SOLUBILITY OF SYRINGIC AND VANILLIC ACIDS IN SUPERCRITICAL CARBON DIOXIDE R. Murga, M.T. Sanz, S. Beltrán* and J.L. Cabeza Departamento de Ingeniería Química. Univeridad de Burgo. Plaza Miael Bañuelo /n. 09001 Burgo. Epaña. Tel.: +34 947 58809. Fax.: +34 947 58831. E-mail: beltran@ubu.e 1. INTRODUCTION Phenolic compound are a wide and heterogeneou group of ubtance that are widely ditributed in vegetable. They are o abundant among uperior plant, that only carbohydrate are on top of them. The interet of thee phenolic compound lie on their known health benefit due to their antioxidant activity and to their ability a free radical cavenger. Thee propertie give them a great potential a active principle in the pharmaceutical indutry and a antioxidant in the food indutry. There i, therefore, an increaing interet in iolating thee compound from their natural matrice. However, mot of them are thermolabile and ome extraction technology that not require high temperature hould be ued. One of the technologie that i being explored i upercritical fluid extraction (SFE), that i becoming an extended operation for eparation of natural ubtance from complex material due to the mild condition needed and other environmental benefit [1]. The efficient deign of any particular SFE proce need the previou knowledge of the olute olubility. In thi work, the olubility of two non-flavonoid, low molecular weight phenolic compound, 4-hydroxy-3,5-dimethoxybenzoic acid (yringic acid) and 4-hydroxy-3- methoxybenzoic acid (vanillic acid), in upercritical CO, at preure from 10 to 50 MPa and temperature from 40 to 60ºC, i preented. Thee two compound preent a Trolox equivalent antioxidant activity (TEAC) of 1.4 mm []. The TEAC i defined a the concentration of Trolox (the water-oluble vitamin E analog) olution with an antioxidant potential equivalent to a 1 mm concentration of the compound under conideration. Two type of model have been ued for correlation of the experimental olubility data: a rigorou thermodynamic method [3] where the Peng-Robinon equation of tate (PR-EOS) wa ued to determine the fugacity of the fluid phae, and two emiempirical denity dependent correlation: a linear correlation between the enhancement factor and the denity of the olvent, a uggeted by Schmitt and Reid [4], and the Chratil correlation [5]. Thee emiempirical correlation are widely ued and, although they are not capable of predicting unknown phae equilibria, they are ueful tool for experimental data correlation. EXPERIMENTAL.1. Chemical The olute ued in thi work, yringic and vanillic acid, were upplied by Sigma, (>98 % purity). The molecular weight (MW) and normal melting point (mp) of thee olute are reported in Table 1 together with the etimated normal boiling point (T b ), critical temperature 1
(T c ) and preure (P c ), and acentric factor (ω). The main chemical tructure of thee compound i repreented next to Table 1 where the R group for each compound i given. Table 1: Phyical propertie of the olute. Solute Syringyc acid Vanillic acid Formula C 9 H 10 O 5 C 8 H 8 O 4 R-group -OCH 3 -H MW 198.18 168.15 mp (101.33 kpa)/k a 480± 483± T b (101.33 kpa)/k b 71.9 840.5 T c /K b 941.40 1105.67 p c /MPa b 4.0 4.5 ω b 1.190 1.3016 a Sigma data bae (www.igma-aldrich.com/aw.nf/mdhelp). b Etimated by the Joback method implemented in PE [6]. R COOH OH OCH 3 The carbon dioxide ued a upercritical olvent (SFC/SFE quality) wa upplied by Air Liquide. HPLC-grade methanol and acetonitrile (Lab-Scan) and puri. p.a. (98 %) formic acid (Fluka) were ued a olvent for ample analye. Gla bead 30/60 meh (Phae Separation) were ued to ditribute the olute in the equilibrium cell, and gla wool (Panreac) wa placed at the top of each of the three tage of the equilibrium cell to prevent olid entrainment... Apparatu and procedure The olubility of the olid wa experimentally determined by the dynamic analytical method in an apparatu that ha been previouly decribed and ued in our laboratory to obtain olubility data of olid olute in SC-CO [7, 8]. Sample analyi wa carried out off-line by uing an HPLC equipped with a diode array detector (DAD) (Hewlett-Packard 1100 erie). The coupling of chromatography and DAD would allow to detect the compound degradation that could take place, or ome of the impuritie of the pure olute that could be preferentially olved by the SC-CO. None of thoe wa the cae with the olute tudied in thi work. Quantification wa made at a wavelength of 80 nm for yringic acid, and 60 nm for vanillic acid with the calibration curve previouly obtained for each compound in the range where the Lambert-Beer law wa valid. 3. RESULTS AND DISCUSSION The experimental olubility data of vanillic acid and yringic acid in SC-CO, at different condition of preure, p, and temperature, T, are plotted in Figure 1, together with their tandard error. The temperature tudied range from 313 to 333 K and the preure from 8.5 to 50 MPa. The reult how that the olubility increaed with preure, at contant temperature, in all cae. The effect of temperature i more complex and retrograde olubility (croover preure) behavior [4] can be oberved for both acid. At preure above the croover preure, the olubility increae with temperature while at preure below the croover preure, the olubility decreae with increaing temperature.
Under the ame condition of temperature and preure, vanillic acid how higher olubility than yringic acid which may be due to the larger polarization of the yringic acid molecule that ha a methoxy group more than the vanillic acid molecule. 100 1.000 y x 10 7 10 1 0,1 (a) y x 10 7 100 10 (b) 10 0 30 40 50 Preión, MPa 10 0 30 40 50 Preión, MPa Figure 1: Experimental olubility of yringic acid (a) and vanillic acid (b) in SC-CO. ( ) 313 K, (?) 33 K, (?) 333 K. The error bar repreent the tandard error of each olubility data. Continuou line repreent the olubility iotherm calculated with the PR-EOS. According to the mot rigorou method for modeling equilibrium data, the olubility of the olid olute in the SCF can be calculated by mean of Eq. (1), which conider a olid phae, formed by the pure olute (), in equilibrium with a fluid phae (here conidered a a dene ga) formed by a mixture of olvent (1) and olute () [3]. y = p φ p φ υ exp ( p p ) RT In eq.(1), p i the aturation (vapor) preure of the pure olid, φ the fugacity coefficient at the ytem preure p, φ the fugacity coefficient at aturation preure p, and υ the olid molar volume, all at the ytem temperature T. The Peng Robinon equation of tate (PR-EOS), whoe parameter for the mixture have been calculated according to the quadratic mixing rule [3], ha been ued for calculating the fugacity coefficient. The PR-EOS give good quantitative fit for a wide variety of ytem, but critical propertie (T c and p c ) and the acentric factor (ω) of olute and olvent are needed for calculating the EOS parameter for the pure component. Thee propertie are available for CO, but not for the olid ubject of thi work; Therefore, they have been etimated by the Joback group contribution method [6] and are lited in Table 1. Thi group contribution method ha alo been ued for prediction of the normal boiling temperature of the olute. The interaction binary parameter, k 1 and l 1, of the PR-EOS have been obtained by fitting the experimental olubility data to Eq. (1). The regreion procedure wa carried out by minimizing the average abolute relative deviation (AARD) between experimental (y exp ) and calculated (y cal ) olubility data, uing the Phae Equilibria 000 (PE 000) program developed by Prof. Brunner group [6]. The value of the interaction parameter k 1, and l 1 for the different temperature are lited in Table together with the correponding AARD value. The olubility curve a predicted by (1) 3
Eq. (1) and the PR-EOS are repreented by continuou line in Fig. 1, where it may be oberved that the model reproduce fairly well the experimental olubilitie far above the critical preure of the olvent, while larger error are found cloe to it. Thi may be due to the large catter that experimental olubility data uually preent in the vicinity of the critical point a a conequence of the high enitivity that denity preent to preure. Table : Reult of the olubility data correlation through the PR-EOS: number of data point ued in the correlation, n, binary interaction parameter, k ij and l ij, and percentage of average abolute relative deviation, AARD. Parameter of PR-EOS Solute n T/K k ij l ij AARD (%) Syringic acid 9 313 0.0670-0.1575 11.05 9 33 0.0575-0.1783 14.78 9 333 0.0493-0.1973 13.41 Vanillic acid 10 313 0.5086 0.4396.44 9 33 0.550 0.4856 5.30 9 333 0.5171 0.4735 4.98 Two emiempirical model, have alo been ued for data correlation. The firt one, propoed by Schmitt and Reid [4] aume a linear functionality between the logarithm of the enhancement factor (ln E), a calculated from Eq. (), and the denity of the olvent, ρ. p / p E = y () The enhancement factor provide a meaure of the extent that preure enhance the olubility of the olid in the ga and it i alway greater than unity. The vapor preure of the olid, neceary to calculate the enhancement factor uing Eq. (), ha been etimated according to the method propoed by Grain [9]. The enhancement factor model i very ueful for data correlation becaue of it implicity and good fit. The reult of data correlation to thi model are preented in Fig and Table 3. It may be oberved in Table 3, that the model reproduce much better the experimental data when the parameter are obtained without including the olubilitie obtained at preure in the vicinity of the critical point of carbon dioxide. ln E 4 0 00 400 600 800 1000 16 3 1 (a) 8 (b) 00 400 600 800 1000 ρ, kg/m 3 ρ, kg/m 3 Figure : Experimental olubility of yringic acid (a) and vanillic acid (b) in SC-CO. ( ) 313 K, (?) 33 K, (?) 333 K. Continuou line repreent the olubility iotherm calculated with the enhancement factor model. ln E 40 36 4
Table 3: Reult of the olubility data correlation conidering a linear dependence between the enhancement factor and the denity of the olvent ( ln E = A + Bρ ): number of data point ued in the correlation, n, parameter A and B, percentage of average abolute relative deviation, AARD, and -quared,. Preure Correlation Parameter Solute n range/mpa T/K A B/(10 m 3 /kg) AARD (%) Syringic acid 9 10-50 313 1.5 1.0 7.86 0.9960 9 10-50 33 11.00 1.0.30 0.9997 9 10-50 333 7.88 1.4 33.60 0.973 8* 15-50 333 11.4 1.0 9.15 0.9914 Vanillic acid 9 8.5-50 313 30.77 0.863 16.06 0.984 8* 10-50 313 9.79 0.97 10.38 0.989 8 10-50 33 8.64 0.871 1.97 0.9885 8 10-50 333 6.47 0.901 7.35 0.9980 * correlation excluding the olubility data obtained at preure in the vicinity of the critical point of carbon dioxide. The econd empirical correlation ued, wa the Chratil correlation [5] that aume the formation of a olvato complex between molecule of the SC-olvent and the olute at equilibrium. Thi model lead to a linear relationhip between the olubility of the olute, C, expreed a (g of olute)/(l of olvent), and the denity of the olvent,?, in g/l, for a given temperature (in K) a follow: ln C = k ln ρ + a / T + b (3) a, b and k are the adjutable parameter of the model. The Chratil equation ha the advantage of having only three parameter to fit all the experimental data, no matter at which temperature they were obtained. Moreover, thi equation doe not require the etimation of the propertie of the pure component. The parameter a, b, and k were obtained by nonlinear regreion of the experimental data uing the Marquardt algorithm, and are reported in Table 4 together with the AARD between experimental and calculated olubility, C. The quality of the correlation i indicated by - quared. A plot of the experimental data and the olubility iotherm calculated with the Chratil model i preented in Fig. 3 where the quality of the correlation can be viually evaluated. Table 4: Reult of the olubility data correlation by the Chratil model: number of data point ued in the correlation, n, parameter of Chratil equation, a, b and k, percentage of average abolute relative deviation, AARD, and -quared,. Preure Parameter of the Chratil equation Solute n range/mpa k a b AARD (%) Syringic acid 7 10-50 7.411-6495 -34.05 10.73 0.9896 Vanillic acid 8 8.5-50 4.641-546 -17.11 9.43 0.998 5
ln C, g/l - -4-6 -8 (a) -4-8 (b) -1 5,6 6,0 6,4 6,8 5,6 6,0 6,4 6,8 ln ρ, kg/m 3 ln ρ, kg/m 3 ln C, g/l Figure 3: Experimental olubility of yringic acid (a) and vanillic acid (b) in SC-CO. ( ) 313 K, (?) 33 K, (?) 333 K. Continuou line repreent the olubility iotherm calculated with the Chratil model. ACKNOWLEDGMENTS Financial upport provided by the CICYT (QUI96-0691), and the EU, CICYT and JCyL (1FD97-1471-QUI) i gratefully acknowledged. REFERENCES [1] MURGA, R., RUIZ, R., BELTRÁN, S., CABEZAS, J.L., J. Agric. Food Chem., Vol. 48, 8, 000, p. 3408-341. [] RICE EVANS, C.A., MILLER, N.J., PAGANGA, G., Free Radical Bio. Med., Vol. 0, 7, 1996, p. 933-956. [3] PRAUSNITZ, J.M., LICHTENTHALER, R.N., GOMES DE AZEVEDO, E., Molecular Thermodynamic of Fluid Phae Equilibria, 3 rd edition, Prentice Hall, Englewood Cliff, N.J. 1999. [4] SCHMITT, W.J., REID, R.C., in PENNINGER, J.M.L., RADOSZ, M., MCHUGH, M.A., KRUKONIS, V.J., (Ed.) Supercritical Fluid Technology. Elevier, Amterdam, 1985, p. 13-137 [5] CHRASTIL, F.J., Phy. Chem., Vol. 86, 198, p. 3016-301. [6] PETKOV, S., PFOHL O., BRUNNER, G., PE A Program to Calculate Phae Equilibria, Herbert Utz Verlag, München. 000. [7] MURGA, R., SANZ, M.T., BELTRÁN S., CABEZAS, J. L., J. Supercrit. Fluid, Vol. 3, 00, p. 113-11. [8] MURGA, R., SANZ, M.T., BELTRÁN S., CABEZAS, J. L., J. Supercrit. Fluid, 003, in pre. [9] LYMAN, W.J., REEHL, W.F., ROSENBLATT, D.H. Handbook of chemical property etimation method, ACS, Wahington, DC, USA. 1990. 6