Concentration by pervaporation of brown crab volatile compounds from dilute model solutions: Evaluation of PDMS membrane

Concentration by pervaporation of brown crab volatile compounds from dilute model solutions: evaluation of PDMS membrane Rodrigo Martínez, María Teresa Sanz ∗ , Sagrario Beltrán Department of Chemical Engineering, University of Burgos, 09001 Burgos. Spain Abstract Pervaporation experiments with PERVAPTM 4060 membrane have been performed to study its ability to concentrate some aroma compounds identified in the brown crab boiling juice from a model dilute aqueous solution. The effect of feed concentration, pervaporation temperature and permeate pressure on the pervaporation performance of the membrane has been analyzed. The results obtained with PERVAPTM 4060 membrane were compared with experiments performed in a previous work with a POMS/PEI membrane. Membrane material seems to be a key factor in the permeate aroma profile. The presence of non-volatile compounds, such as sodium chloride, in brown crab boiling juice slightly improves pervaporation performance in the concentration range studied in this work. Batch pervaporation experiments prove that the flavor loss percentage during long operation time is high, especially for the most ∗ Corresponding author. Tel.: +34 947 258810. Fax: +34 947 258831. E-mail address [email protected] volatile compounds. Additionally, it has been shown that pervaporation can be significantly improved by modifying the permeant circuit by means of two stage condensation step. Keywords: volatile compounds, pervaporation, salt effect, PDMS membrane 1. Introduction Shellfish flavour is a high value food product. The liquid effluent of the seafood industry, produced during boiling process, contains important amounts of flavour components [1]. Brown crabs are found in the Eastern Atlantic and are heavily exploited commercially, being available throughout the year. This work is part of a wider study to consider the conversion of the brown crab effluent produced during boiling into a valuable volatile concentrate. Membrane separation techniques represent a potential pathway for the production of a natural aroma concentrate and a disposable effluent [2, 3]. In a previous study [4], organophilic pervaporation through a polyoctylmethylsiloxane membrane was shown to be able to concentrate some compounds of the brown crab boiling juice from a model dilute aqueous solution. In this work, pervaporation through a polydimethylsiloxane membrane has been performed to compare the ability of the two polymers, POMS and PDMS, to recover some aroma compounds found in the brown crab boiling juice from model dilute aqueous solutions. The effect of some important variables that determine the final aroma profile, such as feed concentration, pervaporation temperature and permeate pressure has been studied. Brown crab boiling juice contains organic compounds and non-volatile compounds. The influence of the presence of non-volatile components, such as salts (e.g.: sodium chloride) was also analysed by varying the salt content in the feed solution. For industrial applications, batch operation is preferred to a continuous operation if the aroma recovery is a short-term operation with relatively small amount of extracts [5]. In this work, batch process has been performed to recover aroma compounds from model solutions. Finally, fractionation experiments have been also performed to improve the pervaporation performance by using a two condensation step at the permeate side. 2. Theory Solution/diffusion model is used to describe the transport of permeating components through the membrane being proportional to the difference in partial vapor pressure at both sides of the membrane according to equation 1 [6]: ( ) pi s iii i ,OVipypxQJ −γ= (1) where Ji is the partial permeation flux of component i, QOVi its pressure-normalized permeation flux (permeance), xi its mole fractio in the feed, γ i its activity coefficient, s i p its saturation vapor pressure at the feed temperature, yi, its mole fraction in the permeate and pp the permeate pressure. In case of pervaporation of dilute aqueous solutions, activity coefficients at infinite dilution ( ∞ γi ) are used as feed-side activity coefficients. In this work, the activity coefficients at infinite dilution in water were obtained from previous work in which activity coefficients at infinite dilution were obtained by using the headspace gas chromatography technique [7]. The separation performance of a pervaporation membrane, in case of dilute aqueous solutions, can be described in terms of the enrichment factor. The enrichment factor of a given component is the relationship between its concentration in the permeate and the feed: f,ip,ii ww=β (2) In dilute systems, as aroma recovery systems, the solvent enrichment factor is close to one; therefore aroma enrichment factors of organic compounds can be considered equal to the corresponding separation factors. There is usually an opposite trend between permeation flux and separation factor, i.e.: when one factor increases the other decreases. This way, a pervaporation separation index (P.S.I.) has been defined as a measure of the separation ability of a membrane [8]: P.S.I. = Jtot · Separation factor (3) To describe the batch operation systems, the following expression inferred by She and Hwang [5] was used: t V C K a x x ln o tot L o,i i                       + −=         (4) She and Hwang [5] introduced the term KL to consider the flavor loss rate and represents how quickly the organic compound is lost due to several reasons such as system leakage, partial condensation or incomplete collection in the pervaporation system. In Eq. 4 xi is the mole fraction of component i in the feed tank, xi,o is the initial mole fraction of component i at the beginning of the process, Vo is the initial volume of the feed solution, and t is the operation time. The parameter a is defined as: V tot i s ii,OV K C ApQ a− γ = ∞ (5) where Ctot is the total mole concentration of the feed solution (for dilute systems Ctot becomes approximately the pure water molar density), A is the membrane area and KV is the total permeation volume flow rate (dV = -KVdt). According to She and Hwang [5] the residue percentage and recovery percentage are calculated by equations (6) and (7) respectively: () () 100 ·t V K 1%residue V toti , L K/CKa1 O V ++        −= (6) ( ) ( ) ( ) 100·t V K 11 KCKa1 Ka1 %erycovre Vtoti,L K/CKa1 O V Vtoti,L V                −− ++ + = ++ (7) 3. Experimental section 3.1. Materials An organophilic dense membrane was used in this study: PERVAPTM 4060 (Sulzer Chemtech, Switzerland), a membrane whose active layer is based on polydimethylsiloxane (PDMS). A multicomponent dilute aqueous solution was prepared with seven selected volatile compounds [4] belonging to different chemical classes: 1-octen-3-ol (Sigma Aldrich, 98 %), 1-penten-3-ol (Sigma Aldrich, 99 %), 3-methylbutanal (Sigma Aldrich, 97 %), hexanal (Sigma Aldrich, 98 %), benzaldehyde (Sigma Aldrich, >=99 %), 2,3-pentadione (Sigma Aldrich, 98 %) and ethyl acetate (Sigma Aldrich, HPLC grade). Some thermodynamic properties of the selected compounds are listed in Table 1. 3.2. Pervaporation experiments Pervaporation experiments were performed with a plate and frame laboratory stainless steel permeation cell (Sulzer Chemtech) with an effective membrane area in contact with the feed mixture of 170 cm2 [9]. The temperature of the feed liquid mixture was kept constant (± 0.5 ºC) by using a thermostat to heat the stirred tank feed reactor. The permeate was condensed on two parallel glass cold traps cooled by liquid nitrogen to ensure that permeate was fully collected. Permeate pressure was regulated with an airinlet located between the condensers and the vacuum pump. For steady state operation the feed reactor has a 5 L capacity. This way, due to the small amount of permeate product, the concentration of the volatile compounds in the feed tank was kept approximately constant along operation. The feed flow rate was set to 70 kg·h-1. This flow was enough to avoid concentration polarization in the concentration range studied in this work. For unsteady state operation the ratio membrane area to initial feed volume (A/Vo) was higher than for steady state operation; thus feed concentration of volatile compounds continuously decreases as pervaporation takes place. In case of fractionation at the permeate side, two condensers were placed in series in one of the two parallel permeate circuits. The first condenser was cooled by a refrigerant bath using a Julabo FP50 cryostat. The second condenser was cooled with liquid nitrogen, acting as a total condenser. The chemical stability of the membrane was checked by measuring pure water flux at reference operating conditions. 3.3. Sample analysis Permeate and feed concentration was measured off-line using a Hewlett Packard (6890) gas chromatograph (GC) equipped with series connected thermal conductivity (TCD) and flame ionization (FID) detectors. Helium, 99.999 % pure, was used as carrier gas. The GC column was a 007 FFAP 25 m × 0.25 mm bonded phase fused silica capillary column. The injector and detectors were kept at 200 ºC and 250ºC respectively. The oven was operated at programmed temperature, from 40ºC to 220ºC. 1-hexanol was used as internal standard for analysis of the sample. 4. Results and discussion. First, the effect of different operating variables, such as feed concentration, pervaporation temperature and permeate pressure, on the pervaporation performance of PERVAPTM 4060 membrane is presented. Further, these results are compared with a previous work [4], where pervaporation was carried out with a POMS/PEI membrane. Afterwards, the results obtained in batch operation are presented. Then, the effect of sodium chloride is analysed by varying the salt concentration in the feed. Finally some preliminary fractionation results by using two condensation steps are discussed. 4.1. Evaluation of PERVAPTM 4060 membrane performance. Effect of feed concentration The effect of feed concentration was investigated in the range of 0.1 to 50 ppm at a fixed pervaporation temperature (26ºC) and permeate pressure (300 Pa). Since experiments were carried out at very dilute concentration, total permeation flux was close to water permeation flux and remained almost constant (0.0054 mol/m2s) whatever the volatile compounds feed concentration. Figure 1 shows the organic compounds permeation fluxes as a function of its feed concentration. It can be observed a linear relationship for all the compounds studied in this work (r2 > 0.99 except for 2,3-pentanedione that shows a r2 = 0.97). Since activity coefficients at infinite dilution are constant, and the polymeric membrane swelling can be considered negligible, the pressure-normalized permeation flux (permeance) of the organic compounds and water, and consequently the separation factor, will remain constant. Table 2 reports the enrichment factors and permeances obtained through the PDMS membrane. Benzaldehyde was found to have the highest enrichment factor and permeance. This way, Baudot and Marin [10] assessed that pervaporation membranes are very permselective for aldehydes containinig a benzene ring, which enhances their hydrophobicity and, consequently, their solubility in the pervaporation membrane. Ethyl acetate and 1-octen-3-ol also exhibit high separation factors. On the other hand, the selectivity for 2,3-pentanedione was found to be the lowest, with a separation factor value of 10. Effect of pervaporation temperature The operating temperature was varied in the range 26 ºC to 40 ºC. The experiments were carried out at three different feed concentrations (1, 5 and 10 ppm) and a fixed permeate pressure (300 Pa). Water permeation flux increased exponentially from 0.0054 mol/m2s to 0.012 mol/m2s in the temperature range studied in this work. Figure 2 shows the effect of temperature on volatile compounds permeation flux at an organic feed concentration of 10 ppm. The effect of feed temperature can be described by an Arrhenius type equation:        −= RT E exp·JJ i,a o,ii (8) where Ea,i is the apparent activation energy of permeation, Ji,o the preexponential factor and T the absolute temperature. Table 3 reports the apparent activation energies for permeation of aroma compounds and water for the PDMS membrane. This parameter characterizes the overall effect of temperature on the permeability and the driving force for permeation [11]. A higher value of the apparent activation energy indicates a more sensitive behaviour towards temperature changes. Table 3 shows that Ea,i for most of the organic compounds is higher than that of water, except for benzaldehyde and ethyl acetate. However, these differences are not high. Therefore, it cannot be generalized for PDMS membrane that enrichment factors of volatile compounds increase with temperature, although for some of the volatile compounds, such as 1-octen-3ol, the increase of the enrichment factor with temperature is remarkable. This behaviour can be observed in Figure 3 that shows the enrichment factor at the four temperatures studied in this work at a fixed feed concentration of 10 ppm for all the volatile compounds. The effect of temperature on selectivity depends on changes of sorption of organic compounds on the membrane and its diffusion through the membrane. According to Feng and Huang [11] the activation energy of permeation, Ep, which characterizes the temperature dependence of the membrane, can be estimated by subtracting the heat of vaporization from the calculated apparent activation energy (Ep = Ea - ΔHv). The calculated activation energy follows the order: Ep,1-penten-3-ol (-8.8 kJ/mol) < Ep,ethylacetate (- 8.6 kJ/mol) < Ep,benzaldehyde (-5.3 kJ/mol) < Ep,hexanal (-1.8 kJ/mol) < Ep, water ( -0.5 kJ/mol) < Ep,1-octen-3-ol (10.5 kJ/mol) < Ep,3-methylbutanal (17.0 kJ/mol) < Ep,2,3pentadione (28.1 kJ/mol). Negative values of Ep indicate that the membrane’s permeability decreases with increasing temperature. Ep can be expressed as the activation energy of permeating compounds to diffuse through the membrane, ED, plus the enthalpy of dissolution, ∆Hs (Ep = ED + ∆Hs) [11]. Enthalpy of dissolution is usually negative due to exothermic Kujawski and Krajewski [17] proposed the applicability of the empirical Setschenov equation: si o,i i ckln =         γ γ ∞ ∞ (9) to describe the increase of organic permeation flux with salt concentration. In Eq. 9 ∞ γi is the activity coefficient at infinite dilution in salty solutions, ∞ γo,i the activity coefficient at infinite dilution in pure water, cs the salt concentration and the proportionality factor, ki, is the salting-out coefficient. This way the ln (Ji, salty solutions/Ji,non-saltysolution) can be plotted as a function of NaCl concentration and compared with data of the activity coefficients obtained in previous work [7]. Table 5 lists the slopes of these plots together with the values previously reported in the study of the activity coefficients at infinite dilution. It can be observed that the values of the slopes obtained for the ln (Ji, salty solutions/Ji,non-saltysolution) were lower than the values previously reported [7]. Therefore, blocking effect cannot be considered negligible. As an example Figure 10 represents the applicability of the Setschenov equation for ethyl acetate. In this Figure, the values of the activity coefficients at infinite dilution obtained in a previous work have been also plotted. 4.5. Multistage condensation of the aroma compounds. In section 4.1 enrichment factors were reported in the range of 10 for 2,3-pentanedione to 39 for ethyl acetate. Marin et al. [18] proposed to modify the downstream section with a multi-stage condensation step to improve the pervaporation performance. In this work, some experiments have been performed by using two condensers placed in series. The components of the permeate vapour have different condensation potentials, thus yielding an additional separation factor [19]. The distribution of water and organic compound in each condenser depends on the temperature of the first condenser, the volatility of the components, the flow of the condensable and inert gases in the permeate side and the stripping effect between condensers [20, 21]. Pervaporation experiments have been carried out at a fixed temperature of 26ºC and permeate pressure of 300 Pa. Feed concentration was about 1 ppm for all the organic compounds. Two different temperatures were tried in the first condenser -4 ºC and - 10ºC. Table 6 shows the permeation flux percentage collected in both condensers as a function of temperature in the first condenser. As it can be observed, the percentage collected in the first condenser increases as the condensation temperature in the first condenser decreases. Due to the efficiency in water removal in the first condenser, enrichment factors of the organic compounds collected in the second condenser improve considerably (Table 7) compared to one condenser. Table 7 also reports the separation factor that would be obtained on the basis of vapour-liquid equilibrium (VLE) at 40ºC. This relative volatility at infinite dilution, ∞ α w,i , is defined as: 𝛼𝛼𝑖𝑖,𝑤𝑤 ∞=𝑝𝑝𝑖𝑖 𝑠𝑠 𝑝𝑝𝑤𝑤 𝑠𝑠𝛾𝛾𝑖𝑖 ∞ (10) where pis and piw are the saturation vapour pressure of organic compounds and water respectively and 𝛾𝛾𝑖𝑖 ∞ is the activity coefficient at infinite dilution. The separation factors obtained with one condenser were lower than the separation that would be achieved on the basis of VLE. However, with a two condensation step the new separation factors obtained are, for some of the compounds, even higher than the values of the relative volatility. Concentration for the most volatile compounds in the permeate collected in the first condenser was very low. That indicates that most of the permeant molecules of these organic compounds were mainly collected in the second condenser probably due to a stripping effect. However, the concentration of the aromatic compounds in the second condenser is less than expected based on mass global balance with the results obtained with one condenser. This fact can be due to the very high loss rates found for these compounds in the studies performed in non-steady state (section 4.3). The non-steady state experiments revealed that the concentration of the high volatile compounds is largely conditioned by the time employed in each run due to the high value of the flavor loss rate. The operation time used in the fractionation experiments was longer than in the experiments carried out with one total condenser, due to the fact that insufficient quantities for the analysis were collected in the second condenser, especially in the experiments carried out at the lowest temperature in the first condenser. 5. Conclusion PERVAPTM 4060 membrane has been found to be effective to recover some key aroma compounds found in the brown crab boiling juice from a model dilute aqueous solution. The effect of feed concentration, pervaporation temperature and permeate pressure on pervaporation performance has been analyzed. An increase of the aroma feed concentration increases organic permeation flux due to an increase in the permeation driving force. Partial and total permeation flux increased with temperature and permeation flux decreased as permeate pressure increased. The results obtained under steady state operation with PERVAPTM 4060 were compared with those one obtained with a POMS/PEI membrane. Total permeation flux was higher for PERVAPTM 4060 membrane than for POMS/PEI. Based on the different effect of temperature for both types of membranes it was concluded that sorption contribution was more important for PERVAPTM 4060 membrane than for POMS/PEI membrane. From batch pervaporation experiments, it was found that the loss of volatile compounds increases with operation time. Loss percentage was high for volatile compounds with low permeation rate and high flavour loss rate. Higher membrane area would help to reduce the loss percentage. The presence of NaCl slightly improved the efficiency of aroma recovery from dilute aqueous solutions. This fact can be attributed to the “salting out”. However the increase in partial permeation flux is lower than the corresponding “salting out” effect concluding that blocking effect cannot be neglected. Modifying the permeate circuit with a two condensation step can be an efficient way of improving the pervaporation performance, since it has been observed a considerable increase of the enrichment factors regarding one condensation step. However, the design must be optimized since important losses of some of the volatile compounds, especially the most volatiles, have been observed. Further studies will be carried out with the industrial brown crab boiling juice to study the ability of pervaporation technique. Nomenclature a = permeation rate constant, m3·s-1 (Eq. 4) A = membrane area, m2 Ctot = total molar concentration in liquid phase, mol·m-3 (Eq. 4) Ea = apparent activation energy of permeation, kJ·mol-1 H = Henry’s law constant, Pa J = permeation flux, mol·s-1·m-2 k = salting out coefficient KL = flavor loss rate, mol·s-1 (Eq. 4) KV = total permeation volume flow rate, m3·s-1 (Eq. 5) QOV = pressure-normalized permeation flux, mol·s-1·m-2·Pa-1 p = pressure, Pa P.S.I. = pervaporation separation index R = gas constant, kJ·mol-1K-1 T = absolute temperature, K V = volume of feed solution, m3 x, y = mol fraction t = time, s γ = activity coefficient β = enrichment factor subscripts i = component p = permeate f = feed upperscripts s = saturation ∞ = infinite dilution Acknowledgments Financial support from the MICINN through CTQ2008-04999-PPQ is gratefully acknowledged. R. Martinez acknowledges the JCyL Education Ministry. References [1] Y.J. Cha, K.R. Cadwallader, H.H. Baek, Volatile flavor components in snow crab cooker effluent and effluent concentrate, J. Food Sci. 58 (1993) 525-530. [2] S. Cros, B. Lignot, P. Bourseau, P. Jaouen, Reverse osmosis for the production of aromatic concentrates from mussel cooking juices: a technical assessment, Desalination 180 (2005) 263-269. [3] S. Cros, B. Lignot, P. Bourseau, P. Jaouen, Technical and economical evaluation of an integrated membrane process capable both to produce an aroma concentrate and to reject clean water from shrimp cooking juices, J. Food. Eng. 77 (2006) 697-707. [4] R. Martínez, M. T. Sanz, S. Beltrán, Concentration by pervaporation of representative brown crab volatile compounds from dilute model solutions, J. Food. Eng. 105 (2011) 98-104. [5] M. She, S. T. Hwang, Recovery of key components from real flavor concentrates by pervaporation, J. Membr. Sci. 279 (2006) 86-93. [6] I. Blume, J.G. Wijmans, R.W. Baker, The separation of dissolved organics from water by pervaporation, J. Membr. Sci. 49 (1990) 253-286. [7] R. Martínez, M.T. Sanz, S. Beltrán, E. Corcuera, Activity coefficients at infinite dilution of aroma compounds in water: effect of temperature and salt concentration, J. Chem. Eng. Data 57 (2012) 1480-1485. [8] R.Y.M. Huang, J.W. Rhim, Separation characteristics of pervaporation membrane separation processes, in Elsevier (Ed.), Pervaporation Membrane Separation Processes, Amsterdam, 1991, pp. 111-180 [9] P. Delgado, M.T. Sanz, S. Beltrán, Pervaporation of the quaternary mixture present during the esterification of lactic acid with ethanol, J. Membr. Sci. 332 (2009) 113-120. [10] A. Baudot, M. Marin, Pervaporation of aroma compounds: comparison of membrane performances with vapour-liquid equilibria and engineering aspects of process improvement, Food Bioprod. Process. 75 (1997) 11-142. [11] X. Feng, R.Y.M. Huang, Estimation of activation energy for permeation in pervaporation processes, J. Membr. Sci. 118 (1996) 127-131. [12] V. García, E. Pongracz, E. Muurinen, R.L. Keiski, Pervaporation of dichloromethane from multicomponent aqueous systems, J. Membr. Sci. 326 (2009) 92102. [13] M. She, S.-T. Hwang, Effects of concentration, temperature, and coupling on pervaporation of dilute flavor organics, J. Membr. Sci. 271 (2006) 16-28. [14] J. Olsson, G. Tragardh, Influence of temperature on membrane permeabilibity during pervaporative aroma recovery, Sep. Sci. Technol. 34 (1999) 1643-1659. [15] R. Clément, Z. Bendjama, Q. T. Nguyen, J. Neel, Extraction of organics from aqueous solutions by pervaporation. A novel method for membrane characterization and process design in ethyl acetate separation, J. Membr. Sci. 66 (1992) 193-203. [16] F. Lipnizki, S. Hausmanns, R.W. Field, Influence of impermeable components on the permeation of aqueous 1-propanol mixtures in hydrophobic pervaporation, J. Membr. Sci. 228 (2004) 129-138. [17] W. Kujawski, S. R. Krajewski, Influence of inorganic salt on the effectiveness of liquid mixtures separation by pervaporation, Sep. Purif. Technol, 57 (2007) 495–501. [18] M. Marin, C. Hammami, D. Beaumelle, Separation of volatile organic compounds form aqueous mixture by pervaporation with multi-stage condensation, J. Food Eng. 28 (1996) 225-238. [19] L.M. Vane, F.R. Alvarez, A.P. Mairal, R.W. Baker, Separation of vapor-phase alcohol/water mixtures via fractional condensation using a pilot-scale dephlegmator: enhancement of the pervaporation process separation factor, Ind. Eng. Chem. Res. 43 (2004) 173-183. [20] F. Gascons Viladomat, I. Souchon, V. Athés, M. Marin, Membrane air stripping of aroma compounds, J. Membr. Sci. 277 (2006) 129-136. [21] A. Baudot, M. Marin, Improved recovery of an ester flavor compound by pervaporation coupled with a flash condensation, Ind. Chem. Eng. Res. 38 (1999) 44584469. Table 1. Physicochemical properties of the volatile compounds. (MW = molecular weight, BP = boiling point, ps = saturation pressure, γ∞ = activity coefficient at infinite dilution, H = Henry’s constant). Compound MW, g/mol BP, ºC p26ºC 𝑠𝑠, Pa γ26º𝐶𝐶 ∞ H 26ºC ·10-4, Pa 1-Octen-3-ol 128.2 174-5 65 1779 11.6 1-Penten-3-ol 86.1 114.4 1348 38 5.1 3-Methylbutanal 86.1 92.5 7035 127 89.3 Benzaldehyde 106.1 178.8 181 504 9.1 2,3-Pentanedione 100.1 108 2918 133 38.8 Hexanal 100.2 128.3 1540 599 92.3 Ethyl Acetate 88.1 77.1 13045 62 80.9 Table 2. Pervaporation parameters of PERVAPTM 4060. Comparison of PERVAPTM 4060 and POMS/PEI membranes (TPV = 26ºC, pp = 300 Pa). PERVAP TM 4060 POMS/PEI Compound β QOV,i·107 (mol/m2sPa) PSI (mol/m2s) β QOV,i ·107 (mol/m2sPa) PSI (mol/m2s) 1-Octen-3-ol 36 18.5 2.0∙10 -1 120 8.6 8.8∙10 -2 1-Penten-3-ol 25 29.6 1.4∙10 -1 25 3.1 1.9∙10 -2 3-Methylbutanal 18 1.1 1.0∙10 -1 5 0.03 3.9∙10 -3 Benzaldehyde 51 35.8 2.8∙10-1 93 7.5 7.5∙10-2 2,3-Pentanedione 10 1.4 6.8∙10 -2 7 0.1 5.2∙10 -3 Hexanal 23 1.4 1.3∙10 -1 22 0.2 1.7∙10 -2 Ethyl Acetate 39 2.7 2.2∙10-1 7 0.04 5.6∙10-3 Water 17.5 2.0 Figure 1. Effect of volatile feed concentration on volatile compound permeation flux (TPV = 26 ºC, pp = 300 Pa) for PERVAPTM4060. 0.E+00 1.E-06 2.E-06 3.E-06 0.E+00 3.E-06 6.E-06 9.E-06 Ji (mol/m2s) xi (mol/mol) 1-octen-3-ol 1-penten-3-ol 3-methylbutanal benzaldehyde 2,3-pentanedione hexanal ethyl acetate Figure 2. Effect of pervaporation temperature on volatile compound permeation flux (Ci. feed ≈ 10 ppm, pp = 300 Pa) for PERVAPTM4060. -18 -17 -16 -15 -14 -13 -12 -11 3.15 3.2 3.25 3.3 3.35 3.4 ln Ji (mol/m2s) 1000/T (K) 1-octen-3-ol 1-penten-3-ol 3-methylbutanal benzaldehyde 2,3-pentanedione hexanal ethyl acetate Figure 3. Effect of pervaporation temperature on the separation factor of the volatile compounds (Ci.feed ≈ 10 ppm, pp = 300 Pa) for PERVAPTM4060. 0 10 20 30 40 50 60 70 Enrichment factor T=26ºC T = 30ºC T = 36ºC T = 40ºC Figure 4. Effect of pervaporation temperature on the flux and driving force ratio for PERVAPTM4060 (Cfeed ≈ 10ppm, pp = 300 Pa). DF =driving force of the process. 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 J40ºC/J26ºC and DF40ºC/DF26ºC J40ºC/J26ºC DF40ºC/DF26ºC Figure 5. Effect of permeate pressure on partial permeation flux for PERVAPTM4060 (Ci. feed ≈ 1 ppm. Tfeed = 26ºC). The continuous lines are to guide the eye. 0.E+00 2.E-08 4.E-08 6.E-08 8.E-08 1.E-07 0500 1000 1500 2000 Ji (mol/m2s) pp (Pa) 1-octen-3-ol 1-penten-3-ol 3-methylbutanal benzaldehyde 2,3-pentanedione hexanal ethyl acetate Figure 6. Effect of permeate pressure on the enrichment factors of the volatile compounds for PERVAPTM4060 (Ci. feed ≈ 1 ppm. T=26°C). 0 10 20 30 40 50 60 70 Enrichment factor 300 Pa 400 Pa 600 Pa 900 Pa 1200 Pa 1800 Pa Figure 7. Effect of membrane material on the permeate aroma profile (T = 26ºC, Ci = 10 ppm pp = 300 Pa. 1-octen-3-ol 1-penten-3-ol 3-methylbutanal benzaldehyde2,3-pentanedione hexanal ethyl acetate Feed solution PDMS, 26ºC, 300 Pa POMS, 26ºC, 300 Pa Figure 8. Evolution of feed concentration over time in the pervaporative recovery of aroma compounds (TPV = 26ºC, pp = 300 Pa, t = 11 h). The continuous lines represent the model proposed by She and Hwang [5]. -7 -6 -5 -4 -3 -2 -1 0 0 5 10 ln x i /x i,0 t (h) 1-octen-3-ol 1-penten-3-ol 3-methylbutanal benzaldehyde 2,3-pentanedione hexanal ethyl acetate Figure 9. Influence of salt concentration on the enrichment factor of the organic compounds for PERVAPTM4060 (Cfeed = 1ppm, Tfeed = 40ºC. Pp = 300 Pa). 0 10 20 30 40 50 60 70 80 Enrichment factor 0 mol/kg 0.34 mol/Kg 0.43 mol/kg 0.51 mol/kg 0.85 mol/kg Figure 10. Comparison between the effect of salt concentration on ethyl acetate partial permeation flux (○) and activity coefficients at infinite dilution (▲). 0.0 0.2 0.4 0.6 0.8 1.0 1.2 00.5 11.5 2 ln (Ji,salty/Ji,non-salty), ln(γi∞/γi,o∞) CNaCl (mol/kg)