SciELO - Scientific Electronic Library Online

 
vol.16 issue1PREPARATION OF TANTALUM CARBIDE FROM AN ORGANOMETALLIC PRECURSORA PUBLISHED KINETIC MODEL EXPLAINS THE VARIATION IN NITROGEN CONTENT OF Pichia guilliermondii DURING ITS BATCH CULTIVATION ON DIESEL OIL author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand

Journal

Article

Indicators

Related links

Share


Brazilian Journal of Chemical Engineering

Print version ISSN 0104-6632On-line version ISSN 1678-4383

Braz. J. Chem. Eng. vol. 16 n. 1 São Paulo Mar. 1999

http://dx.doi.org/10.1590/S0104-66321999000100002 

HIGH-PRESSURE VAPOR-LIQUID EQUILIBRIUM DATA FOR BINARY AND TERNARY SYSTEMS FORMED BY SUPERCRITICAL CO2, LIMONENE AND LINALOOL

 

S. A. B. VIEIRA DE MELO*, P. PALLADO, G. B. GUARISE and A. BERTUCCO
Università di Padova, Istituto di Impianti Chimici
Via Marzolo, 9, I-35131, Padova (PD) - Italy

 

(Received: June 25, 1998; Accepted: November 23, 1998)

 

 

Abstract - The feasibility of deterpenating orange peel oil with supercritical CO2 depends on relevant vapor-liquid equilibrium data because the selectivity of this solvent for limonene and linalool (the two key components of the oil) is of crucial importance. The vapor-liquid equilibrium data of the CO2-limonene binary system was measured at 50, 60 and 70oC and pressures up to 10 MPa, and of the CO2-linalool binary system at 50oC and pressures up to 85 bar. These results were compared with published data when available in the literature. The unpublished ternary phase equilibrium of CO2-limonene-linalool was studied at 50oC and up to 9 MPa. Selectivities obtained using these ternary data were compared with those calculated using binary data and indicate that a selective separation of limonene and linalool can be achieved.
Keywords: vapor-liquid equilibrium, supercritical CO2, orange oil, experimental data.

 

 

INTRODUCTION

Orange peel oils have an important role as a flavoring ingredient in the food and perfumery industries. They are a complex mixture of more than 200 components which can be grouped in three fractions: hydrocarbon terpenes (unsaturated compounds) responsible for off-flavor characteristics; oxygenated compounds, which give the desirable citrus fresh accent; and nonvolatile residues.

Prior to industrial processing, these oils have to be deterpenated (removal of terpenes) to improve their solubility in alcoholic solvents and to prevent autoxidation of the terpenes to undesirable off-note substances (Barth et al., 1994).

Supercritical carbon dioxide is an alternative procedure for deterpenating orange peel oil due to its advantages over traditional techniques, since mild temperatures are required and it leaves no solvent residues (Temelli et al., 1988)

Orange peel oil is considered a binary mixture of its two key compounds, limonene and linalool, representing the terpene and the oxygenated fractions, respectively (Temelli et al., 1990; Sato et al., 1996)

To evaluate the feasibility of using SC-CO2 to separate limonene from linalool, it is necessary to obtain vapor-liquid equilibrium (VLE) data at different conditions of pressure and temperature.

Binary phase equilibrium data and solubility data for limonene and/or linalool with supercritical CO2 can be found in the literature (see Table I), although not all of them meet the VLE needs for process calculation. In a previous work presented elsewhere (Vieira de Melo et al., 1996), preliminary vapor-liquid equilibrium data were shown for the CO2+limonene and CO2+linalool binary systems at 60oC and at 50oC, respectively, and pressures up to 10 MPa. Note in Table I that only liquid-phase molar fractions were reported by Vieira de Melo et al. (1996) for these systems, but in the present work the respective vapor-phase molar fractions are also given. Thus, the liquid-phase molar fraction values that appear in Tables III and IV have already been published.

 

Table1: Literature review of vapor-liquid equilibrium for carbon dioxide-limonene and carbon dioxide-linalool systems.

Reference

T range (oC)

P range (bar)

NDP

Remarks

CO2-Limonene

       

Stahl and Gerard (1985)

10-50

40-120

-

g

Matos et al. (1989)

45-50

85-98

18

x*, y*

Di Giacomo et al. (1989)

35-50

30-100

26

y

Gironi and Lamberti (1995)

42

24-85

7

y

Marteau et al. (1995)

37-50

70-98

38

x*, y*

Iwai et al. (1996)

40-60

40-102

15

x, y

Vieira de Melo et al. (1996)

60

75-100

4

x

CO2-Linalool

       

Iwai et al. (1994)

40-60

40-110

15

x, y

Vieira de Melo et al. (1996)

50

66-95

3

x

 

 

Table 2: Vapor-liquid equilibrium for CO2-limonene at 50oC. P is total pressure; x and y are liquid and vapor molar fractions of CO2, respectively.

P (bar)

x (CO2)

y (CO2)

70.0

0.5654

0.9984

75.9

0.6176

0.9978

80.3

0.6594

0.9963

85.6

0.6876

0.9964

96.8

0.8690

0.9913

97.8

0.8966

0.9933

 

 

Table 3: Vapor-liquid equilibrium for CO2-limonene at 60oC. P is total pressure; x and y are liquid and vapor molar fractions of CO2, respectively.

P (bar)

x (CO2)

y (CO2)

75.4

0.5253

0.9931

81.8

0.5782

0.9984

89.0

0.6099

0.9943

100.7

0.6626

0.9939

 

 

Table 4: Vapor-liquid equilibrium for CO2-limonene at 70oC. P is total pressure; x and y are liquid and vapor molar fractions of CO2, respectively.

P (bar)

x (CO2)

y (CO2)

51.4

0.3212

0.9979

69.6

0.4488

0.9951

83.7

0.5058

0.9966

89.5

0.5263

0.9948

93.9

0.5845

0.9898

105.5

0.6358

0.9723

 

In this work, additional phase equilibrium data are reported for the CO2+limonene binary system at 50, 60 and 70oC and the CO2+linalool binary system at 50oC and at pressures up to 10 MPa, and for the CO2+limonene+linalool ternary system at 50oC and pressures up to 9 MPa. The selectivity values calculated from the ternary VLE experimental results were compared with the values of the ratio between the partition coefficients obtained from the cited binary systems. These values indicate that it is possible to obtain a selective separation of limonene and linalool. Nonetheless, this does not exclude the possibility of using an entrainer like ethanol to enhance this selectivity.

Remarks: NDP: number of data points; x: liquid-phase mass or molar fractions are given; y: vapor-phase mass or molar fractions are given; g: data are given on graphs only; x*: liquid-phase molar fractions are given without vapor-phase molar fractions at the same tie-line; y*: vapor-phase molar fractions are given without liquid-phase molar fractions at the same tie-line.

 

EXPERIMENTAL

Apparatus

The high-pressure vapor-liquid equilibrium measurements are made using a circulation-type apparatus, as shown schematically in Figure 1. The main piece of the equipment is an equilibrium cell of approximately 100 cm3 (i.d. = 25 mm and h = 200 mm), placed in a temperature-controlled water-bath. This cell has two ports at the bottom and two at the top, which allow one or both phases to circulate. Recirculation is achieved by a compressed-air actuated piston-pump located inside the water bath so as to have the same temperature as the cell. All the relative lines and valves are also placed in this bath to avoid thermal disturbances. The bath is operated by manual control so as to minimize cell temperature fluctuations to within one tenth of one degree during an experimental run. System temperatures are monitored by a platinum thermoresistance sensor (PT100W ) calibrated within ± 0.1oC, located in the cell and connected to a millivolt multimeter. Pressure readings are made by a Data Instruments model AB-5000 pressure transducer with an accuracy of 0.1 bar.

 

Image7.gif (16897 bytes)

Figure 1: High-pressure vapor-liquid equilibrium circulation-type apparatus.

 

Each experimental run begins by evacuating the equilibrium cell and the relative lines, once the desired temperature has been attained. The water bath was turned on on the previous day. First, a small amount (about 20 cm3) of the liquid heavy component or a binary mixture (for ternary experiments) is loaded into the cell through valves V7 and V3. Carbon dioxide is pumped into the cell from a cooled storage unit until the desired pressure is reached. All feeds enter the apparatus through the same line.

After all components have been fed into the cell and the apparatus is at the desired temperature and pressure, the air-compressed pump is started up to circulate the vapor phase for a period of about 30 minutes. Then, the settling period begins to ensure complete separation of the phases in the cell. It usually lasts approximately 30 to 60 minutes. Once the two phases are well separated, cell temperature and pressure are recorded and the system is ready for sampling.

Sampling is initiated by evacuating the liquid and vapor sample loops connected to the capillary sample lines. A small amount of vapor phase is removed through the three-way valve V5 to the sample loop of the six-way valve V9. It is assumed that the portion of this phase contained in the recirculation line is representative of the vapor phase in equilibrium with the liquid phase inside the cell. The liquid phase sample is taken directly from the bottom of the cell through valve V6. To ensure that a representative sample is removed, the capillary line between the cell and this valve is purged and after this it is filled with a liquid whose composition near is that of equilibrium. After the purged material between valve V6 and six-way valve V13 has been removed, the liquid sample system is ready for sampling. Moreover, valve V6 is placed as close as possible to the cell.

Small pressure drops during sampling are inevitable, since the cell is not a variable-volume one. The loop volumes of 1 cm3 and 0.5 cm3 for the vapor and liquid samples, respectively, are small enough to ensure a negligible pressure drop in the cell while sampling. A maximum 0.5 bar pressure drop accompanies vapor sample removal, and a maximum 0.2 bar drop results for liquid sample removal. Furthermore, liquid samples are taken first because this causes less disturbance to the cell than taking of vapor samples. The pressure drop does not appear to affect the composition of the phases in equilibrium, since the reproducibility of the results is checked with good agreement.

Therefore, the option of a simpler fixed volume cell, instead of a more complex variable volume one as in this case, does not considerably affect theaccuracy of results. After the sampling period is completed, valves V9 and V13 are switched to allow both samples to decompress through micrometer valves V11 and V15. All the valves and lines related to the sampling and expansion sections are heated properly to prevent any condensation. Decompression begins by opening the expansion valves very slowly. The decompressed samples are transferred to glass traps placed quite close to the expansion valves, cooled at –20oC with a salt-ice mixture and filled with a calibrated amount of appropriate organic solvent (n-octane) to completely recover the heavy component. The organic solvent has been previously saturated with CO2 to avoid its dissolution in the sample. The expanded-CO2 volume is measured by means of a burette connected to a liquid reservoir at atmospheric pressure. Water saturated with CO2 is chosen to prevent its absorption. After recording the temperature and volume of the gas, supercritical CO2 is flowed from the by-pass line through the sampling loops and the expansion valves to remove and completely recover the heavy component from the loops.

The solution of n-octane containing the heavies (limonene and/or linalool) is analyzed in a GC model 14-A Shimadzu using a Flame Ionization Detector (FID) and a Chromosorb Column W-17W SA-476 with a dimethyl siliconic fluid as a stationary phase.

Materials

Limonene (97%) and n-octane (99%) were purchased from Aldrich Co. and linalool (98%) was supplied by Teranol Co. (Lalden, Switzerland). These materials were not further purified. Carbon dioxide with a high degree of purity (99.9%, Rivoira srl) was used as received.

 

RESULTS AND DISCUSSION

Following the experimental procedure described above, VLE data for CO2-limonene were measured at 50, 60 and 70oC. These results are shown in Tables II, III and IV, respectively. Tables V and VI give binary and ternary VLE data at 50oC for CO2-linalool and CO2-limonene-linalool, respectively.

Figure 2 shows the Pxy phase diagram for CO2-limonene at 50oC, and a portion of this diagram is zoomed in Figure 3 to allow a better view of the vapor-phase data points. The present VLE data are in good agreement with those reported by Matos et al. (1989) and very recently by Iwai et al. (1996).

 

Image8.gif (6802 bytes)

Figure 2: Binary vapor-liquid equilibrium for CO2(1)-limonene(2) at 50oC: (X1: ¡ , Y1: l ) present work; (X1: +,Y1: x) Matos et al. (1989); (X1: n ,Y1: o ) Iwai et al. (1996). X1 and Y1 are liquid and vapor molar fractions, respectively.

 

 

Image9.gif (6830 bytes)

Figure 3: Binary vapor-liquid equilibrium for CO2(1)-limonene(2) at 50oC: (Y1: ¡ ) present work; (Y1: +) Matos et al. (1989); (Y1: o ) Iwai et al. (1996). Y1 is the vapor molar fraction.

 

Figures 4 and 5 give representative binary phase diagrams of CO2-limonene at 60 and 70oC, respectively. VLE data published by Iwai et al. (1996) are also included for comparison in Figure 4. It is shown that there are some divergences between their data and these given here, mainly at more elevated pressures. Precision of composition values is directly influenced by disturbances in equilibrium during sampling. An alternative to reduce pressure drop while sampling is to withdraw a smaller amount of each phase, but this implies later difficulties in detecting the amount of solute (often very small) removed from the vapor phase.

 

 

Image10.gif (6133 bytes)

Figure 4: Binary vapor-liquid equilibrium for CO2(1)-limonene(2) at 60oC: (X1: ¡ , Y1: l ) present work; (X1: n , Y1: o ) Iwai et al. (1996). X1 and Y1 are liquid and vapor molar fractions, respectively.

 

 

Image11.gif (6042 bytes)

Figure 5: Binary vapor-liquid equilibrium for CO2(1)-limonene(2) at 70oC: (X1: n , Y1: o ) present work. X1 and Y1 are liquid and vapor molar fractions, respectively.

 

There are no published data in the literature for comparison at 70oC, but it can be seen in Figure 6 that this isotherm is qualitatively consistent with those at 50 and 60oC.

 

Image12.gif (7180 bytes)

Figure 6: Binary vapor-liquid equilibrium for CO2(1)-limonene(2) at 50 (X1: n , Y1: o ), 60 (X1: l , Y1: ¡ ) and 70oC (X1: +, Y1: x). X1 and Y1 are liquid and vapor molar fractions, respectively.

 

Figure 7 shows the VLE diagram for CO2-linalool at 50oC. The present data are plotted against those of Iwai et al. (1994). There is a good agreement for the liquid phase but some discrepancies were detected for the vapor phase.

 

Image13.gif (6141 bytes)

Figure 7: Binary vapor-liquid equilibrium for CO2(1)-linalool(2) at 50oC: (X1: ¡ , Y1: l ) present work; (X1: n , Y1: o ) Iwai et al. (1994). X1 and Y1 are liquid and vapor molar fractions, respectively.

 

Looking at the equilibrium diagrams presented above, it can be seen that the amount of supercritical solvent that dissolves in the liquid phase is considerable. At constant temperature, raising total pressure of the system increases the amount of CO2 that solubilizes in limonene or linalool and, consequently, increases the degree of expansion of this phase. At a constant pressure, an increase in temperature gives a reduction in the amount of CO2 dissolved in the liquid phase. So, the amount of carbon dioxide that is solubilized in the liquid phase is of utmost importance for designing purposes because it affects the extractor internal volume needed for a given amount of feed material.

For the vapor phase, note that the amount of solute solubilized by the supercritical fluid is noticeable only near the mixture’s critical point. At a constant temperature, raising the pressure increases the density of the supercritical fluid, that is, its solvating power becomes greater and more solute is transferred from the liquid to the vapor phase. At a constant pressure, one observes that raising the temperature increases the solubility of limonene or linalool in the vapor phase. In this case, the reduction in solvent density, that is, the decrease in its solubilization capacity due to an increase in temperature, is overcome by the effect of increasing the solute vapor pressure. Phase diagrams for both the CO2-limonene and the CO2-linalool systems were plotted at 50oC in Figure 8.

 

Image14.gif (6021 bytes)

Figure 8: Binary vapor-liquid equilibrium for CO2(1)-limonene(2) (X1: n , Y1: o ) and CO2(1)-linalool(2) (X1: l , Y1: ¡ ) at 50oC. X1 and Y1 are liquid and vapor molar fractions, respectively.

 

 

Table V: Vapor-liquid equilibrium for CO2-linalool at 50oC. P is total pressure; x and y are liquid and vapor molar fractions of CO2, respectively.

P (bar)

x (CO2)

y (CO2)

74.9

0.6333

0.9987

80.3

0.7052

0.9993

83.7

0.7255

0.9995

 

 

Table VI: Vapor-liquid equilibrium for CO2-limonene-linalool at T=50 oC. P is total pressure; x and y are liquid and vapor molar fractions of CO2, limonene and linalool, respectively. b is selectivity, defined by ylim/ylol / xlim/xlol.

P (bar)

xCO2

xlim

xlol

yCO2

ylim

ylol

b

75.4

0.6592

0.2026

0.1382

0.9987

0.0011

0.0002

3.75

80.8

0.6637

0.2024

0.1339

0.9987

0.0011

0.0002

3.64

87.6

0.5568

0.2312

0.2120

0.9981

0.0014

0.0005

2.57

89.0

0.8265

0.1158

0.0577

0.9979

0.0017

0.0004

2.12

 

The results indicate the possibility of a selective separation of limonene and linalool. For instance, the ratio between the partition coefficients of limonene (klim = ylim/xlim) and linalool (klol = ylol/xlol) at 80.3 bar is 4.54. This factor is often called selectivity, although by definition this term is more appropriate when the 3 components are present in the same mixture.

Selectivity calculated directly from the ternary VLE data for CO2-limonene-linalool at 50oC is plotted versus pressure in Figure 9. At a constant temperature, raising the pressure decreases the selectivity due to an increase in the solvating capacity of the supercritical solvent (solvent higher density effect). The present data (selectivity > 2) are indicative of a possible selective deterpenation of orange peel oil. For instance, the value of selectivity between limonene and linalool at 50oC and 80.8 bar is 3.64, which is less than that calculated from the binary VLE data. This is evidence that molecular interactions between limonene and linalool cannot be neglected. It also suggests that relevant errors can be made when only binary VLE behavior analysis is included in the designing of a supercritical extraction process.

 

Image15.gif (5947 bytes)

Figure 9: Selectivity between limonene and linalool in CO2 at 50oC calculated from ternary VLE data (l ) as shown in table VI.

 

Uncertainty in the results obtained for all systems discussed, defined by the experimental error, was a maximum of 6% in the vapor molar fraction and a maximum of 3% in the liquid molar fraction for all components. These errors were assumed to be mainly due to GC analysis, and this is a common practice for this kind of experimental determination. Further, their values are with in the typical accepted range found in the literature for high-pressure vapor-liquid equilibrium measurements.

 

CONCLUSIONS

The present study of phase behavior for CO2-limonene, CO2-linalool and CO2-limonene-linalool concentrated mainly on two aspects: examination of the possibility of a selective separation of the terpenic (limonene) and oxygenated (linalool) fractions of orange peel oil using supercritical CO2 and obtaining the VLE data needed for modeling and simulating the oil deterpenation process in a supercritical fluid extraction column.

Experimental results show that not only binary VLE data should be considered in modeling multicomponent systems at high pressures, but some attempts also should be made to obtain information from ternary VLE data. For instance, values for selectivity between limonene and linalool from the ternary VLE data are higher than those from the binary VLE data. This fact suggests those from the binary VLE data. This fact suggests that molecular interactions between limonene and linalool cannot be neglected because accurate thermodynamic modeling of this kind of behavior is crucial in a developing separation process.

The selectivity values obtained at 50oC are indicative that a selective separation of limonene and linalool can be achieved.

 

ACKNOWLEDGEMENTS

S.A.B. Vieira de Melo gratefully acknowledges the financial support received in Italy, which was provided by the CNPq (Brazilian Council of Research and Technology).

 

REFERENCES

Barth, D., Chouchi, D., Della Porta, G., Reverchon, E. and Perrut, M., Desorption of Lemon Peel Oil by Supercritical Carbon Dioxide: Deterpenation and Psoralens Elimination, J. Supercritical Fluids, 7, 177 (1994).         [ Links ]

Di Giacomo, G., Brandani, V., Del Re, G. and Mucciante, V., Solubility of Essential Oil Components in Compressed Supercritical Carbon Dioxide, Fluid Phase Equilibria, 52, 405 (1989).         [ Links ]

Gironi, F. and Lamberti, L., Solubility of Lemon Oil Components in Supercritical Carbon Dioxide. Proc. 3rd Italian Conference on I Fluidi Supercritici e le loro Applicazioni, 165, Trieste (1995).         [ Links ]

Iwai, Y., Hosotami, N., Morotomi, T., Koga, Y. and Arai, Y., High-Pressure Vapor-Liquid Equilibria for Carbon Dioxide + Linalool, J. Chem. Eng. Data, 39, 900 (1994).         [ Links ]

Arai, Y., High-Pressure Vapor-Liquid Equilibria for Carbon Dioxide + Limonene, J. Chem. Eng. Data, 41, 951 (1996).         [ Links ]

Marteau, Ph., Obriot, J. and Tufeu, R., Experimental Determination of Vapor-Liquid Equilibria of CO2 + Limonene and CO2 + Citral Mixtures, J. Supercritical Fluids, 8, 20 (1995).         [ Links ]

Matos, H.A., Azevedo, E.G., Simões, P.C., Carrondo, M.T. and Nunes da Ponte, M., Phase Equilibria of Natural Flavours and Supercritical Solvents, Fluid Phase Equilibria, 52, 357 (1989).         [ Links ]

Sato, M., Goto, M. and Hirose, T., Supercritical Fluid Extraction on Semibatch Mode for the Removal of Terpene in Citrus Oil, Ind. Eng. Chem. Res., 35, 1906 (1996).         [ Links ]

Stahl, E. and Gerard, D., Solubility Behavior and Fractionation of Essential Oils in Dense Carbon Dioxide, Perfumer & Flavorist, 10, 30 (1985).

Temelli, F., Braddock, R.J., Chen, C.S. and Nagy, S., Supercritical Carbon Dioxide Extraction of Terpenes from Orange Essential Oil. Supercritical Fluid Extraction and Chromatography, Charpentier, B.A. and Sevenants, M.R., eds., ACS Symposium Series 366, 109, Washington (1988).         [ Links ]

Temelli, F., O'Connell, J.P., Chen, C.S. and Braddock, R.J., Thermodynamic Analysis of Supercritical Carbon Dioxide Extraction of Terpenes from Cold-Pressed Orange Oil, Ind. Eng. Chem. Res., 29, 618 (1990).         [ Links ]

Vieira de Melo, S.A.B., Pallado, P., Bertucco, A. and Guarise, G.B., High-Pressure Phase Equilibria Data of Systems Containing Limonene, Linalool and Supercritical Carbon Dioxide. Process Technology Proceedings (12): High Pressure Chemical Engineering, Elsevier Sci. Publisher, 411 (1996).         [ Links ]

Vieira de Melo, S.A.B., Supercritical CO2 Deterpenation of Orange Peel Oil. Ph.D. diss., PEQ-COPPE, Federal University of Rio de Janeiro (1997) (in Portuguese).         [ Links ]

 

* To whom correspondence should be addressed.

Current address: DEQ-EP, Federal University of Bahia, R. Aristides Novis, 2, Federação, Salvador / BA, 40210-630, Brazil. E-mail: sabvm@ufba.br

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License