Effect of Ethanol and Temperature on Partition Coefficients of Ethyl

This is an open access article published under a Creative Commons Non-Commercial No Derivative Works (CC-BY-NC-ND) Attribution License, which permits copying and redistribution of the article, and creation of adaptations, all for non-commercial purposes.

Article Cite This: J. Chem. Eng. Data XXXX, XXX, XXX−XXX

pubs.acs.org/jced

Effect of Ethanol and Temperature on Partition Coefficients of Ethyl Acetate, Isoamyl Acetate, and Isoamyl Alcohol: Instrumental and Predictive Investigation Ali Ammari* and Karin Schroen Department of Agrotechnology and Food Science, Laboratory of Food Process Engineering, Wageningen University & Research, Bornse Weilanden 9, 6708 WG Wageningen, The Netherlands Downloaded via VOLUNTEER STATE COMMUNITY COLG on July 24, 2019 at 14:24:57 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: For alcoholic beverages such as beer, downstream processing for either dealcoholization or off-flavor removal requires both quantitative data and suitable predictive methods. Along with experimental investigations, we use a method initially developed for studying the solubility of gases in two or more miscible liquid solvents to monitor the effect of ethanol on air−water partition coefficients of three major flavors found in beer, namely, isoamyl alcohol, ethyl acetate, and isoamyl acetate. In the ethanol concentration range between 0 and 0.1 mole fraction, a slight, rather linear increase in the Henry’s solubility coefficient was observed. This overall behavior can be captured well using Henry coefficients for aqueous binary and ternary systems together with the Wohl expansion for excess Gibbs free energy coupled with the oneparameter Margules equation. Based on the developed model, the Wohl’s expansion parameter for ethanol−water is introduced as the solvent−solvent interaction parameter. The van ’t Hoff parameters for temperature dependence of Henry coefficients for binary water−flavor solutions are determined in the range of 30 to 60 °C.

1. INTRODUCTION Volatile flavor compounds are small molecules with relatively similar physiochemical characteristics, such as hydrophobicity or boiling points. These molecules are present at low concentrations in a complex matrix present in beverages; therefore, separation or addition of these compounds to enhance flavor profiles or develop new products is challenging. The market for alcohol-free beer is growing rapidly in Western Europe, with a worldwide market projected to reach 25 billion U.S. dollars by 2024.1 Alcohol-free beer can be produced using yeasts that do not produce alcohol, which affects the flavor composition compared to regular beer due to a different bioconversion. Alternatively, alcohol may be removed from regular beer, but in the process, the flavors are expected to be removed as well, depending on their physiochemical interaction with the aqueous matrix. For more information on the process details, we would like to refer to a recent review on production processes by Brányik et al.2 In current processes for the preparation of low-alcohol beer, two approaches are mostly used: single-stage and cost-effective dealcoholization and multistage ethanol separation and flavor recovery followed by reconstitution of beer.2 In brewing industries, stripping is a popular postproduction technique for both flavor control and dealcoholization due to its well-defined and mild operational temperature and other conditions. Depending on the stripping gas polarity, different volatiles © XXXX American Chemical Society

can be separated. For example, esters have higher affinity for carbon dioxide, whereas components with an alcoholic group can be removed through steam stripping. Through stripping, it is (almost) impossible to target one single component due to physiochemical similarities between volatile compounds. This implies that control of beer flavor needs to take place in multistage processes. For instance, a stripping column separates the component of interest that gives an off-flavor, and a consecutive stage recovers and recycles the “on-flavor” components that are removed with the target molecule. It is evident that the effectiveness of the latter stage depends on the previous stages, and the interaction of compounds needs to be charted in order to monitor the retention of flavors. Methods for the prediction of nonideality and the phase behaviors of simple systems comprising of liquid, gas, or vapor have been reasonably well documented. For aqueous mixtures, comprising electrolytes and flavors, even commercial chemical engineering software can be used to predict nonideality, although these databases use binary interaction parameters of which it is questionable whether they are valid for such dilute systems or not. Furthermore, the effect of the food matrix on Received: November 25, 2018 Accepted: June 21, 2019

A

DOI: 10.1021/acs.jced.8b01125 J. Chem. Eng. Data XXXX, XXX, XXX−XXX

Journal of Chemical & Engineering Data

ij GE yz jj z jj RT zzz = A12 x1x 2 + A13x1x3 + A 23x 2x3 k {123

flavor has been limited to sensory evaluation. Data on quantitative analytical methods are very scarce and limited in scope. In the current work, we quantify the effect of ethanol and temperature on the partitioning of three major flavors found in typical pilsner beer produced in the Netherlands and Belgium (light ale). When considering a gas−liquid system at equilibrium, the ratio of the concentration of flavors in the gas phase and liquid phase is constant. This constant is called the Henry’s law constant (HLC) and can be defined in either solubility or volatility terms K cc =

C 1 = G H cc CL

+ (β0 + β1x1 + β2x 2)x1x 2x3

where K and H are volatility and solubility coefficients, respectively, superscript cc indicates a dimensionless value based on concentrations, and CG and CL are the concentrations of the flavor in the gas and vapor phases, respectively. Binary HLC has been tabulated for various components,3 and the database is still expanding. By definition, in a binary water− flavor system at equilibrium, the chemical potential (μ) of flavor i is equal in both phases.

fi ̂

L

ij = xiγifi 0 expjjj j k

∫

(3)

where p, pT, and ps are the given, total, and saturated pressures, respectively, R is the gas constant, T is the temperature, x and y are mole fractions in the liquid and gas phases, respectively, and γ is the activity coefficient. The partition coefficient based on mole fraction can be defined as follows Ki =

xi

=

L

γ1 ≡

pis γi

γ2* ≡

pT

(4)

f1̂

x 2f10

ij j expjjjj− j k

Hicc

=

ij xi × ρsol yz jj ∑ x × Mw zz j j = k p ×y { =

ij ρsol yz jj ∑ x × Mw zz j{ k j

RT

RT

T

i

pis γi

f2̂

cc x 2H21

L

and the Henry’s solubility coefficient (Hcc) as follows C Li CGi

(7) 0

L

yx

pR

yz viL dpT zzz z RT {

where p is the reference pressure, f i is the liquid reference fugacity of species i, and v is the partial molar volume of the species in the solution. Since water (1) is present in the largest quantity followed by ethanol (3) as a secondary solvent, the symmetric standard state convention is assumed to be valid to find their activity coefficients. This means that fugacity for solvents in their pure state is taken as a reference, complying with Raoult’s law. The flavor (2) is referenced to the state of “infinite dilution” therewith complying with Henry’s law9 through an unsymmetric fugacity referencing method. For our three components, eq 7 now becomes

and by definition

yi

pT

R

(2)

ij yp yz μi 0 (p , T ) + RT lnjjjj i sT zzzz = μi 0 (p , T ) + RT ln(γixi) j p z k i {

(6)

where x is the mole fraction of components, and β0, β1, and β2 are adjustable parameters.6 As indicated in the literature,7 this approach does not lead to an invariance problem for a system consisting of water, ethanol, and ethyl acetate (one of our selected flavors), and we expect that given the low flavor concentration used, this will hold for all flavors under investigation. The Gibbs energy and hence chemical potential are defined in relation to internal energy and entropy for which absolute values are unknown but can be approached using the fugacity concept.8 For a system at the equilibrium condition, the fugacity in the liquid phase fLî is defined as

(1)

μiG = μi L

Article

γ3 ≡

f3̂

x3f 30

∫

ij j expjjjj− j k

jij expjjjj− j k

pT

p1s

∫

p1s

∫

pT

p1s

y v1dP zzz zz RT zz {

pT

(8)


(9)


(10)

The asterisk denotes that the activity coefficient is normalized using the unsymmetric convention. The reference pressure for all components is taken as equal to the dominant solution vapor pressure, which is that of water. For solute molecules, O’Connell−Prausnitz10 showed that

(5)

where ρ is the solution density, and Mw is the molar mass of compounds. Equation 5 will be used in this work to derive HLC from activity coefficients available in the literature and databases. To determine activity coefficients, excess Gibbs free energy is one of the useful approaches. Wohl presented a general expression for the Gibbs free energy based on a power series expansion of the effective volume fractions of the solution combinations.4 Several methods have been developed based on simplified Wohl expansion, such as Wilson, van Laar, UNIQUAC, Margules, etc., in which interaction parameters (A) of components i and j are used. In this work, Wohl’s expansion is coupled with the one-parameter (two-suffix) Margules equation for a ternary system of water (1), flavor (2), and ethanol (3) and defined as follows5

γ2* = γ2 × exp( −A12 )

(11)

and A 23 = A12 +

cc H23 cc H21

(12)

Since for flavor compounds, by definition, G /RT = ∑i[xi ln (γi)], integration of eq 6 gives E

ln γ2* = A12 [x1(1 − x 2) − 1] + A 23x3(1 − x 2) − A13x1x3 (13)

By assuming x2 = 0, substitution of eq 12 in eq 13 yields B



Article

Table 1. Chemicals and Their Properties chemical name

CAS

purity (GC)

molecular wt (g/mol)

boiling point (°C)

log Pa

supplier

ethanol (SeccoSolv) ethyl acetate (Chromasolv) acetone (SupraSolv) isoamyl alcohol (Emsure) isoamyl acetate (Emplura) water

64-17-5 141-78-6 67-64-1 123-51-3 123-92-2

≥99.9 ≥99.7 ≥99.8 ≥99 ≥99

46.07 88.11 58.08 88.148 130.19 18.013

78.37 77.1 56 131 141 100

−0.31 0.73 −0.24 1.42 2.25

Merck Sigma-Aldrich Merck Merck Merck Milli Q-Plus system

a

Experimental data.11

Figure 1. Effect of ethanol on HLC of (a) ethyl acetate, (b) isoamyl acetate, and (c) isoamyl alcohol regressed by eq 15 using experimental data. (d) Ethanol−water interaction parameter A13 as a function of ethanol mole fraction. Markers are experimental data, and the values beneath them are concentrations as volume percentage.

lim ln γ2* = x3 ln

x2 → 0

cc H23 cc − A13x1x3 H21

is based on the fact that the partition coefficient of a volatile compound in a solution is not a function of the volume of the solution. However, a larger volume of a solution makes volatiles more concentrated in the headspace. This change in concentration can be detected by various headspace analyses. By using a gas chromatography method, the change in the reciprocal value of peak areas against the ratio of the vial total volume over the liquid phase volume becomes a linear plot with a slope of a′ and intercept of b′. As described by Kolb and Ettre,12 HLC can be determined through

(14)

By recalling the definition of the activity coefficient of the flavor in the solution (eq 9) and considering Hcc 2M = f 2/x2 in a dilute system, HLC of the flavor Hcc 2M can be found using the following equation cc x1 cc x3 H2ccM = (H21 ) × (H23 ) × exp( −A13x1x3)

(15)

Equation 15 suggests that the HLC of a flavor is a function of the mole fraction of both water and ethanol as well as solvent−solvent interaction (A13), which in itself is a function of ethanol concentration.

H cc =

2. MATERIAL AND METHODS 2.1. Chemicals. All chemical are listed in Table 1 and used as supplied without further purification. 2.2. Static Headspace Analysis. All HLC were determined using phase ratio variation (PRV). This method

a′ b′

(16)

Therefore, in this work, various sample volumes of 0.1, 0.2, 0.5, 1, 2, and 5 mL were transferred to standard 20 mL headspace screw-neck vials supplied by VWR and were incubated for 60 min. After that, 1 mL of the headspace sample was taken by a CombiPAL autosampler equipped with C



Article

Table 2. Parameters for Eq 15 at T = 30 °Ca experimental values component ethyl acetate isoamyl acetate isoamyl alcohol

i

xflavor −05

−06

9.182 × 10 (1.836 × 10 ) 6.015 × 10−05 (1.203 × 10−06) 8.269 × 10−05 (1.654 × 10−06)

ii

predicted data iv

Hcc 21

Hcc 23

γ21iii

γ23iii

Hcc 21

100.49 (5.53) 36.10 (1.79) 1125.31 (12.01)

965.3 1181.3 8237.6

63.7 2578.8 147.6

6.6 76.4 20.1

107.3 34.4 1142.2

v, iii

Hcc 21

128.4 62.4 1722.7

v, iii

Hcc 23

3197.5 35453.7 192787.7

γ21v

γ23v

66.4 1945.8 129.2

2.66 3.42 1.15

a The parameters were determined using (i) experimental data, (ii) extrapolated eq 5, (iii) eq 5, (iv) the literature,20 and (v) UNIFAC (Dortmund) standard uncertainty values for experimental points, u(Hcc 21).

Figure 2. Binary HLC for (a) ethyl acetate, (b) isoamyl acetate, and (c) isoamyl alcohol in water as a function of temperature.

Henry solubility coefficients of flavor compounds with corresponding regression lines and ethanol−water interaction parameters A13 (using eq 15). In the range of 0−0.05 mole fractions, the effect of ethanol on retention of flavors is minor, which was in accordance with Conner et al.13 who reported that activity coefficients for esters were not affected significantly by ethanol concentrations below 17% (v/v) ≈ (0.0622 mole fraction), whereas they decreased at higher ethanol concentrations, which means a higher Henry’s solubility constant (Hcc) in our case. They attribute this to the formation of ethanol clusters that reduce hydrophobic interactions, leading to partitioning into these ethanol-rich clusters.14−18 The value of the interaction parameter A13 showed a fairly linear increase with ethanol volume fraction. Please note that for all components, attention must be given due to the fact that the Hcc 23 values are extrapolated (ultimately to a volume fraction of 1 for ethanol) as illustrated in the appendix (Figure S1). This co-determines the slope of A13; however, because we did not see any significant difference in the value of A13 when comparing flavors, we expect that these values are reliable. It is important to note that applying binary parameters for

a Hamilton-Gastight 1002 syringe and was injected into the same gas chromatograph (GC) mentioned earlier. The syringe was heated 10 °C above the injection temperature to avoid condensation of vapor. To create the −100 °C trap required for analysis, the GC was coupled with a CryoFocus-4 cold trap. The initial temperature of the GC oven was kept at 40 °C for 30 s and was increased to maximum 160 °C with a temperature-increase rate of 10 °C/s. We used DB-WAXetr, a high-polarity polyethylene glycol (PEG) column from Agilent with a flame ionization detector (FID). All flavor concentrations in this study were produced by mixing 0.5 mL of flavor, taken using a 1 mL (±1%) GSM gas tight syringe, with water in a 1000 mL (±0.4) volumetric flask, creating 0.05 vol %. HLC was determined at 30, 40, 50, and 60 °C for the binary flavor−water systems.

3. RESULTS AND DISCUSSION 3.1. Effect of Ethanol on Henry Solubility Coefficients of Flavors. Figure 1 is constructed by experimental data and the proposed model (eq 15) with parameters presented in Table 2. It illustrates the effect of ethanol concentration on D



Article

Table 3. Experimental HLC Data Using Headspace Analysis Binary Water−Flavor at Different Temperaturesa Hcc(−) component

xflavor

30 °C

40 °C

50 °C

60 °C

ethyl acetate isoamyl acetate isoamyl alcohol

9.182 × 10−05 (1.836 × 10−06) 6.015 × 10−05 (1.203 × 10−06) 8.269 × 10−05 (1.654 × 10−06)

100.49 (12.12) 36.097 (5.79) 1125.31 (412.01)

62.98 (8.84) 20.52 (2.68) 593.56 (117.95)

40.89 (5.10) 10.77 (1.26) 308.20 (38.63)

26.55 (6.03) 5.65 (1.09) 160.03 (39.52)

The standard uncertainties for the temperature of the incubator u(T) = 0.5 °C and u(Hcc) are given in the table.

a

where Hcc⊖ is the HLC at the reference temperature T⊖, ΔsolH is the enthalpy of dissolution, and R is the universal gas constant. The negative or positive value in the exponent depends on how the HLC is defined. When applying nonlinear regression to the experimental data (Figure 2) using the leastsquares method, the values tabulated in Table 4 are obtained.

multicomponent systems by extending usual quadratic mixing rules in equations of state to higher-order polynomials potentially suffers from the fact that these models are not invariant when a component is divided into two or more identical subcomponents.19 There are modified mixing rules for multicomponent systems to overcome this problem.7 In this work, composition A13 serves not only as a so-called solvent−solvent interaction parameter in the model but also carries structural characteristics of an ethanol−water mixture. Even though we can argue about its degree of linearity, we believe that this parameter is immune to the abovementioned complications because of two main reasons. First, flavor-matrix systems are so dilute that they hardly have any interactions with the solvents and each other; therefore, their mixtures can be considered as ternary. Second, dissimilarity between the two solvents (water and ethanol) and solute (flavor) is large enough that it does not fall in the category of what is called “identical subcomponents” addressed by Michelsen and Kistenmacher.19 3.2. Temperature Dependence of Henry’s Law Constant. Temperature is known to significantly alter the HLC particularly for those components with a low enthalpy of dissolution. Figure 2 shows HLC of the flavor compounds at four experimental temperatures, 30, 40, 50, and 60 °C, together with their corresponding regressed and predicted data. As also shown in Table 3, the HLC has a higher standard deviation at low temperatures, and this may originate from the relatively low solubility of the flavors, which influences equilibrium quality. Figure 2a compares our experimental data with that of Kutsuna et al.21and Fenclová et al.,22 both using a column-stripping method. The method of Hilal et al.,20 which uses SPARC (SPARC Performs Automated Reasoning in Chemistry)23 vapor pressure coupled with activity coefficient models, relatively underestimates the HLC of ethyl acetate in water; however, it is in good agreement with the prediction method of Mackay et al.,24 which is based on the ratio of vapor pressure over the solubility of isoamyl acetate in water. Our data, especially at higher temperatures, is in better agreement with Meylan and Howard’s25 predictive methods, which are based on bond contribution values. In the absence of experimental data, we compared three different predictive methods: those of Nirmalakhandan and Speece26 using quantitative structure−activity relationship (QSAR), Kühne et al.27 using their novel model based on two-dimensional structure for organic compounds, and Hilal et al.,20 which has already been mentioned above. Classically, the temperature dependency of the Henry’s law constant is described by the approaching van ’t Hoff introduced for equilibrium constants. The definition of Hcc was as follows i Δ Hi1 1 yy H cc(T ) = H cc ⊖ × expjjjj( −) sol jjj − ⊖ zzzzzzz R kT T {{ k

Table 4. Henry’s Law Constants at the Reference Temperature (25 °C) for the Binary Water−Flavor System Regressed from Gas−Liquid Equilibrium Data Given in Table 3 this work Hcc⊖

Δsol H R

ethyl acetate

123

4500

isoamyl acetate

49

5500

isoamyl alcohol

1628

7000

component

literature Hcc⊖

Δsol H R

reference

146 154 89 45 64 64 1710 1834 1140

5900 5500 4800 5000 5000 5000 7600 8200 7600

21 28 20 25 24 20 26 27 20

To investigate the effect of experimental uncertainties on the uncertainty of driven parameters using eq 17 the concept of the propagation of uncertainties29 is applied and presented in Table 5.

4. CONCLUSIONS The applied predictive method allows us to describe air−water partition coefficients of flavors in ethanoic solutions. The retention of isoamyl acetate and ethyl acetate increased slightly with increasing amount of ethanol, whereas these effects were Table 5. The Uncertainty of Propagation for Driven Data (see Appendix B) component

ethyl acetate

isoamyl acetate

isoamyl alcohol

(17) E

temp [°C]

W (Hcc⊖)

30 40 50 60 30 40 50 60 30 40 50 60

6.09 18.46 16.64 29.56 2.11 6.63 5.38 7.63 45.91 368.98 245.90 469.40

W

(

Δsol H R

)

1057.73 883.73 487.05 647.02 1048.24 830.74 463.49 552.75 684.35 1252.84 497.34 705.89



Article

(11) Souza, E. S.; Zaramello, L.; Kuhnen, C. A.; Junkes, B. D. S.; Yunes, R. A.; Heinzen, V. E. F. Estimating the Octanol/Water Partition Coefficient for Aliphatic Organic Compounds Using SemiEmpirical Electrotopological Index. Int. J. Mol. Sci. 2011, 12, 7250− 7264. (12) Kolb, B.; Ettre, L. S. Static Headspace-Gas Chromatography: Theory and Practice; John Wiley & Sons: 2006, 308−311. (13) Conner, J. M.; Birkmyre, L.; Paterson, A.; Piggott, J. R. Headspace Concentrations of Ethyl Esters at Different Alcoholic Strengths. J. Sci. Food Agric. 1998, 77, 121−126. (14) Cipiciani, A.; Onori, G.; Savelli, G. Structural Properties of Water-Ethanol Mixtures: A Correlation with the Formation of Micellar Aggregates. Chem. Phys. Lett. 1988, 143, 505−509. (15) D’Angelo, M.; Onori, G.; Santucci, A. Self-Association of Monohydric Alcohols in Water: Compressibility and Infrared Absorption Measurements. J. Chem. Phys. 1994, 100, 3107−3113. (16) Price, W. S.; Ide, H.; Arata, Y. Solution Dynamics in Aqueous Monohydric Alcohol Systems. J. Phys. Chem. A 2003, 107, 4784− 4789. (17) Asenbaum, A.; Pruner, C.; Wilhelm, E.; Mijakovic, M.; Zoranic, L.; Sokolic, F.; Kezic, B.; Perera, A. Structural Changes in EthanolWater Mixtures: Ultrasonics, Brillouin Scattering and Molecular Dynamics Studies. Vib. Spectrosc. 2012, 60, 102−106. (18) Gereben, O.; Pusztai, L. Hydrogen Bond Connectivities in Water-Ethanol Mixtures: On the Influence of the H-Bond Definition. J. Mol. Liq. 2016, 220, 836−841. (19) Michelsen, M. L.; Kistenmacher, H. On CompositionDependent Interaction Coefficeints. Fluid Phase Equilib. 1990, 58, 229−230. (20) Hilal, S. H.; Ayyampalayam, S. N.; Carreira, L. A. Air−Liquid Partition Coefficient for a Diverse Set of Organic Compounds: Henry’s Law Constant in Water and Hexadecane. Environ. Sci. Technol. 2008, 42, 9231−9236. (21) Kutsuna, S.; Chen, L.; Abe, T.; Mizukado, J.; Uchimaru, T.; Tokuhashi, K.; Sekiya, A. Henry’s Law Constants of 2,2,2Trifluoroethyl Formate, Ethyl Trifluoroacetate, and Non-Fluorinated Analogous Esters. Atmos. Environ. 2005, 5884. (22) Fenclová, D.; Blahut, A.; Vrbka, P.; Dohnal, V.; Böhme, A. Temperature Dependence of Limiting Activity Coefficients, Henry’s Law Constants, and Related Infinite Dilution Properties of C4-C6 Isomeric n-Alkyl Ethanoates/Ethyl n-Alkanoates in Water. Measurement, Critical Compilation, Correlation, and Recommended data. Fluid Phase Equilib. 2014, 375, 347−359. (23) Hilal, S.; Karickhoff, S. Verification and Validation of the SPARC Model; US Environmental Protection Agency, 2003, No. March, 44. (24) Mackay, D.; Shiu, W.-Y.; Ma, K.; Lee, S. C. Handbook of Physical-Chemical Properties and Environmental Fate for Organic Chemicals; 2nd Edition, Introduction and Hydrocarbons, CRC press: 2006; Vol. 3. (25) Meylan, W. M.; Howard, P. H. Bond Contribution Method for Estimating Henry’s Law constants. Environ. Toxicol. Chem. 1991, 10, 1283−1293. (26) Nirmalakhandan, N. N.; Speece, R. E. QSAR Model for Predicting Henry’s Constant. Environ. Sci. Technol. 1988, 22, 1349− 1357. (27) Kühne, R.; Ebert, R.-U.; Schüürmann, G. Prediction of the Temperature Dependency of Henry’s Law Constant from Chemical Structure. Environ. Sci. Technol. 2005, 39, 6705−6711. (28) Fenclová, D.; Dohnal, V.; Vrbka, P.; Laštovka, V. Temperature Dependence of Limiting Activity Coefficients, Henry’s Law Constants, and Related Infinite Dilution Properties of Branched (C3 and C4) Alkanols in Water. Measurement, Critical Compilation, Correlation, and Recommended Data. J. Chem. Eng. Data 2007, 52, 989−1002. (29) Kline, S. J.; Mcclintock, F. A. Describing Uncertainties in a Single Sample Experiment. Mech.Eng. 1953, 75, 3−8.

much stronger for isoamyl alcohol, which eventually were completely retained in ethanol. We found that the ethanol concentration dependency of parameter A13 plays a pivotal role in describing Henry’s law constants and found a similar dependency for the flavors tested, which may help translate our findings to those of other flavors. The observed dependency of A13 may originate from the structural changes reported by others.13−17 As expected, temperature affected the partition coefficient of flavors, which we successfully covered through a van ’t Hoff approach.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jced.8b01125.

■

Henry’s solubility coefficient of flavors (ethyl acetate, isoamyl acetate, and isoamyl alcohol) graphed and tabulated for different water−ethanol concentrations; the propagation of uncertainties for the regressed parameters used in temperature dependency of HLC (PDF)

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: +31(0)317 483396. ORCID

Ali Ammari: 0000-0003-2557-0037 Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS This project was funded by the Institute for Sustainable Processes (ISPT) in the Netherlands. REFERENCES

(1) Sellbyville, D. Worldwide Non-alcoholic Beer Market worth over $ 25 billion by 2024: Global Market Insights, Inc. https://globenewswire. com/news-release/2018/03/20/1442488/0/en/Worldwide-Nonalcoholic-Beer-Market-worth-over-25-billion-by-2024-Global-MarketInsights-Inc.html (accessed Sep 16, 2018). (2) Brányik, T.; Silva, D. P.; Baszczyňski, M.; Lehnert, R.; Almeida E Silva, J. B. A Review of Methods of Low Alcohol and Alcohol-Free Beer Production. J. Food Eng. 2012, 108, 493−506. (3) Sander, R. Compilation of Henry’s Law Constants (Version 4.0) for Water as Solvent. Atmos. Chem. Phys. 2015, 15, 4399−4981. (4) Wohl, K. Thermodynamic Evaluation of Binary and Ternary Liquid Systems. Trans. Am. Inst. Chem. Eng. 1946, 42, 215−249. (5) Villamañań , R. M.; Martín, M. C.; Villamañań , M. A.; Chamorro, C. R.; Segovia, J. J. Thermodynamic Properties of Binary and Ternary Mixtures Containing Di-Isopropyl Ether, 2-Propanol, and Benzene atT= 313.15 K. J. Chem. Eng. Data 2010, 55, 2741−2745. (6) Jackson, S. L. Extension of Wohl’s Ternary Asymmetric Solution Model to Four and n Components. Am. Mineral. 1989, 74, 14−17. (7) Mathias, P. M.; Klotz, H. C.; Prausnitz, J. M. Equation-of-State Mixing Rules for Multicomponent mixtures: The Problem of invariance. Fluid Phase Equilib. 1991, 67, 31−44. (8) Smith, J. M. Introduction to Chemical Engineering Thermodynamics. J. Chem. Educ. 1950, 584. (9) Brelvi, S. W.; O’Connell, J. P. Prediction of Unsymmetric Convention Liquid-phase Activity Coefficients of Hydrogen and Methane. AIChE J. 1975, 21, 157−160. (10) O’Connell, J. P.; Prausnitz, J. M. Thermodynamics of Gas Solubility in Mixed Solvents. Ind. Eng. Chem. Fund. 1964, 3, 347−351. F



■

Article

NOTE ADDED AFTER ASAP PUBLICATION This paper was published ASAP on July 17, 2019, with incorrect Supporting Information. The corrected version was reposted on July 19, 2019.

G


Effect of Ethanol and Temperature on Partition Coefficients of Ethyl

Recommend Documents