269

Table 1Suppliers, CAS number, purity, water content, heat of fusion (�fusH), melting points (Tfus), heat of transition (�trsH) and transition points (Ttrs) for the chemicalsused

Compound Supplier CASnumber

Purity(% GC)

Water content(mass ppm)

�fusH (J mol−1) Tfus (K) �trsH (J mol−1)DDBa

Ttrs (K)DDBa

DDBa This work DDBa This work

Benzene Scharlau 71-43-2 99.9 4.7 9,944 – 278.68 279.10 – –Cyclohexane Scharlau 110-82-7 99.9 8.9 2,628 – 279.75 280.09 6740.7 185.952,2-Dimethoxybutane Aldrich 3453-99-4 98.2 430 – 9,345 – 174.03 – –1,1-Diethoxyethane Aldrich 105-57-7 99.7 278 – 10,954 173.15 167.04 – –Diethoxymethane Acros 462-95-3 99.9 3.6 15,062 – 206.65 208.35 – –Toluene Scharlau 108-88-3 99.9 6.1Dimethyl carbonate Acros 616-38-6 99.9 5.62-Propanol Rotipuran 67-63-0 99.7 260Methanol Merck 67-56-1 99.9 30

a Dortmund Data Bank[24].


2.2. Measurement of azeotropic data

For the measurement of azeotropic data, a commercially available microspinning wire band columnwith electronically controlled reflux ratio from NORMAG GmbH (Hofheim, FRG) was used to distil themixtures as described earlier[17]. The reflux ratio is realized on the basis of vapor dividing principle. Forpressures below atmospheric pressure, the desired pressure is kept constant with the help of a vacuum pumpin connection with a manostat Kobold type DCM1. The temperature was determined using a resistancethermometer with an accuracy of±0.1 K and the pressure by means of a sensor (Druck Limited, typePDCR) with an accuracy of±0.05 kPa. For the analytical determination of the azeotropic composition gaschromatography was used. The required factors to determine the composition from recorded peak areaswere obtained using prepared test mixtures of exactly known composition. The accuracy of the azeotropiccomposition determination,yaz, is within±0.2 mol%. Approximately, 40 cm3 of the binary mixture withestimated azeotropic composition was distilled at constant pressure and a small pressure drop at nearlytotal reflux for approximately 60 min. Then, the composition of the distillate was determined by GC. Forall investigated systems, homogeneous pressure maximum azeotropes were obtained. To verify whetherthe system shows azeotropic (separation factorα12 = 1) and not quasi-azeotropic (α12 ≈ 1) behavior,the experiments were always repeated starting with a different feed composition.

2.3. SLE measurements

The SLE data of all the systems were measured by the synthetic visual technique which is described else-where[18–20]. The equilibrium cell (total volume 160 cm3) is glass made. It is inserted in a three-jacketvessel. The exterior jacket, kept under vacuum, prevents the condensation of water vapor from atmosphereon the glass surface at low temperatures. Therefore, the visual observation is not disturbed. The cryostat(RL6 CP Lauda model) medium (usually pure ethanol with low water content) flows through the centraljacket and transfer the heat to the equilibrium cell via the contact medium (also ethanol). A nitrogenatmosphere in the equilibrium cell avoids contamination with water.

First, the sample inserted in the equilibrium cell is supercooled in liquid nitrogen for a short time,whereby a small amount of the liquid crystallizes in the form of fine particles. Then, the equilibriumcell is installed in the apparatus. The mixture is heated slowly within a defined rate (e.g. 0.3 K h−1). Themelting point for a given composition is determined by visual detection of the temperature at which thesolid phase just disappears. By varying the composition of the mixture synthetically prepared, the liquidusline over the whole composition range can be measured. Using the apparatus, measurements between 193and 280 K have been performed. The temperature was determined with a calibrated platinum resistancethermometer (model KT5614, Hart Scientific). For the conversion of the measured resistance a digitalthermometer unit (model 1560, Hart Scientific) was used. The accuracy of the melting temperaturemeasurement by this method for the studied mixtures was found to be of±0.1 K. The accuracy of thecomposition was± 0.005 in mole fraction.

2.4. Enthalpy of fusion measurements

As it will be shown later, for the correlation and/or prediction of SLE data for eutectic systems,besides the melting temperature the enthalpies of fusion of the pure compounds are required. For2,2-dimethoxybutane and 1,1-diethoxyethane no such data were available in the literature. Subsequently,

M. Teodorescu et al. / Fluid Phase Equilibria 204 (2003) 267–280 271

they had to be determined in this work. At the same time the melting points for the pure chemicals weremeasured with the same procedure. Only for 1,1-diethoxyethane a value of 173.15 K was found in theDortmund Data Bank[14]. This data bank also contains recommend basic pure component propertiesoften derived from a large number of published data. The heat of fusion measurements have been per-formed by means of a Tian–Calvet heat flow batch calorimeter from SETARAM, France (model BT 2.15II), with a temperature range from 77.15 to 473.15 K. The principle of the measurement method hasbeen described in detail by Calvet and Prat[21]. The calorimeter consists of two thermal flux meters,each constructed by a series of 480 thermocouples surrounding a cylindrical cavity. The flux meters arearranged symmetrically around two identical cells in an aluminum block located in the cavity. The signaldelivered by the power difference of the two flux meters is proportional to the heat effects occurring inthe cells. The temperature of the calorimeter block can be regulated using liquid nitrogen and electricalheating and is monitored using a Pt 100 resistance thermometer located between the two cells. To accountfor the difference between the registered temperature and the sample temperature from the measuring cella calibration run was performed in advance as described elsewhere[22]. The two identical cells used forthe heat of fusion measurements were standard cells (five bars) made of stainless steel with 12.5 cm3 totalvolume and inlet volume of 2 cm3 and with PTFE sealing. While one of the cells was empty the otherone was filled with a defined amount of the sample. The whole calorimeter block was evacuated properlyby vacuum pump to avoid condensation effects and than the liquid nitrogen was introduced in order tocool down the system below the melting temperature of the investigated substance. After the crystalliza-tion occurred, the calorimeter was heated up using a minimum heating rate of 0.1 K min−1. The meltingprocess was registered by the heat flow signal and the area of the corresponding peak was converted toheat of fusion by integration. From the tangent line to the peak corresponding the appearance of the firstdrop of liquid in the sample, the melting temperature was determined. According to the manufacturer’sspecification, the experimental error for the determination of enthalpy of fusion is within±0.5%.

3. Results and discussion

The experimental azeotropic data (T, P, yaz) for the investigated systems methanol+diethoxymethane,2-propanol+ diethoxymethane, diethoxymethane+ dimethyl carbonate and 2,2-dimethoxybutane+toluene are given inTable 2together with published data[15,16]. In all the cases, homogeneous pressuremaximum azeotropes were observed. For the system 2,2-dimethoxybutane–toluene, azeotropic behaviorwas clearly confirmed by further experiments just at approximately atmospheric pressure (101.44 kPa).For sub-atmospheric pressures, the distillate composition was the same as the feed composition withinthe accuracy of the gas chromatographic method. These values are given inTable 2. However, no clearconfirmation was obtained by further experiments. This indicates that at the marked pressures inTable 2the azeotrope, if present, is either very narrow or tangent. Nevertheless, a more accurate method is requiredfor the composition analysis.

The experimental data for methanol and 2-propanol with diethoxymethane systems were comparedwith those predicted by the group contribution method Modified UNIFAC (Dortmund) using the “ether”group. The results are shown inFigs. 1 and 2in the form ofT–y1az diagrams. As can be seen, there is aqualitative agreement between experiment and prediction for the system 2-propanol+ diethoxymethaneand poor agreement for the system methanol+ diethoxymethane. The Clausius–Clapeyron and Gibbs–Helmholtz equations provide quantitative information about the alteration of the azeotropic composition


Table 2Experimental azeotropic data

System Experimental data T (K) Published data Reference

P (kPa) y1az P (kPa) y1az

Methanol (1)+ diethoxymethane (2)336.35 101.32 0.8579 [15]

101.52 0.8127 336.0348.89 0.7769 318.2018.47 0.7302 298.23

2-Propanol (1)+ diethoxymethane (2)352.75 101.32 0.6525 [16]

98.61 0.6107 351.4537.11 0.4459 326.4712.54 0.3160 302.58

Diethoxymethane (1)+ dimethyl carbonate (2)359.15 101.32 0.5647 [15]

100.42 0.5563 358.7139.76 0.6379 331.7711.56 0.6968 302.73

2,2-Dimethoxybutane (1)+ toluene (2)101.44 0.9180 380.1531.21∗ 0.9770 344.196.03∗ 0.9473 306.04

∗ Not clearly confirmed azeotropic points. The azeotrope is either very narrow or tangent.

Fig. 1. Experimental and predicted azeotropic data for the binary system methanol (1)+ diethoxymethane (2): (�) experimentaldata from literature[15]; (�) our data; (—) Modified UNIFAC (Dortmund) prediction using the “ether” group.


Fig. 2. Experimental and predicted azeotropic data for the binary system 2-propanol (1)+ diethoxymethane (2): (�) experimentaldata from literature[16]; (�) our data; (—) Modified UNIFAC (Dortmund) prediction using the “ether” group.

with temperature. In most cases, a correct temperature dependence is predicted, which in majority ofcases depend mainly on the slope of the vapor pressure data and only to a smaller extent on the tem-perature dependence of the activity coefficient (partial molar excess enthalpies)[14]. For the systemsdiethoxymethane+ dimethyl carbonate and 2,2-dimethoxybutane+ toluene, no prediction was possible,since either the group interaction parameters were not available (in the case of dimethyl carbonate) orvapor pressures data have not been yet determined (in the case of 2,2-dimethoxybutane). However, toimprove the results of Modified UNIFAC (Dortmund) additional measurements for the determinationof parameters for a new “acetal” group fitted simultaneously to VLE,HE, γ∞, SLE, etc. for systemscontaining acetal compounds, are required. Subsequently, SLE data for the binary systems of benzeneand cyclohexane (as representative for aromatic hydrocarbons and naphthenes with the same carbonnumber) with three different acetal compounds, namely diethoxymethane, 2,2-dimethoxybutane and1,1-diethoxyethane have been measured. For these systems, no experimental data have been published.

The experimental SLE data for the binary benzene (1)+ diethoxymethane (2), benzene (1)+ 2,2-dimethoxybutane (2) and benzene (1)+ 1,1-diethoxyethane (2) systems are listed inTable 3and thosefor cyclohexane (1)+ diethoxymethane (2), cyclohexane (1)+ 2,2-dimethoxybutane (2) and cyclohexane(1) + 1,1-diethoxyethane (2) are given inTable 4. A relation for the calculation of SLE can be derivedstarting from the isofugacity criterion, e.g. as shown by Gmehling and Kolbe[23]. With some usefulsimplifications it leads to the following formula which was already described elsewhere[18]:

lnxLi γL

i = −�fusHi

RT

(1 − T

Tfus,i

)− �trsHi

RT

(1 − T

Ttrs,i

)(1)

wherexLi is the mole fraction of componenti in the liquid phase,γL

i the activity coefficient of componenti in the liquid phase,�fusHi the enthalpy of fusion of componenti, Tfus,i the melting temperature ofcomponenti, T the absolute temperature, andR the molar gas constant. If a solid phase transition occurs,


Table 3Experimental solid–liquid equilibrium (SLE) data for the binary systems benzene+ diethoxymethane, benzene+ 2,2-dimethoxybutane and benzene+ 1,1-diethoxyethane

xL1 T (K) xL

1 T (K) xL1 T (K)

Benzene (1)+ diethoxymethane (2)0.000 208.3 0.406 230.4 0.752 261.90.062 206.9 0.457 236.2 0.799 265.00.100 206.0 0.498 239.8 0.846 268.50.213 203.7 0.549 244.8 0.898 272.10.254 210.9 0.601 249.4 0.949 275.70.306 218.3 0.649 253.7 1.000 279.20.358 224.8 0.698 257.6

Benzene (1)+ 2,2-dimethoxybutane (2)0.000 174.0 0.550 244.5 0.807 266.10.268 211.7 0.598 248.8 0.843 268.20.305 217.1 0.649 253.5 0.897 272.00.386 227.5 0.698 257.7 0.950 275.60.498 239.3 0.749 261.7 1.000 279.0

Benzene (1)+ 1,1-diethoxyethane (2)0.000 167.0 0.454 234.2 0.753 261.60.209 199.7 0.500 238.3 0.797 264.60.259 208.4 0.549 243.7 0.847 268.30.305 215.4 0.598 248.2 0.897 271.90.357 222.4 0.648 252.7 0.945 275.30.409 228.6 0.698 257.0 1.000 279.1

the molar enthalpy of transition�trsHi must be considered below the transition temperatureTtrs,i by thesecond term on the right side ofEq. (1). Since the activity coefficient depends on temperature as wellas on concentration,Eq. (1)must be solved iteratively. For the correlation of the activity coefficients theNRTL model has been used by means of the following objective function,F:

F =∑

n

(γ1,calc − γ1,exp

γ1,calc

)2

(2)

wheren represents the number of the experimental points andγ1,calc,γ1,exprepresent the activity coefficientfor component 1 calculated by the model or from experimental data.

Figs. 3–8contain the experimental data together with the correlations by means of the NRTL model andthe predictions of the group contribution method Modified UNIFAC (Dortmund) using the parameters forthe “ether” main group. At the same time the results are compared with the predictions assuming idealbehavior. In all the cases, eutectic behavior is assumed. The binary parameters for the NRTL model�gij

as well as the non-randomness parametersαij together with the root-mean-square deviations (RMSDs)are given inTable 5. For all studied systems (except the system cyclohexane+ 2,2-dimethoxybutane)the obtained RMSDs are smaller than 0.5 K. The systems benzene+ dimethoxyethane (Fig. 3) (eu-tectic point atxL

i ≈ 0.2), benzene+ 2,2-dimethoxybutane (Fig. 4) (eutectic pointxLi ≈ 0.07) and

benzene+ 1,1-diethoxyethane (Fig. 5) (eutectic pointxLi ≈ 0.07) show nearly ideal behavior. This


Table 4Experimental solid–liquid equilibria (SLE) data for the binary systems cyclohexane+ diethoxymethane, cyclohexane+ 2,2-dimethoxybutane and cyclohexane+ 1,1-diethoxyethane

xL1 T (K) xL

1 T (K) xL1 T (K)

Cyclohexane (1)+ diethoxymethane (2)0.000 208.4 0.374 200.4 0.750 231.80.052 207.0 0.419 199.4 0.798 239.20.105 205.7 0.500 198.2 0.845 247.50.151 204.9 0.547 206.0 0.899 257.80.206 203.8 0.597 211.2 0.948 268.40.255 203.1 0.649 217.5 1.000 280.10.305 202.3 0.699 224.3

Cyclohexane (1)+ 2,2-dimethoxybutane (2)0.000 174.0 0.694 218.7 0.947 268.80.549 193.9 0.748 228.2 1.000 280.10.597 201.6 0.799 238.10.650 210.8 0.858 250.0

Cyclohexane (1)+ 1,1-diethoxyethane (2)0.000 167.0 0.702 221.8 0.895 256.40.549 197.9 0.752 229.3 0.948 268.80.598 205.3 0.797 237.3 1.000 280.00.650 213.7 0.848 247.1

Fig. 3. Experimental and calculated solid–liquid equilibria data for the system benzene (1)+ diethoxymethane (2): (�) experi-mental; (—) NRTL correlation; (- - -) Modified UNIFAC (Dortmund) prediction using the “ether” group; (· · · ) ideal behavior.


Fig. 4. Experimental and calculated solid–liquid equilibria data for the system benzene (1)+ 2,2-dimethoxybutane (2): (�)experimental; (—) NRTL correlation; (- - -) Modified UNIFAC (Dortmund) prediction using the “ether” group; (· · · ) idealbehavior.

Fig. 5. Experimental and calculated solid–liquid equilibria data for the system benzene (1)+ 1,1-diethoxyethane (2): (�)experimental; (—) NRTL correlation; (- - - ) Modified UNIFAC (Dortmund) prediction using the “ether” group; (· · · ) idealbehavior.


Fig. 6. Experimental and calculated solid–liquid equilibria data for the system cyclohexane (1)+ diethoxymethane (2): (�)experimental; (—) NRTL correlation; (- - - ) Modified UNIFAC (Dortmund) prediction using the “ether” group; (· · · ) idealbehavior.

Fig. 7. Experimental and calculated solid–liquid equilibria data for the system cyclohexane (1)+ 2,2-dimethoxybutane (2):(�) experimental; (—) NRTL correlation; (- - - ) Modified UNIFAC (Dortmund) prediction using the “ether” group; (· · · ) idealbehavior.


Fig. 8. Experimental and calculated solid–liquid equilibria data for the system cyclohexane (1)+ 1,1-diethoxyethane (2): (�)experimental; (—) NRTL correlation; (- - - ) Modified UNIFAC (Dortmund) prediction using the “ether” group; (· · · ) idealbehavior.

Table 5Results of the solid–liquid equilibrium data correlation by means of the NRTL model

System �g12 (J mol−1) �g21 (J mol−1) α12 RMSDa (K)

Benzene (1)+ diethoxymethane (2) 1004.0 −900.73 0.3089 0.2Benzene (1)+ 2,2-dimethoxybutane (2) −540.06 467.24 0.3034 0.3Benzene (1)+ 1,1-diethoxyethane (2) −1785.4 2126.5 0.3018 0.2

Cyclohexane (1)+ diethoxymethane (2) 958.89 721.58 0.3008 0.3Cyclohexane (1)+ 2,2-dimethoxybutane (2) 335.83 525.48 0.3017 1.1Cyclohexane (1)+ 1,1-diethoxyethane (2) 392.03 860.34 0.3013 0.7

a RMSD =√

(1/n)∑

n

(Texp − Tcalc

)2.

behavior is well predicted by Modified UNIFAC (Dortmund) using the “ether” group. The systemscyclohexane+diethoxymethane (Fig. 6) (eutectic pointxL

i ≈ 0.48), cyclohexane+ 2,2-dimethoxybutane(Fig. 7) (eutectic pointxL

i ≈ 0.21) and cyclohexane+ 1,1-diethoxyethane (Fig. 8) (eutectic pointxLi ≈

0.13) show slightly positive deviation from ideal behavior. The predictions of Modified UNIFAC (Dort-mund) using the ether group are not too bad.

4. Conclusions

Azeotropic data for the binary systems methanol+ diethoxymethane, 2-propanol+ diethoxymethane,diethoxymethane+ dimethyl carbonate and 2,2-dimethoxybutane+ toluene have been measured andcompared, when possible, with the prediction results of the group contribution method Modified UNIFAC


(Dortmund). The comparison reveals satisfactory agreement in the case of azeotropic data for 2-propanolmixtures, but poor agreement in the case of methanol mixtures. Additionally, liquidus lines were measuredfor the six binary systems of benzene and cyclohexane with diethoxymethane, 2,2-dimethoxybutane, and1,1-diethoxyethane. The experimental SLE data have been correlated successfully using the NRTL model.They were also compared with those predicted by the Modified UNIFAC (Dortmund) model using the“ether” group. The results of the comparison are good for the systems containing benzene which showsalmost ideal behavior and are not too bad for the systems containing cyclohexane which exhibit slightlypositive deviation from ideal behavior. The need of an acetal group has been already demonstrated andthe present measurements confirm this conclusion.

With the data measured in this work, the experimental data base for acetal systems has been enlarged.The experimental data will be used as supporting information for fitting temperature dependent groupinteraction parameters for a new “acetal” group for the Modified UNIFAC (Dortmund) model. The database will be supplemented soon with excess enthalpies at high temperature and activity coefficients atinfinite dilution for similar systems.

List of symbolsF objective functiong parameter for NRTL model (J mol−1)GC gas chromatographyH molar enthalpy (J mol−1)N number of the experimental data pointsP pressure (kPa)R molar gas constant (8.3144 J mol−1 K−1)SLE solid–liquid equilibriumT absolute temperature (K)VLE vapor-liquid equilibriumx liquid phase mole fractiony vapor phase mole fraction

Greek lettersα non-randomness parameter for the NRTL modelα12 separation factorγ activity coefficient� variation of property

SuperscriptsE excess propertyL for the liquid phase∞ at infinite dilution

Subscripts1,2,i, j components 1, 2,i, jaz at the azeotropic pointcalc calculated value


exp experimental valuefus fusiontrs transition

Acknowledgements

The authors would like to thank the “Fond der Chemischen Industrie (FCI)” for the financial supportand R. Bölts for technical assistance. Furthermore, we would like to thank the DDBST GmbH (Oldenburg,Germany) for providing the latest version of the Dortmund Data Bank.

References

[1] V.M.T.M. Silva, A.E. Rodrigues, AIChE J. 48 (3) (2002) 625–634.[2] R.J. Meyer, J.V. Metzger, C. Kehiaian, H.V. Kehiaian, Thermochim. Acta 38 (1980) 197–209.[3] M.R. Tine, H.V. Kehiaian, Fluid Phase Equilib. 32 (1987) 211–248.[4] H.V. Kehiaian, M.R. Tine, Fluid Phase Equilib. 59 (1990) 233–245.[5] H.V. Kehiaian, M.R. Tine, L. Lepori, E. Matteoli, B. Marongiu, Fluid Phase Equilib. 46 (1989) 131–177.[6] H.S. Wu, S.I. Sandler, AIChE J. 35 (1989) 168–172.[7] H.S. Wu, S.I. Sandler, Ind. Eng. Chem. Res. 30 (1991) 881–889.[8] H.S. Wu, S.I. Sandler, Ind. Eng. Chem. Res. 30 (1991) 889–897.[9] D. Constantinescu, R. Wittig, J. Gmehling, Fluid Phase Equilib. 191 (2001) 99–109.

[10] J.R. Rarey-Nies, D. Tiltmann, J. Gmehling, Chem. Eng. Technol. 61 (1989) 407–410.[11] J. Gmehling, Pure Appl. Chem. 71 (6) (1999) 939–949.[12] U. Weidlich, J. Gmehling, Ind. Eng. Chem. Res. 26 (1987) 1372–1381.[13] J. Gmehling, R. Wittig, J. Lohmann, R. Joh, Ind. Eng. Chem. Res. 41 (2002) 1678–1688.[14] J. Gmehling, J. Menke, J. Krafczyk, K. Fischer, Fluid Phase Equilib. 103 (1995) 51–76.[15] M. Lecat, Ann. Chim. 2 (1947) 158–202.[16] M. Lecat, Ann. Soc. Sci. Bruxelles 61 (1947) 153–162.[17] J. Gmehling, R. Bölts, J. Chem. Eng. Data 41 (1996) 202–209.[18] A. Jakob, R. Joh, C. Rose, J. Gmehling, Fluid Phase Equilib. 113 (1995) 117–126.[19] C. Fiege, R. Joh, M. Petri, J. Gmehling, J. Chem. Eng. Data 41 (1996) 1431–1433.[20] R. Wittig, D. Constantinescu, J. Gmehling, J. Chem. Eng. Data 46 (2001) 1490–1493.[21] E. Calvet, H. Prat, Recent Progress in Microcalorimetry, Pergamon Press, Oxford, 1963.[22] L. Becker, O. Aufderhaar, J. Gmehling, J. Chem. Eng. Data 45 (2000) 661–664.[23] J. Gmehling, B. Kolbe, Thermodynamik, VCH, Weinheim, 1992.[24] http://www.ddbst.de.

http://www.ddbst.de

Documents

Azeotropic and solid–liquid equilibria data for several binary organic systems containing one acetal compound