Separation Performance of BioRenewable Deep Eutectic Solvents

Article pubs.acs.org/IECR

Separation Performance of BioRenewable Deep Eutectic Solvents Sergey P. Verevkin,*,†,‡ Aleksandra Yu. Sazonova,§ Alla K. Frolkova,§ Dzmitry H. Zaitsau,‡ Igor V. Prikhodko,∥ and Christoph Held*,⊥ †

Department of Physical Chemistry, University of Rostock, D-18059 Rostock, Germany Department of Physical Chemistry, Kazan Federal University, 420008 Kazan, Russia § Lomonosov Moscow State University of Fine Chemical Technology, 119571 Moscow, Russia ∥ Saint Petersburg State University, Institute of Chemistry, 199034 St. Petersburg, Russia ⊥ Technische Universität Dortmund, Department BCI, Laboratory of Thermodynamics, D-44227 Dortmund, Germany ‡

S Supporting Information *

ABSTRACT: Deep eutectic solvents (DESs) have been regarded as promising cost-effective and environmentally benign alternatives to conventional volatile organic solvents. The screening and selection of the suitable solvent for separation is an important part of the process design. Limiting activity coefficients provide a useful tool for the optimal choice of the selective solvent. For the first time, activity coefficients at infinite dilution have been measured in DESs as a solvent for 23 solutes (aliphatic and aromatic hydrocarbons, alcohols, ketones, ethers, and esters). The DESs were constituted from choline chloride and glycerol in molar ratios of 1:1 and 1:2. The measurements were carried out with the help of gas−liquid chromatography in the temperature range 298−358 K. Using experimental results, selectivity of different separation cases was assessed. To verify the separation performance of DESs the perturbed-chain statistical associating fluid theory (PC-SAFT) was employed for the first time. This method appears to be powerful tool for screening of suitable precursors and evaluation of separation performance at temperatures relevant for practical applications. It has turned out that the separation performances of DESs are comparable to those of ionic liquids, but DESs are cheaper, because they are constituted from natural and renewable nontoxic bioresources. with a melting temperature of −36 °C.4 The strong association between hydrogen-bond donating hydroxyl groups of glycerol with the [Cl] anion provided by choline chloride ([Ch][Cl]) is the reason for considerable reduction of the melting point. Moreover, the strong interaction between components drastically decreases their reactivity, making them inert in most cases. Recently, DESs have emerged as attractive alternatives to conventional ionic liquids (ILs).3−7 They share the advantages of ILs such as thermal and chemical stability, low vapor pressure, and designability.5−7 While retaining the useful properties of ILs, DESs can be prepared more cheaply (from large scale industrial reagents) and easily (by simple stoichiometric mixing of the hydrogen-bond donor (HBD) with the appropriate hydrogen-bond acceptor (HBA)), than ILs. The portfolio of possible reagents as well as the range of possible compositions is almost unlimited.7 With such prerequisites, DESs are currently attracting significant attention as “greener” media for material synthesis and processing.7 For example, the DESs were successfully applied for separation of esters produced by the lipase-catalyzed (trans)esterification of 5-hydroxymethylfurfural.8 By using this approach, high ester purities (>99%) and efficiencies (up to >90% ester recovery) in separations were obtained by using choline chloride-based DES.

1. INTRODUCTION Biofuels are alternative substitutes for fossil fuel. Biofuels like ethanol, butanol, and biodiesel are made from agricultural crops. Biodiesel is industrially produced from renewable sources via reaction with alcohols; glycerol appears as a byproduct, representing ca. 10 wt % of the total output. In the past few years, the world glycerol production has surpassed 2 million metric tons, glycerol coming from the biodiesel industry representing more than two-thirds of the total outcome.1 Utilization of the glycerol is one of the much researched topics in present times. Glycerol has been recently proposed as a green solvent for catalysis, organic synthesis, separations, and materials chemistry.2 As compared to the conventional petrochemical-derived solvents which are usually volatile organic compounds (VOCs), the biomass-based glycerol exhibits many advantages such as biodegradability and low vapor pressure at elevated temperatures. However, the possible use of glycerol as a solvent is aggravated with the high viscosity and the chemical reactivity of hydroxyl groups promoting side reactions. These obstacles could be partly overcome by mixing the biomass-based glycerol with a second (usually solid) component leading to formation of a hydrogen bonded liquids with melting points significantly below room temperature. These mixtures represent a new type of solvent, ionic liquid analoguesdeep eutectic solvents.3 Deep eutectic solvents (DES) can be considered as a new subset of ionic liquids (IL), where a hydrogen-bond acceptor (e.g., high-temperature melting quaternary ammonium salt choline chloride) by mixing with a hydrogen-bond donor (e.g glycerol) results in a solvent © XXXX American Chemical Society

Received: January 26, 2015 Revised: March 13, 2015 Accepted: March 18, 2015

A

DOI: 10.1021/acs.iecr.5b00357 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research

compositions. Thus, a priori prediction of DES physical properties for screening their separation performance is highly desired. The PC-SAFT equation of state is one of the successful models that has been applied extensively to calculate phase equilibria and physical properties of classical solvents, ILs, as well as biomolecules of low molecular weight (sugars, osmolytes, amino acids).19 PC-SAFT is especially suitable to nonspherical and asymmetric molecules like ILs and to components that strongly build hydrogen bonds. These features are especially important for modeling of DES because the hydrogen bonding is known to be the dominant factor making a liquid out of high-temperature melting precursors. Using the experimental γ∞ i values measured in this work, we have successfully tested the ability of the PC-SAFT model to predict trends in separation performance of DESs under study. This model can be used for a priori screening of effectiveness of HBA and HBD combinations for separation processes.

The choline chloride based DESs were also tested as extraction solvents in the liquid−liquid separation of azeotropic mixtures.9 The data obtained show that DESs surpass the performance of existing extraction solvents, leading to an increase in efficiency and a reduction in energy consumption of the overall process.9 The selection of the most suitable selective solvent for separation processes such as distillation, extraction, or membrane processes is important for the economical design. Optimization of separation processes is based on the thermodynamic data especially on values of activity coefficients at infinite dilution γ∞ required to assess effectiveness of i separation in a liquid−gas or liquid−liquid systems. Infinite of a solute i quantitatively dilution activity coefficient γ∞ i characterizes the behavior of a single solute molecule completely surrounded by a solvent. The value of γ∞ i is a practical measure of the intensity of the solute−solvent interactions. Consequently, it indicates a maximum of deviation from ideality in the solution and provides incisive information regarding solute−solvent interactions in the absence of solute− solute interactions. From the practical viewpoint, a reliable knowledge of the activity coefficients at infinite dilution γi∞ is of crucial importance for the design and optimization of separation processes, since the largest efforts are usually required to remove the last traces of impurities. Moreover, knowledge of values is important for screening of selective solvents γ∞ i (entrainers) for extractive distillation and extraction or to solve separation problems like azeotropy and miscibility gaps. Furthermore, activity coefficients provide helpful information for testing applicability and development of the theoretical models describing interactions on a physical basis. Activity coefficients γ∞ i of volatile solutes in solvents with low volatility can be measured by gas−liquid chromatography,10 static,11 or dilutor techniques.12 In 2001, our pioneering study of the γ∞ i values for the aprotic ionic liquid 4-methyl-N-butylpyridinium tetrafluoroborate10 has launched systematic measurements of activity coefficients for these neoteric solvents. In 2011, we reported the very first γ∞ i values for the protic ionic liquid ethylammonium nitrate.13 A very recent paper by Lei et al.14 has reviewed over 220 works on activity coefficients at infinite dilution concerning ILs. However, to the best of our knowledge, no data on activity coefficients γ∞ i for any DES are available up to date. This contribution complements and extends our previous work on the thermodynamics of neoteric solvents.10,11,13,15−18 The aim of this study was an experimental and computational study of activity coefficients γ∞ i in order to evaluate separation performance of exemplary DESs based on choline chloride and glycerol in the stoichiometric compositions 1:1 and 1:2. Choline chloride, an essential nutrient, is industrially produced on a large scale. It is also known to be nontoxic and biodegradable. The possibility for utilization of glycerol originated from renewable sources has motivated the study of this particular hydrogen-bonding donor in the DES composition. Gas−liquid chromatography (GC) was chosen as the preferable technique for measurements of activity coefficients γ∞ i . A carefully prepared GC column allows the measurement of a large number of reliable γ∞ i values in a rather short time using the well established data treatment procedure. Nevertheless, the capacity of the experiment is limited in contrast to the unlimited number of possible variations of HBA and HBD precursors for DESs upon changes in their types and

2. EXPERIMENTAL SECTION 2.1. Synthesis of DESs. Choline chloride (Sigma-Aldrich, >98%) and glycerol (Acros, 99+%) were used as received. We checked the purity of the commercial choline chloride by the titration with the AgNO3. It was better than 98%. About 1% of water in the sample was detected by the KF-Titration. The synthesis was carried out in a round-bottomed flask. Raw materials were mixed at a certain molar ratio (1:1 and 1:2) of HBA and HBD. After precursors become liquid with heating, the system was stirred at 80 °C for 3 h. The deep eutectic mixtures prepared in this way were labeled in this work as DES (1:1) and DES (1:2). 2.2. Measurements of Limiting Activity Coefficients. A retention time of a volatile solute i measured on the GC column filled with a certain amount of low-volatile solvent (DES or IL) is considered as a measure of intermolecular interactions of a single molecule i with the bulk phase of solvent. This retention time is proportional to the activity 10 coefficients at infinite dilution γ∞ The Hewlett-Packard gasi . chromatograph equipped with a flame ionization detector. Nitrogen was used as a carrier gas. A GC column (stainless steel) with a length of 45 cm with an inside diameter of 0.40 cm was used. Chromosorb W/AW-DMCS 100/120 mesh was used as a solid support for the DES in the GC column. The column preparation and the packing method were described by us previously.10 The amount of stationary phase (DES) on the support was 30 mass percent. Such an amount was sufficient to prevent the adsorption ability of the solid support. The main impurity in the commercial choline chloride was water (about 1%); however it was completely removed from the prepared DESs during the preconditioning of the GC column under a nitrogen flow prior to the beginning of the experiment. Volumes of the samples injected into the GC probes were 0.5− 2 μL, and such amounts can be considered to be at infinite dilution on the column. The temperature of the column was maintained constant to within 0.1 K. Each experiment was repeated at least twice to establish repeatability. Retention times were generally reproducible to within 0.01 min depending upon the temperature and the individual solute. Details on the experimental procedure are given in the Supporting Information. The estimated overall error in limiting activity coefficient was less than 3%, taking into account the possible errors in determining the column loading, the retention times, and solute vapor pressure of solutes under B


Article


Table 1. Experimental Values of Activity Coefficients at Infinite Dilution of Non-Polar Solutes in Different Solvents at 298.15 K and 1 bar solvent/solute

n-C8H18

n-C10H22

n-C12H24

n-C14H30

benzene

toluene

cyclohexane

DES 1:1 DES 1:2 glycerol26 [Ch][Cl] [C4mim][NTf2]15 [C8mim][BF4]16 [C2H5NH3][NO3]11 PEG-40027

618 304 752 193 33.8 22.0 2165 22

846 504 1119 330 64.5 36.4 2566 37

1337 728 1541 678 127 78.0 3165 61

5011 1096 2101 1207 236 153 4403 103

27.1 18.8 66.5 86.4 0.88 1.2 14.3 1.1

60.1 44.3 148 153 1.4 1.5 26.3 1.5

228 131 248 143 9.3 8.1

study. Experimental results on γ∞ i measurements are given in short form in Tables 1−4 and detailed in the Supporting Information (Tables S1−S8).

thermodynamic relations.21 Interaction parameters or correlating factors are not used in this work. The classical mixing rules were applied as described in ref 21. 3.1. Pure Components. PC-SAFT pure component parameters for a DES based on choline chloride and glycerol were obtained in this work (see Table S9 in Supporting Information). These parameters were supposed to be valid independently of the molar ratio between choline chloride and glycerol. Experimental densities available in the temperature range 298−323 K at pressures from 1 up to 400 bar22 were used to derive PC-SAFT parameters. The deviation between experimental and calculated with PC-SAFT densities was at the level of 0.2% (Table S9). 3.2. Binary Mixtures. On the basis of the pure-component parameters, PC-SAFT can be applied to predict activity coefficients at infinite dilution of binary mixtures. We used the standard Lorentz−Berthelot combining rules for dispersion energy and segment diameter in a mixture i and j without using binary fitting parameters. Table 5 compares experimental data with prediction results for activity coefficients at infinite dilution in DESs and some selected ILs at 298.15 K.

Table 2. Experimental Values of Activity Coefficients at Infinite Dilution of n-Alcohols in Different Solvents at 298.15 K and 1 bar solvent/solute

MeOH

EtOH

PrOH

BuOH

PeOH

DES 1:1 DES 1:2 glycerol27 [Ch][Cl] [C4mim][NTf2]15 [C8mim][BF4]16 [C2H5NH3][NO3]13 PEG-40030

0.6 1.0 1.9 5.8 1.3 1.1 1.5 0.4

2.2 1.8 3.9 18.5 1.9 1.8 2.3 0.5

5.1 4.0 7.8 20.2 2.7 2.2 2.7 0.6

9.4 8.0 14.4 22.9 3.0 2.3 3.6 0.7

33.9 22.4 28.4 28.0 3.2 2.4 4.8 0.8

6.1

3. THERMODYNAMIC MODELING The PC-SAFT model is a state-of-the-art, engineering-like equation of state. It is designed for modeling mixtures of all types of substances: gases, solvents, and polymers. PC-SAFT is suitable for the calculation of phase equilibria and thermophysical properties of pure components and mixtures. It has been tested against experimental data for numerous systems and found to give excellent results. When compared to other equations of state, we find PC-SAFT to be more precise for correlation of experimental data and more predictive when applied to mixtures. It is very reliable for extrapolations over the regions where parameters were fitted. The aim of this work is to apply the classical PC-SAFT20 to predict the activity coefficients of bulk chemicals of different classes (alcohols, alkanes, aromatics, polar components) at infinite dilution in solvents such as DES, ILs, or classical organic solvents. This requires pure-component PC-SAFT parameters in a first step. Activity coefficients are then obtained by calculating the Helmholtz energy of the considered system and using classical

4. RESULTS AND DISCUSSION 4.1. Experimental Limiting Activity Coefficients. Measurements of activity coefficients γ∞ i with the GC method are generally based on the assumption that solvent (DES or IL) has no measurable vapor pressure in the temperature range under study of 298−353 K. This assumption is completely correct for ILs,8,23−25 but it has to be proved for measurements with DES as the solvent. The exact and constant quantity of the solvent on the solid support is the prerequisite for correct calculations of activity coefficients. Thus, the main attention in the experimental part of this work has been paid to reproducibility of the retention time measurements. In order to detect the possible elution of the stationary phase DES with the nitrogen stream, the measurement of retention time of ndecane was repeated systematically every 2−3 h. No changes of

Table 3. Experimental Values of Activity Coefficients at Infinite Dilution of Common Solutes in Different Solvents at 298.15 K and 1 bar solvent/solute DES 1:1 DES 1:2 glycerol [Ch][Cl] [C4mim][NTf2]15 [C8mim][BF4]16 [C2H5NH3][NO3]13

acetone 3.6 5.029 26.3 0.4 0.8

MIBK

ACN

EtOAc

THF

pyridine

66.2 35.0 1231 50.8

2.9 2.1 3.727 93.3 0.5 0.6

27.9 16.9

11.2 6.0

4.6 3.0

38.8 0.8 1.4 5.4

27.1

12.7

C


Article

Industrial & Engineering Chemistry Research Table 4. Values of Selectivity Factors Sij of Typical Separation Examples solute i

alkane/alcohola

alkane/benzeneb

ester/alcoholc

cyclohexane/benzene

DES 1:1 DES 1:2 [Ch][Cl] Glycerol [C4mim][NTf2] [C8mim][BF4] [N1,1,1,4][NTf2] [C4mim][NO3] [C4mim][SCN] [C2mim][EtSO4] [C2mim][N(CN)2] [C2H5NH3][NO3] PEG-400

90 (102) 62 (358) 14 78 98 (7) 71 (14) 149 2604 1893 2040 3200 2110 53

31 (304) 27 (302) 3.8 17 14 (7) 12 (13) 15 29 24 9.5 29 9.7 34

3.0 (3.8) 2.1 (4.0) 1.7

8.4 (2.9) 6.4 (2.9) 1.7 3.7 10(1.6) 6.7 (2.2) 4.2 15 29 15 0.4

0.4 0.8 0.4 5.8 6.4 2.5

5.5

ref this this this this 13 14 19 25 23 20 24 11 28

work work work work

a

Using n-decane/n-butanol model system. bUsing n-decane/benzene model system. cUsing ethyl acetate/n-butanol model system. dValues in parentheses were predicted by PC-SAFT.

solvent. Are these interactions for DESs much different in comparison to the aprotic and the protic ILs or in another promising solvent polyethylene glycol (PEG)? The nonpolar aliphatic and aromatic hydrocarbons (see Table 1) are interacting with both DESs very weakly. For example, γ∞ i values for n-octane in typical common imidazolium based aprotic ILs (such as [C4mim][NTf2] and [C8mim][BF4]) and in PEG-400 are between 22 and 34. At the same time, γ∞ i values for n-octane in both DESs are 10 to 20 times higher than in these solvents (see Table 1), but they are still far away from the γ∞ i = 2165 for the typical protic ionic liquid [C2H5NH3][NO3]. It is interesting to note that γ∞ i values of aliphatic hydrocarbons in both DESs are roughly between those in the pure [Ch][Cl] and the pure glycerol. In contrast, for the aromatic hydrocarbons, γ∞ i values are significantly lower in comparison to those in pure precursors, apparently due to intensive interactions of π orbitals with the hydrogen-bonded network. It is apparent from Table 2 that the polar alcohols are strongly interacting with both DESs. For example, γ∞ i values for ethanol in the imidazolium based aprotic ILs as well as in the protic ionic liquid [C2H5NH3][NO3] are approximately at the same level, around γ∞ i = 2, but these values are still in pure competition with the PEG-400, where γ∞ i values are typically below unity. As expected, γ∞ i values of alcohols in both studied DESs are 2−8 times lower than those in [Ch][Cl], but surprisingly they are about twice lower than even γ∞ i values of alcohols in the pure glycerol and in PEG-400 are between 22 and 34. Experimental γ∞ i values for selected common solutes are collected in Table 3, and they follow the common sense. Methyl isobutyl ketone and ethyl acetate interact with DESs at the moderate level. Acetone, acetonitrile, pyridine, and tetrahydrofuran are intensively interacting with both DESs but significantly less intensively than with ILs taken for comparison in this table. Finally, it has turned out from results measured in this work that tendencies in γ∞ i values for solutes under study in DESs are generally the same as in the aprotic and the protic ILsthe activity coefficients of the series of n-alkanes and n-alkanols increase with increasing chain length. Correlations of ln(γ∞ i ) values with the number of carbon atoms for homologous series in DESs at 298 K show good linearity (see Figure 1). This fact confirms consistency of the experimental data measured in the

Table 5. Comparison of Experimental and Predicted Values of Activity Coefficients at Infinite Dilution of a Number of Solutes in Different Solvents at 298.15 K and 1 bar solvent/solute

n-C10H22

benzene

BuOH

CH3CN

EtAc

DES 1:1 (exp.) (pred.) DES 1:2 (exp.) (pred.) [C4mim][NTf2] (exp.) (pred.) [C8mim][BF4] (exp.) (pred.)

846 828 504 2750 64 11 36 34

27 5.1 19 9.1 0.9 1.5 1.2 2.7

9.4 8.3 8.0 7.7 3.0 1.6 2.3 2.4

2.9 1.3 2.1 1.9 0.5 0.8 0.6 0.9

28 54 17 57 0.8 1.4

the retention times were observed over 2−5 days of continuous operation; thus the experimental results should be considered as reliable. A series of nonpolar hydrocarbons such as n-alkanes, cyclohexane, benzene, and toluene (see Table 1), as well as for representatives of typically polar compounds like linear C1− C5 alcohols (see Table 2), have been studied over the temperature range 298−353 K in the DES (1:1) and DES (1:2). Additionally, we measured γi∞ values for selected common solutes: acetone, acetonitrile, chloroform, esters, ketones, pyridine, and tetrahydrofuran. For comparison, the same series of solutes was also studied on the pure [Ch][Cl] used as the stationary phase. For comparison, we also collected γ∞ i values for the same series of solutes in the pure glycerol available in the literature.26−28 The experimental values of γ∞ i have been approximated by the linear regression ln γ∞i = a + b/ T. The coefficients a and b, as well as the values of γ∞ i(298 K) calculated with these coefficients, are also given in Tables S1− S3. Values for the partial molar excess enthalpy at infinite dilution HE,∞ were obtained from the slope of ln γ∞ i i versus 1/T (see Table S4). A comparison of values of γ∞ of typical i nonpolar and polar solutes at 298 K in DES (1:1) and DES (1:2), in pure [Ch][Cl], and in pure glycerol derived in this work and those from the literature are given in the Tables 1−3. Activity coefficients γ∞ i are considered as a common measure for the intensity of intermolecular interactions between a solvent j (DES, IL, pure compound) and any solute i. From the practical point of view, the γ∞ i values at the level from 10 to 100 indicate weak interactions. The γ∞ i values at a level below 10 are clear evidence of the strong interactions between solute and D


Article


given in the Supporting Information (see Tables S10 and S11). It cannot be expected that quantitative results are obtained since (1) usually binary interaction parameters are applied for such complex mixtures and (2) some experimental data might be accurate within 30% only. Thus, the results in Table 4 are highly encouraging since PC-SAFT is generally able to predict the order of magnitude of most activity coefficients. It is especially important that the modeling was successful for very high activity coefficients (as in the case of alkanes) as well as for very low activity coefficients (as in the case of CH3CN). The corresponding prediction results for activity coefficients of selected solutes at infinite dilution in DES are presented in Tables S10−S12. It can be observed that the PC-SAFT results are in acceptable agreement with the experiment. 4.3.2. Temperature Dependence of γ∞ i . In order to evaluate the separation capacity at different temperatures, the activity coefficients at infinite dilutions at temperatures different from 298.15 K serve as valuable indicators. We have performed predictions for a series of solutes in DES at broad temperature ranges, and the results are summarized in Tables 6 and S12. Values of γ∞ i predicted with the PC-SAFT show the same trend as the experimental data, and γ∞ i values decrease with increasing temperature.

Figure 1. Values of ln(γ∞ i ) at 298 K as a function of the number of carbon atoms of solutes in DES (1:1) (•, n-alkanes; ■, n-alcohols) and in DES (1:2) (○, n-alkanes; □, n-alcohols).

Table 6. Comparison of Experimental and Predicted Values of Activity Coefficients at Infinite Dilution for a Series of Solutes in DES (ratio 1:1) at Temperatures Between 298 and 323 K at 1 bar

present work and gives the possibility to assess values of γ∞ i (298 K) for other members of the homologues series. 4.2. Experimental Selectivity of the DESs for Extraction Processes. Qualitative interpretation of intensity of interactions between solutes and DESs derived from experimental γ∞ i values can be easily quantified for the practical application by using the selectivity factor calculated according to the relation:

Sij =

∞

γi /γj∞

where i and j represent two components to be separated by the deep eutectic solvent. The selectivity factor Sij is broadly used for evaluation of the capability of DES in extraction processes or in order to assess the suitability of the DES as an entrainer in extractive distillation. The Sij values calculated in Table 4 provide an important indication of the ranking of solvents in terms of their efficiency in the separation of some mixtures typically occurring in chemical technology. In this case, the large Sij values refer to the better separation of compounds i and j. As is apparent from Table 4, the DESs are effective enough for alkane/alcohol and alkane/benzene separation. The separation efficiency for ester/ alcohol mixtures is in the middle field among the solvents listed in Table 4. Analysis of Sij values reveals that DESs are mostly suitable for the separation of nonpolar and polar organic solutes. This trend is similar to that inherent for ionic liquids. However, in contrast to very expensive ILs, the DESs are produced from cheap precursors readily available in the chemical industry. Consequently, the less expensive DESs could compete for ILs by replacing conventional entrainers applied for the separation processes. 4.3. Modeling Results. 4.3.1. Limiting Activity Coefficients Predicted by PC-SAFT. A priori designing of DESs for specific separation tasks and predicting their thermodynamic properties is highly desired. As can be observed from examples in Table 5, in most cases results of prediction agree qualitatively with measured limiting activity coefficients. The detailed tables with a comparison of experimental and predicted results are

solute i

298

303

nC8H18 (exptl.) predicted toluene (exptl.) predicted EtAc (exp.) predicted acetone (exptl.) predicted

618 494 60.1 12.6 27.9 33.9 5.5 5.0

566 432 59.2 12.0 27.5 32.0 5.4 4.9

308

313

323

57.8 11.5 26.8 30.3 5.3 4.8

551 335 56.7 11.0 26.0 28.7 5.2 4.6

467 266 53.8 7.1 25.1 26.0 4.8 4.4

4.3.3. Selectivity at Infinite Dilution. Activity coefficient predicted with PC-SAFT listed in Tables S10 and S11 were used for estimation of selectivities Sij at infinite dilution for some separation tasks given in Table 4. The predicted selectivities are in agreement with the experimental results. Such good consistency of predicted temperature dependences of γ∞ i as well as selectivities Sij proves PC-SAFT as a powerful tool for prediction thermodynamic properties of DESs at temperatures relevant for practical applications.

5. CONCLUSIONS Values of limiting activity coefficients were measured for the first time for two typical representatives of the novel type of solvents called deep eutectic solvents. These values have allowed assessment of the separation performance of DESs and comparison of it with properties of the parent ionic liquids. It was shown that DESs can be suitable solvents for different practical separation tasks. The separation performances of DESs can be comparable to those of ionic liquids, but DESs are cheaper, because they are constituted from natural and renewable nontoxic bioresources. Test modeling of the DESs thermodynamic properties with the PC-SAFT have revealed that this method is the powerful tool for screening of suitable E


Article


(13) Verevkin, S. P.; Zaitsau, D. H.; Tong, B.; Welz-Biermann, U. New for old. Password to the thermodynamics of the protic ionic liquids. Phys. Chem. Chem. Phys. 2011, 13, 12708−12711. (14) Lei, Z.; Dai, C.; Zhu, J.; Chen, B. Extractive distillation with ionic liquids: A review. AIChE J. 2014, 60, 3312−3329. (15) Heintz, A.; Casas, L. M.; Nesterov, I. A.; Emel’yanenko, V. N.; Verevkin, S. P. Thermodynamic Properties of Mixtures Containing Ionic Liquids. 5. Activity Coefficients at Infinite Dilution of Hydrocarbons, Alcohols, Esters, and Aldehydes in 1-Methyl-3-butylimidazolium Bis(trifluoromethyl-sulfonyl) Imide Using Gas-Liquid Chromatography. J. Chem. Eng. Data 2005, 50, 1510−1514. (16) Heintz, A.; Verevkin, S. P. Thermodynamic Properties of Mixtures Containing Ionic Liquids. 6. Activity Coefficients at Infinite Dilution of Hydrocarbons, Alcohols, Esters, and Aldehydes in 1Methyl-3-octyl-imidazolium Tetrafluoroborate Using Gas−Liquid Chromatography. J. Chem. Eng. Data 2005, 50, 1515−1519. (17) Heintz, A.; Vasiltsova, T. V.; Safarov, J.; Bich, E.; Verevkin, S. P. Thermodynamic Properties of Mixtures Containing Ionic Liquids. 9. Activity Coefficients at Infinite Dilution of Hydrocarbons, Alcohols, Esters, and Aldehydes in Trimethyl-butylammonium Bis(trifluoromethylsulfonyl) Imide Using Gas−Liquid Chromatography and Static Method. J. Chem. Eng. Data 2006, 51, 648−655. (18) Sumartschenkowa, I. A.; Verevkin, S. P.; Vasiltsova, T. V.; Bich, E.; Heintz, A. Experimental Study of Thermodynamic Properties of Mixtures Containing Ionic Liquid 1-Ethyl-3-methylimidazolium Ethyl Sulfate Using Gas−Liquid Chromatography and Transpiration Method. J. Chem. Eng. Data 2006, 51, 2138−2144. (19) Nann, A.; Mündges, J.; Held, C.; Verevkin, S. P.; Sadowski, G. Molecular Interactions in 1-Butanol + IL Solutions by Measuring and Modeling Activity Coefficients. J. Phys. Chem. B 2013, 117, 3173− 3185. (20) Gross, J.; Sadowski, G. Perturbed-Chain SAFT: An Equation of State Based on a Perturbation Theory for Chain Molecules. Ind. Eng. Chem. Res. 2001, 40, 1244−1260. (21) Held, C.; Cameretti, L. F.; Sadowski, G. Measuring and Modeling Activity Coefficients in Aqueous Amino-Acid Solutions. Ind. Eng. Chem. Res. 2011, 50, 131−141. (22) Leron, R. B.; Hill, W. D. S.; Li, M.-H. Densities of a deep eutectic solvent based on choline chloride and glycerol and its aqueous mixtures at elevated pressures. Fluid Phase Equilib. 2012, 335, 32−38. (23) Mutelet, F.; Revelli, A.-L.; Jaubert, J.-N.; Sprunger, L. M.; Acree, W. E., Jr.; Baker, G. A. Partition Coefficients of Organic Compounds in New Imidazolium and Tetralkylammonium Based Ionic Liquids Using Inverse Gas Chromatography. J. Chem. Eng. Data 2010, 55, 234−242. (24) Domanska, U.; Laskowska, M. Measurements of activity coefficients at infinite dilution of aliphatic and aromatic hydrocarbons, alcohols, thiophene, tetrahydrofuran, MTBE, and water in ionic liquid [BMIM][SCN] using GLC. J. Chem. Thermodyn. 2009, 41, 645−650. (25) Foco, G.; Bermejo, M. D.; Kotlewska, A. J.; van Rantwijk, F.; Peters, C. J.; Bottini, S. B. Activity Coefficients at Infinite Dilution in Methylimidazolium Nitrate Ionic Liquids. J. Chem. Eng. Data 2011, 56, 517−520. (26) Ge, M.-L.; Ma, J.-L.; Wu, C.-G. Activity Coefficients at Infinite Dilution of Alkanes, Alkenes, and Alkyl Benzenes in Glycerol Using Gas−Liquid Chromatography. J. Chem. Eng. Data 2010, 55, 1714− 1717. (27) Vincent, J. D.; Srinivas, K.; King, J. W. Characterization of the Solvent Properties of Glycerol Using Inverse Gas Chromatography and Solubility Parameters. J. Am. Oil Chem. Soc. 2012, 89, 1585−1597. (28) Fermeglia, M.; Braiuca, P.; Gardossi, L.; Pricl, S.; Halling, P. J. In Silico Prediction of Medium Effects on Esterification Equilibrium Using the COSMO-RS Method. Biotechnol. Prog. 2006, 22, 1146− 1152. (29) Scheepers, J. J.; Muzenda, E. GlycerolA Viable Solvent for Absorption of Highly Polar Solutes I: Behaviour of Molecular Interactions. J. Clean Energy Technol. 2015, 3, 282−286.

precursors and evaluation of separation performance at temperatures relevant for practical applications.

■

ASSOCIATED CONTENT

* Supporting Information S

Experimental details and additional data. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] (for experimental measurements (Rostock) and data evaluation). *E-mail: [email protected] (for PC SAFT modelling (Dortmund)). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work has been partly supported by the Russian Government Program of Competitive Growth of Kazan Federal University. I.V.P. is grateful to Technische Universität Dortmund for financial support for his research stay at TU Dortmund.

■

REFERENCES

(1) García, J. I.; García-Marína, H.; Piresa, E. Glycerol based solvents: synthesis, properties and applications. Green Chem. 2014, 16, 1007− 1033. (2) Gu, Y.; Jérôme, F. Glycerol as a sustainable solvent for green chemistry. Green Chem. 2010, 12, 1127−1138. (3) Abbott, A. P.; Boothby, D.; Capper, G.; Davies, D. L.; Rasheed, R. K. Deep Eutectic Solvents Formed between Choline Chloride and Carboxylic Acids: Versatile Alternatives to Ionic Liquids. J. Am. Chem. Soc. 2004, 126, 9142−9147. (4) Tang, B.; Row, K. H. Recent developments in deep eutectic solvents in chemical sciences. Monatsh. Chem. 2013, 144, 1427−1454. (5) Francisco, M.; van den Bruinhorst, A.; Kroon, M. C. LowTransition-Temperature Mixtures (LTTMs): A New Generation of Designer Solvents. Angew. Chem., Int. Ed. 2013, 52, 3074−3085. (6) Ruß, C.; König, B. Low melting mixtures in organic synthesis − an alternative to ionic liquids? Green Chem. 2012, 14, 2969−2982. (7) Smith, E. L.; Abbott, A. P.; Ryder, K. S. Deep Eutectic Solvents (DESs) and Their Applications. Chem. Rev. 2014, 114, 11060−11082. (8) Krystof, M.; Perez-Sanchez, M.; Dominguez de Maria, P. LipaseCatalyzed (Trans)esterification of 5-Hydroxymethylfurfural and Separation from HMF Esters using Deep-Eutectic Solvents. ChemSusChem. 2013, 6, 630−634. (9) Oliveira, F. S.; Pereiro, A. B.; Rebelo, L. P. N.; Marrucho, I. M. Deep eutectic solvents as extraction media for azeotropic mixtures. Green Chem. 2013, 15, 1326−1330. (10) Heintz, A.; Kulikov, D. V.; Verevkin, S. P. Thermodynamic Properties of Mixtures Containing Ionic Liquids. 1. Activity Coefficients at Infinite Dilution of Alkanes, Alkenes, and Alkylbenzenes in 4-Methyl-n-butylpyridinium Tetrafluoroborate Using Gas− Liquid Chromatography. J. Chem. Eng. Data 2001, 46, 1526−1529. (11) Verevkin, S. P.; Safarov, J.; Bich, E.; Hassel, E.; Heintz, A. Thermodynamic properties of mixtures containing ionic liquids: Vapor pressures and activity coefficients of n-alcohols and benzene in binary mixtures with 1-methyl-3-butyl-imidazolium bis(trifluoromethyl-sulfonyl) imide. Fluid Phase Equilib. 2005, 236, 222−228. (12) Kato, R.; Gmehling, J. Activity coefficients at infinite dilution of various solutes in the ionic liquids MMIM methylsulfate, MMIM methoxyethylsulfate, MMIM dimethylphosphate, N ethylpyridinium bis(trifluoromethylsulfonyl) imide and pyridiniumethoxyethylsulfate (MMIM = 1 methyl 3 methylimidazolium). Fluid Phase Equilib. 2004, 226, 37−44. F


Article

Industrial & Engineering Chemistry Research (30) Bighi, C.; Betti, A.; Dondi, F.; Francesconi, R. Gas chromatographic: behavior of Ci-C5 alcohols on Carbowax 400. J. Chromatogr. A 1969, 42, 176−182. (31) Díaz, E.; Cazurro, A.; Ordóñez, S.; Vega, A.; Coca, J. Determination of solubility parameters and thermodynamic properties in hydrocarbon-solvent systems by gas chromatography. Braz. J. Chem. Eng. 2007, 24, 293−306.

G


Separation Performance of BioRenewable Deep Eutectic Solvents

Recommend Documents