19536
J. Phys. Chem. B 2005, 109, 19536-19541
Hydrophobicity vs Hydrophilicity: Effects of Poly(ethylene glycol) and tert-Butyl Alcohol on H2O as Probed by 1-Propanol Kumiko Miki,†,§ Peter Westh,‡ and Yoshikata Koga*,§ Department of Liberal Arts and Basic Sciences, College of Industrial Technology, Nihon UniVersity, Narashino, Chiba 275-8575, Japan; Department of Life Science and Chemistry, Roskilde UniVersity, PO Box 260, Roskilde DK-4000, Denmark; and Suitekijuku (Water Drop Institute), VancouVer, B.C., Canada V6R 2P5, and Department of Chemistry, The UniVersity of British Columbia, VancouVer, B.C., Canada V6T 1Z1 ReceiVed: April 29, 2005; In Final Form: August 15, 2005
E The enthalpic interaction between 1-propanol (1P) molecules, H1P-1P , was evaluated in 1P-poly(ethylene glycol) (PEG)-H2O and 1P-tert-butyl alcohol (TBA)-H2O ternary mixtures. The model-free and E experimentally accessible quantity, H1P-1P , indicates the effect of an additional 1P on the actual enthalpic E reflects the situation of 1P in the mixture. It was shown earlier that the composition dependence of H1P-1P E process how 1P modifies H2O. This H1P-1P pattern changes in the presence of a third component, PEG or TBA. The effects of PEG or TBA on the molecular organization of H2O were elucidated from these induced changes. Together with previous similar studies for the effects of methanol (ME), 2-propanol (2P), ethylene glycol (EG), and glycerol (Gly), we suggest a method and hence a possible scaling for sorting out hydrophobicity vs hydrophilicity of these alcohols by the changes induced to the loci of the maxima in E H1P-1P . We show that hydrophilicity scales with the number of oxygen, regardless of whether O is the ether -O- or the hydroxyl -OH. Hydrophobicity also scales with the number of carbon atoms for alcohols without a methyl group. For those with methyl groups, the hydrophobicity seems proportional to the total number of carbon with a different proportionality factor from those without methyl group.
Introduction Aqueous solutions of amphiphiles are omnipresent in living matters and no doubt play important roles in biofunctions. We have made extensive thermodynamic studies on aqueous monools,1-8 the simplest example of amphiphiles. We studied the effect of a hydrophobic moiety on H2O comparatively by keeping the hydrophilic part, -OH, fixed. These works confirmed that an alcohol enhances the hydrogen bond network of H2O in its vicinity (“iceberg formation”). Concomitantly, however, the hydrogen bond probability of bulk H2O away from icebergs is reduced. In water-rich solutions, the hydrogen bond probability is still high enough, so that the hydrogen bond network is connected throughout; i.e., the hydrogen bond percolation is intact. As the concentration of alcohol increases, the hydrogen bond probability of bulk H2O decreases to a threshold upon which the bond percolation is lost, and the solution consists of two kinds of clusters, one rich in H2O and the other in alcohol. Alcohols with larger hydrophobic moieties require smaller mole fractions to bring the system to this threshold. Furthermore, what we call the enthalpic alcoholE alcohol interaction, HAL-AL , determined experimentally is larger at the threshold for alcohols with larger hydrophobic moieties. (The subscript AL stands for alcohol.) From these and other observations, we concluded that the mixing scheme described above in the H2O-rich region is mainly due to “hydrophobicity” of the mono-ols, i.e., the alkyl moiety. †
Nihon University. Roskilde University. Suitekijuku (Water Drop Institute) and The University of British Columbia. * Corresponding author: Tel (604) 822-3491, Fax (604) 822-2847, e-mail
[email protected]. ‡ §
On the other hand, we conducted similar thermodynamic studies on aqueous ethylene glycol (2C/2OH, meaning an alkyl group with two carbons and two hydroxyl groups),9 glycerol (3C/3OH),10,11 and 1,2- and 1,3-propanediols (3C/2OH)12,13 to learn the effect of the hydrophilic moiety, -OH. A general picture has emerged that in the H2O-rich region -OH forms hydrogen bonds with the existing network of H2O. By so doing, a characteristically large fluctuation inherent to liquid H2O is retarded progressively by breaking the H donor/acceptor symmetry and breaking hydrogen bonds away from the solute. At about 0.1 mole fraction for all the poly-ols we studied, the integrity as liquid H2O is lost, and a different mixing scheme, consisting of H2O-rich and alcohol-rich clusters, sets in. In studying the effect of poly-ols, we used the detailed thermodynamic behavior of 1-propanol (1P) as a probe to elucidate the effect of a third component (a poly-ol) in 1P-a third component-H2O systems.10,12,14-18 This three-component method has been proved to be useful in elucidating effects of salts,19,20 an ionic liquid,21 and other compounds10,12,14-18 in addition to the poly-ols mentioned above. Here we apply the same methodology for studying the effect of poly(ethylene glycol) (PEG) on H2O. PEG’s are known to have a strong favorable interaction with H2O and is often attributed to a good match of the inter-oxygen distance in liquid H2O and that between the neighboring ether oxygens in the PEG helicoidal chain.22 We are interested in the effect of the ether O in the chain of -CH2-O-CH2- as opposed to that by the end -OH, by comparing the results with those for ethylene glycol and/or glycerol, which contain the end -OH only. As will be shown below, the effects of both on H2O turned out to be similar. For the purpose of reference, we also apply the same methodology
10.1021/jp0522462 CCC: $30.25 © 2005 American Chemical Society Published on Web 09/21/2005
Hydrophobicity vs Hydrophilicity
J. Phys. Chem. B, Vol. 109, No. 41, 2005 19537
for tert-butyl alcohol (TBA), which is perhaps the mono-ol with the largest hydrophobic moiety that is soluble in the entire composition range, except for 2-butoxyethanol. The latter is not taken into consideration here since it contains one ether oxygen which may complicate the present issue, a comparison between the end -OH and the ether -O-. We briefly summarize the methodology.10,12,14-18 We measure the excess partial molar enthalpy of 1P, HE1P, in 1P-A-H2O by perturbing the amount of 1P only in the mixture and determine the thermal response of the entire system. Namely
HE1P )
( ) ∂HE ∂n1P
(1) nA,nw,p,T
where ni stands for the amount of the ith compornent, i ) 1P, W () H2O), or A. The partial derivative is taken with respect to n1P, keeping all the other variables constant. HE1P is a second-order derivative of G, since HE is basically the first derivative. As eq 1 suggests, it is the response of the system in terms of HE when the amount of 1P, n1P, is perturbed; hence, it is the actual contribution of 1P of the entire system in terms of enthalpy. Or it signifies the actual enthalpic situation of 1P in the system. Experimentally, we approximate the above partial derivative with (δHE/δn1P) by titrating a small amount of n1P, δn1P, into the cell containing the mixture and determine the thermal response δHE. This approximation was shown acceptable when the mole ratio of the titrant to the titrate in the cell is below 0.5%.23 We take one more derivative of HE1P with respect to n1P graphically without resorting to any fitting function and obtain E E . H1P-1P is what we call the 1P-1P enthalpic interaction H1P-1P the third-order derivative of G and it is written as E ≡N H1P-1P
( )
( )
∂HE1P ∂HE1P ) (1 - x1P) ∂n1P ∂x1P
(2)
where x1P is the mole fraction of 1P and N is the total amount of the mixture. The above definition, the first identity of eq 2, E suggests that H1P-1P is the effect of an additional 1P on the actual enthalpic situation of existing 1P. Therefore, it signifies the degree of 1P-1P interaction in terms of enthalpy. As discussed in the earlier studies on binary aqueous mono-ols,1-8 E the x1P dependence of H1P-1P has the unique pattern that will be reproduced below in Figures 2-6 for the zero mole fraction of a third component. On addition of a third component, the E H1P-1P pattern changes. From this induced change we make an attempt at elucidating the effect of the third component on the molecular organization of H2O. The second equality of eq 2 is used to graphically differentiate HE1P. Namely, we draw a smooth curve through all the data points of HE1P vs x1P by the aid of a flexible ruler. We read the HE1P data off the smooth curve drawn at the interval of δx1P ) 0.002. We approximate the partial derivative on the right of eq 2 by (δHE1P/δx1P) with δx1P ) 0.008. The goodness of this approximation was discussed at some length earlier.10 Experimental Section Poly(ethylene glycol)s (Aldrich Chemical, molecular weight 203 and 600) (abbreviated as PEG200 and PEG600), tert-butyl alcohol (ACROS Organics, 99.5%), and 1-propanol (ACROS Organics, 99.5%) were used as supplied. All of these chemicals are hygroscopic. Due care was exercised to avoid contamination from moisture. A stock aqueous solution of poly(ethylene glycol)
Figure 1. Excess partial molar enthalpy of 1-propanol, HE1P, at 25 °C in the mixed solvent of PEG and H2O at various initial concentrations, x0PEG. (a) PEG(200) and (b) PEG(600).
and tert-butyl alcohol was prepared using freshly distilled water immediately after opening a bottle in a dry atmoshphere. For each series of enthalpy measurement, PEG-water or TBAwater mixed solvents were prepared by diluting the stock solution gravimetrically. The excess partial molar enthalpy of 1P, HE1P, was determined at 25 °C using a homemade titration calorimeter of a design similar to an LKB Bromma 8700 calorimeter.24 The volume of mixed solvent in the cell was about 80 cm3, and 1P was successively titrated until the mole fraction of 1P, x1P, reached around 0.11-0.14. The uncertainty in HE1P is estimated as (0.03 kJ mol-1. Results and Discussion In Figure 1, the excess partial molar enthalpies of 1P, HE1P, against x1P are shown for 1P-PEG200-H2O (Figure 1a) and 0 is the initial mole for 1P-PEG600-H2O (Figure 1b). xPEG fraction of PEG in the mixed solvent. Panels a and b of Figure 0 1 indicate that in the water-rich mixed solvent, xPEG < 0.03 for 0 PEG200 (8.4C/5.2O) and xPEG < 0.01 for PEG600(26.4C/
19538 J. Phys. Chem. B, Vol. 109, No. 41, 2005
Miki et al.
Figure 3. Excess partial molar enthalpy of 1-propanol, HE1P, in mixed solvent of TBA and H2O at 25 °C.
E Figure 2. Enthalpic 1P-1P interaction, H1P-1P , eq 2, in 1P-PEGH2O, at 25 °C. (a) PEG(200) and (b) PEG(600).
E Figure 4. Enthalpic 1P-1P interaction, H1P-1P , in 1P-TBA-H2O at 25 °C.
14.2O), the x1P dependences of HE1P seem similar to that for the binary system (1P-H2O). The initial concave increase turns into 0 composiconvexity with an inflection point. At higher xPEG tions than these thresholds, the shape of the curves change gradually in that the initial concavity is absent within the range 0 0 of measurements. At still higher xPEG , i.e., xPEG > 0.12 for 0 E PEG200 and xPEG > 0.04 for PEG600, H1P decreases monotonically. We note in passing that these threshold values for PEG200 are all about 3 times (600/200) as large as those for PEG600. E , will reveal these trends more The next derivative, H1P-1P clearly. Figure 2 shows the enthalpic 1P-1P interaction, E H1P-1P , against x1P. The peaks correspond to the inflection points in Figure 1. Considering the uncertainty shown in the figure, an accurate determination of the locus of the maximum is problematic. Therefore, we drew two straight lines below and above each peak and defined the vertex of these two lines as the point X. In this manner, ambiguity of locating the inflection point may be circumvented. As we showed earlier,1,2 the locus of point X indicates the beginning of the transition of mixing scheme from the water-rich region to the intermediate region. In the water-rich region, the hydrogen-bond network of H2O is
enhanced in the vicinity of 1P, i.e., “iceberg formation”, but at the same time the hydrogen-bond probability of bulk H2O away from the “iceberg”-clad 1P is reduced. As x1P increases, the hydrogen bond probability of the bulk H2O decreases to the bond percolation threshold at point X. Beyond this point, the hydrogen bond network begins to lose its connectivity. The system then consists of two kinds of clusters, one rich in 1P and the other in H2O; i.e., the different mixing scheme is now operative in the intermediate region.1,2 The equivalent results for 1P-TBA-H2O are shown in E . The general trend Figure 3 for HE1P and in Figure 4 for H1P-1P E in the pattern change in H1P-1P shows a clear difference between PEG (Figure 2) and TBA (Figure 4). In terms of the locus of point X, TBA, a more hydrophobic solute than 1P, drives point X toward smaller values of x1P and higher values E , while PEG, more hydrophilic than 1P, reduces the of H1P-1P E and that of x1P, though the latter effect is value of H1P-1P weaker. To facilitate comparison further, we recall our earlier similar studies for 1P-2-propanol (2P)-H2O,16 shown in Figure 5, and for 1P-glycerol (Gly)-H2O,10 shown in Figure 6. Figure 5 indicates that the almost equally hydrophobic 2P as 1P1,2,5 E shifts the pattern of H1P-1P parallel to smaller values of x1P
Hydrophobicity vs Hydrophilicity
J. Phys. Chem. B, Vol. 109, No. 41, 2005 19539
E Figure 5. Enthalpic 1P-1P interaction, H1P-1P , in 1P-2-propanol (2P)-H2O at 25 °C. Reevaluated using the HE1P data from ref 16.
Figure 7. Loci of point X at various initial mole fraction of the third E component: (a) in terms of H1P-1P and (b) in terms of x1P. E Figure 6. Enthalpic 1P-1P interaction, H1P-1P , in 1P-glycerol (Gly)-H2O at 25 °C. Data from ref 10.
E without changing the value of H1P-1P much. This was argued to be the signature of equal hydrphobicity of the two propanol isomers. Namely, the existing 2P in the mixed solvent of 2PH2O already modified H2O in the same manner as 1P, and hence 1P needs a lesser amount in modifying the remaining way to drive the system to point X. Figure 6 shows that the effect of glycerol, a more hydrophilic solute than 1P, on the pattern of E E H1P-1P is primarily to reduce the value of H1P-1P with a minor effect of reducing x1P. A similar study for 1P-urea-H2O14 E indicated that urea shifts the value of H1P-1P without changing the locus of point X in terms of x1P. We argued that urea, with no hydrophobic moiety, connects onto the existing hydrogen bond network of H2O and keeps the hydrogen bond connectivity intact but reduces the degree of fluctuation inherent in liquid H2O, using the results of a study on the fluctuation nature of aqueous solutions.7 Indeed, a simulation study on binary ureaH2O suggested that the presence of urea in H2O rigidifies the dynamics of urea and H2O.25 Thus, it seems to be fair to state E vs x1P pattern that the shift in the locus of point X in the H1P-1P may be used as a signature for hydrophobicity and hydrophiE at point X licity. Namely, a decrease in the value of H1P-1P
suggests hydrophilicity, and that of decreasing x1P is a sign of hydrphobicity. We stress, however, that this measure is only relative to the hydrophobicity or hydrophilicity of the probing 1P. We plot the loci of point X against the initial mole fractions of alcohols in Figure 7. For mono-ols, i.e., ME, 2P, and TBA, it is clear that the effect of hydrophobicity is to reduce the value of x1P (Figure 7b), as expected from the previous findings discussed above. In addition, however, the data for TBA, which is more hydrophobic than 1P used as a probe, show an increase E in the interaction function, H1P-1P , while the less hydrophobic ME lowers its value at point X, as shown in Figure 7a. Thus, within the group of the same hydrophilic moiety, -OH, the E effects on the value of H1P-1P depend on the relative hydrophobicity/hydrophilicity scale. For the group of EG, Gly, and PEG’s, Figure 7a indicates their hydrophilic effects are stronger for larger solutes with more O’s, if the reduction of the value E is due to hydrophilicity. At the same time, the effect of H1P-1P of hydrophobicity, that of reducing the value of x1P at point X, is stronger for larger solutes with more C’s. Hence, we replot Figure 7 with the abscissa of the mole fraction of O or that of C. Namely, for PEG600 (26.4C/14.2O) for example, we used 0 0 x0carbon ) 26.4 xPEG and x0oxygen ) 14.2xPEG , or for TBA (4C/OH) 0 0 0 0 xcarbon ) 4xTBA and xoxygen ) xTBA, etc. The results are shown in
19540 J. Phys. Chem. B, Vol. 109, No. 41, 2005
Miki et al. with the number of carbon atoms in the solute.28 However, for relative ranking of hydrophobicity and hydrophilicity of amphiphiles and ions, the choice of the organic solvent makes a devastating effect on the scale.29,30 Furthermore, the logarithm of the partition coefficient is equal to the excess chemical potential difference of the solute in H2O from that in the chosen organic solvent. We use here the behavior of the third derivatives of G as opposed to the first-derivative quantity, chemical potential; hence, more subtle difference could be detected by the present method. Indeed, the present method distinguishes the methyl group containing solutes from those without. Thus, we suggest here a useful method and a possible scale in sorting out hydrophobicity vs hydrophilicity of amphiphiles in aqueous solutions. This scale would have a close bearing to the kosmotropicity vs chaotropicity of various “cosolutes” in biologically important aqueous solutions.31 More generally, the kosmotropic/chaotropic scale of cations and anions in the Hofmeister effect32,33 could also be related to the present hydrophobicity/hydrophilicity scale. Indeed, preliminary indications of such relation were published recently.19,20 A fuller account of the relationship between the present hydrophobicity/ hydrophilicity scale with the Hofmeister series is forthcoming. Acknowledgment. We acknowledge support by Nihon University, the Carlsberg Foundation, and the Danish Technical Research Council. We also thank Subramanian Iyer and Foon Yip, Department of Chemistry, the University of British Columbia, for their support. Supporting Information Available: Table S1 of excess partial molar enthalpies of 1P in 1P-PEG H2O for PEG(200) and PEG(600); Table S2 of excess partial molar enthalpies of 1P in 1P-TBA-H2O. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes
E Figure 8. Loci of point X. (a) H1P-1P against the oxygen mole 0 fraction, xoxygen. (b) x1P against the carbon mole fraction, x0carbon. See text.
Figure 8. Figure 8a indicates that the hydrophilicity in terms of E at point X for EG, Gly, PEG200, and the decrease in H1P-1P PEG600 seems to scale with the number of O atoms. This suggests that within this methodology the ether -O- and the end -OH in PEG work in a similar manner in modifying H2O in the H2O-rich region. We note in passing that there is a suggestion by a quasi-static neutron scattering that the interaction of the ether oxygen with H2O is stronger than the end -OH because the bonds connected to the latter break at high temperature.26 The hydrophobicity in terms of the decrease in x1P value of point X seems to be scaled by the number of C, and the plots seem to converge into two lines, one with and the other without methyl groups, as shown in Figure 8b. A recent examination indicated that the standard free energies of transfer from water to hexadecane of alkanes, alkenes, alkadienes, and arenes scale with the number of C-H bonds.27 However, rescaling the abscissa to the mole fraction of C-H bonds instead of that of carbon, x0carbon, does not make the two lines in Figure 8b converge. Tanford used the logarithm of the partition coefficient of a solute in H2O and in an organic solvent as a scale for hydrophobicity and showed a linear relationship
(1) Koga, Y. J. Phys. Chem. 1996, 100, 5172. (2) Koga, Y. Netsusokutei (J. Jpn. Soc. Cal., Therm. Anal.) 2003, 30, 54 (available in a pdf form on request to the author). (3) Koga, Y.; Westh, P.; Nishikawa, K. Can. J. Chem. 2003, 81, 150. (4) Hu, J.; Haynes, C. A.; Wu, A. H. Y.; Cheung, C. M. W.; Chen, M. M.; Yee, E. G. M.; Ichioka, T.; Nishikawa, K.; Koga, Y. Can. J. Chem. 2003, 81, 141. (5) Tanaka, S. H.; Yoshihara, H. I.; Ho, A. W.-C.; Lau, F. W.; Westh, P.; Koga, Y. Can. J. Chem. 1996, 74, 713. (6) Liltorp, K.; Westh, P.; Koga, Y. Can. J. Chem., in press. (7) Koga, Y. Can. J. Chem. 1999, 77, 2039. (8) Koga, Y.; Nishikawa, K.; Westh, P. J. Phys. Chem. A 2004, 108, 3873. (9) Koga, Y. J. Solution Chem. 2003, 32, 803. (10) Parsons, M. T.; Westh, P.; Davies, J. V.; Trandum, Ch.; To, E. C. H.; Chiang, W. M.; Yee, E. G. M.; Koga, Y. J. Solution Chem. 2001, 30, 1007. (11) To, E. C. H.; Davies, J. V.; Tucker, M.; Westh, P.; Trandum, Ch.; Suh, K. S. H.; Koga, Y. J. Solution Chem. 1999, 28, 1137. (12) Parsons, M. T.; Koga, Y. J. Phys. Chem. B 2002, 106, 2981. (13) Parsons, M. T.; Lau, F. W.; Yee, E. G. M.; Koga, Y. J. Solution Chem. 2003, 32, 137. (14) To, E. C. H.; Hu, J.;.Haynes, C. A.; Koga, Y. J. Phys. Chem. B 1998, 102, 10958. (15) Matsuo, H.; To, E. C. H.; Wong, D. C. Y.; Sawamura, S.; Taniguchi, Y.; Koga, Y. J. Phys. Chem. B 1999, 103, 2981. (16) Hu, J.; Chiang, W. M.; Westh, P.; Chen, D. H. C.; Haynes, C. A.; Koga, Y. Bull. Chem. Soc. Jpn. 2001, 74, 809. (17) Chen, D. H. C.; Liu, A. P. C.; Koga, Y. Fluid Phase Equilib. 2001, 189, 31. (18) Koga, Y. J. Therm. Anal. Calorim. 2002, 69, 705. (19) Koga, Y.; Westh, P.; Nishikawa, K. J. Phys. Chem. A 2004, 108, 1635. (20) Koga, Y.; Westh, P.; Nishikawa, K. J. Phys. Chem. A 2004, 108, 8533.
Hydrophobicity vs Hydrophilicity (21) Miki, K.; Westh, P.; Nishikawa, K.; Koga, Y. J. Phys. Chem. B 2005, 109, 9014. (22) Giordano, R.; Magazu, M. S.; Maisano, G.; Majorico, P.; Migliadfo, P.; Vasi, C.; Wanderlingh, U. Physica B 1995, 213/214, 515. (23) Koga, Y. In ComprehensiVe Handbook of Calorimetry and Thermal Analysis; Sorai, M., et al., Eds.; John Wiley & Sons: New York, 2004; Sec. 3.1.3, pp 195-199. (24) Koga, Y. Can. J. Chem. 1988, 66, 1187. (25) Idrissi, A.; Sokolic, F.; Perera, A. J. Chem. Phys. 2000, 112, 9479. (26) Janelli, M. P.; Magazu, S.; Maisano, G.; Majolino, D.; Migliado, P. Phys. Scr. 1994, 50, 214. (27) Kyte, J. Biophys. Chem. 2003, 100, 193.
J. Phys. Chem. B, Vol. 109, No. 41, 2005 19541 (28) Tanford, C. The Hydrophobic Effect: Formation of Micelles and Biological Membranes; John Wiley & Sons: New York, 1973. (29) Marcus, Y.; Kamlet, M. J.; Taft, R. W. J. Phys. Chem. 1988, 92, 3613. (30) Tayler, E. T.; Tsai, R.-S.; Testa, B.; Carrupt, P.-A.; Leo, A. J. Pharm. Sci. 1991, 80, 590. (31) Somero, G. N., Osmond, C. B., Bolis, C. L., Eds.; Water and Life; Springer-Verlag, Berlin, 1992. (32) Collins, K. D.; Washabaugh, M. W. Q. ReV. Biophys. 1984, 18, 323. (33) Cacace, M. G.; Landau, E. M.; Ramsden, J. J. Q. ReV. Biophys. 1997, 30, 241.
