J. Phys. Chem. C 2007, 111, 3753-3755
3753
Specific Cation Effects at the Hydroxide-Charged Air/Water Interface Patrice Creux,*,† Jean Lachaise,† Alain Graciaa,† and James K. Beattie‡ Laboratoire des Fluides Complexes, UMR 5150 CNRS-TOTAL-UPPA, BP 1155, 64013 Pau cedex, France, and School of Chemistry, UniVersity of Sydney, NSW 2006, Australia ReceiVed: January 4, 2007
The zeta potentials of individual gas bubbles in water have been measured with a spinning tube zetameter as a function of added electrolyte concentrations. The bubbles have a negative charge at pH >3 due to the preferential adsorption of hydroxide ions. Addition of alkali halide electrolytes from 10-5 to 10-2 M reduces the zeta potential due to double layer compression. The effect is independent of the anion among the sodium chloride, bromide and iodide salts, indicating that these do not compete with hydroxide ion at the surface. Lithium chloride is more effective than cesium chloride at reducing the zeta potential, indicating that electrostatic effects are more important than dispersion forces at these low ionic strengths.
Introduction
Experimental Section
Since the 19th century, it has been known from electrophoretic observations that air bubbles in neutral surfactant-free water acquire a spontaneous negative charge. In 1914 McTaggart extended the 1861 observations of Quincke with a cylindrical rotating cell to counteract the bubble buoyancy and speculated that the negative charge was due to the selective adsorption of hydroxide ion.1 Since then, measurements of the pH dependence of the zeta potential of gas bubbles have confirmed this view. Graciaa et al. eliminated the artifact of electroosmotic flow in the rotating cell by coating the inside surface with a polymer and made the first reliable measurement of the zeta potential of a bubble in neutral water at pH 7 of about -65 mV.2 This value has recently been confirmed with a rising bubble method.3 The pH dependence of the zeta potential gives an isoelectric point around pH 3-4 and reaches a maximum negative value of about -110 mV above pH 9.3,4 The pristine oil/water interface behaves similarly.5 For hexadecane-in-water emulsions, the total surface charge of the adsorbed hydroxide ions is largely neutralized by alkali metal counterion condensation.6 Surprisingly, however, there is little counterion specificity; addition of lithium, sodium, or cesium chlorides each reduce the zeta potentials by double layer compression by about the same extent up to 10 mM salt concentrations.7 To further investigate this question of the “spontaneous surface charge” at the air-water interface, we have examined the effects on the zeta potentials of single bubbles of different alkali metal cations at concentrations of 0.1 µM to 10-2 M and at different pH’s. The zeta potential measurements were made with the spinning cell zetameter.2,4,8 The rotation of the cell counteracts the effect of buoyancy on the gas bubble, while a thick polymer (DEAE-Dextran) lining on the cell wall eliminates electroosmotic counter flow. The translation of a bubble induced by an electric field applied along the rotation axis is monitored with an optical device to measure the electrophoretic speed.
Materials. Water was purified by a Millipore milli-Q 185 E system to give a resistivity of 18.2 MΩ cm. Contamination by CO2 was avoided by bubbling nitrogen or helium and keeping the sample out of contact with atmospheric air. The measurement cell was coated as described previously,2 then cleaned by soaking with a 10% solution of nitric acid for 48 h, and then abundantly rinsed with pure water. Electrolytes were from Aldrich with purity grades better than 99% for hydrochloric acid, 99.5% for sodium chloride and 99.9% for cesium chloride, lithium chloride, sodium iodide, sodium bromide, sodium hydroxide, lithium hydroxide, and cesium hydroxide. The gases were AGA HPLC grade, with nitrogen used in all cases except where described otherwise. Methods. The zeta potentials were calculated from the measured electrophoretic mobility with the expression of Sherwood (1), as described previously:2
* Corresponding author. E-mail:
[email protected]. † Laboratoire des Fluides Complexes. ‡ University of Sydney.
µa ζ)
xΩν U
3.6E
(1)
where Ω is the spinning cell speed, ν the kinematic viscosity, µ the dynamic viscosity, the dielectric constant of the aqueous solution, E the electrical field applied between the two electrodes, U the electrophoretic mobility, and a the bubble radius. The zeta potentials correspond to the average of at least ten measurements obtained by changing the polarity of the electrodes; their accuracy is approximately (3 mV. Measurements made after 10, 20, 30. and 60 s were the same within this experimental error, indicating no dynamic effects affected the measurements. Results Different Gases in Neutral Water. The zeta potentials of bubbles of four different nonreactive gases (nitrogen, oxygen, helium, and industrial air) were measured and found to be the same (-63 ( 2 mV) within experimental error. In the subsequent experiments described below nitrogen gas was used. Different Monovalent Anions. The effect on the zeta potentials of increasing the electrolyte concentration with sodium
10.1021/jp070060s CCC: $37.00 © 2007 American Chemical Society Published on Web 02/13/2007
3754 J. Phys. Chem. C, Vol. 111, No. 9, 2007
Creux et al.
Figure 1. Zeta potential at air/sodium salt solution interfaces (square shape, NaI; disc shape, NaBr; diamond shape, NaCl.). Figure 3. Zeta potential of gas bubbles immersed in aqueous solutions at pH 9 (disc shape, LiCl; diamond shape, NaCl; square shape, CsCl). The pH was adjusted with solutions of the respective alkali metal hydroxide.
Figure 2. Zeta potential at air/alkaline solutions interfaces (diamond shape, NaOH; square shape, LiOH; triangle shape, CsOH).
salts of different anions is shown in Figure 1. As the sodium halide electrolyte concentrations are increased above about 10-5 M the zeta potential decreases (becomes less negative) in a fashion typical of double layer compression. The slope of the curves is almost the same for the chloride, bromide and iodide salts, with the absolute value of the negative zeta potential decreasing from chloride to bromide to iodide. Different Monovalent Cations. In contrast, addition of NaOH increases the negative zeta potential (Figure 2), clearly demonstrating the specific adsorption of hydroxide ion at the interface. The effect of addition of different alkali metal shows a slightly greater negative zeta potential with CsOH than with NaOH, but the effect is almost within the experimental error. In contrast, as the salt concentration is increased above 10-5 M, specific ions effects appear (Figure 3). The zeta potential is reduced on addition of the salts in a fashion typical of double layer compression, by about 15 mV per factor of 10 increases in salt concentrations. Lithium chloride has the greatest effect and cesium chloride the least, although there is not a large difference among the alkali metal cations. The effects are similarly apparent at pH 7 (Figure 4). The initial zeta potential is somewhat lower, consistent with its pH dependence.3,4 Again Li+ is most effective and Cs+ least, with a somewhat more marked difference among these cations. At pH 5 the initial zeta potential is considerably lower at -40 mV. The order of the effect of the cations remains the same, although the slopes of the plots are lower and there is some hint of saturation of the effect as the zeta potential drops below -10 mV (Figure 5).
Figure 4. Zeta potential at air/chloride salts solutions interfaces at initial pH 7 (disc shape, LiCl; diamond shape, NaCl; square shape, CsCl).
Figure 5. Zeta potential of gas bubbles immersed in aqueous solutions at pH 5 (disc shape, LiCl; diamond shape, NaCl; square shape, CsCl). The pH was adjusted with HCl.
Discussion The observation of a negative charge on gas bubbles in aqueous solutions is consistent with and confirms many previous reports. In contrast, a recent comprehensive review by Jungwirth
Hydroxide-charged Air/Water Interface and Tobias describes theoretical simulations that indicate a much stronger preference of the hydrated proton than the hydroxide ion for the surface.9 This is contrary to the experimental evidence of a negative zeta potential even in weakly acidic solutions. An isoelectric point of pH 3-4 indicates a preference of hydroxide over protons of 104-103, an observation made by Exerova many years ago.10 In addition, the surface preference for Li+ over Cs+ as the countercation indicates that electrostatic effects are more important than dispersion effects. Again, this is contrary to some inferences from theoretical models,11 which indicate a preference for polarizable anions at the surface, but consistent with the repeated views of Bostro¨m and Ninham who observe that polarizability effects only dominate at high electrolyte concentrations where electrostatic forces are screened.12 This is corroborated by the present results (Figure 1) which show no significant difference among chloride, bromide and iodide in affecting the surface charge. Even at 10 mM concentrations these anions do not compete with hydroxide ion at the surface in pH 7 water. So far, however, theoretical studies have ignored this intrinsic charge at the interface and hence have not grasped the essential physics of the problem. Models of the air/water interface without the presence of the hydroxide ions ignore the strong effects of the hydration of the anion on the hydrogen bond structure of the surface water. We do not know what the surface charge density is at the air/water interface, but from the similarity between the pH dependence of the zeta potentials of the air/ water3 and the oil/water6 interfaces, and of their salt dependence,5 it is reasonable to assume that it is of the same magnitude as that measured for the oil/ water interface, about 0.05 C m-2 (5 µ cm-2). This implies a hydroxide ion every 3 nm2, that is separated by less than 2 nm, or 5-7 water molecules. Hence the presence of the hydroxide ions at the interface will have a profound effect on the water structure there. Hydroxide and hydronium ions are unique in aqueous solutions as the autolysis products of the solvent water. Geissler et al. have described a molecular dynamics simulation of the autoionisation process.13 In their simulations dissociation begins with a fluctuation in the solvent electric field that subsequently leads to the charge-separated state. At the hydrophobic interface, however, the electric field may be generated by the orientation of the water dipole moments; it has been estimated to be as large as 0.5 V over a few molecular diameters, leading to a
J. Phys. Chem. C, Vol. 111, No. 9, 2007 3755 field gradient of the order of 109 V m-1.14,15 This could lead to enhanced autolysis and the preferential adsorption of hydroxide ion with repulsion of protons from the surface. As mentioned above, oil droplets in water have very similar surface properties to air bubbles, with a similar pH dependence of their zeta potentials and a similar effect of electrolyte concentration on the calculated double layer charge. One of us recently reported an absence of specific cation effects on hexadecane emulsion droplets.7 The difference from the present results could arise from the different concentration regimes used. In the emulsion study a 2 vol % oil concentration was employed, resulting in a large interfacial surface area. For example, for such an emulsion with 0.6 micron droplets and a surface charge density of 5 µC cm-2, the counterion concentration in the double layer would be 0.1 mM, a significant fraction of the total electrolyte concentration. In contrast, the isolated bubble adsorbs a negligible proportion of the counterions. It would be of interest in subsequent work to repeat some of the present measurements with isolated oil drops,8 to determine if there are observable differences between the air/water and oil/water interfaces. Acknowledgment. This work was supported by the Australian Research Council References and Notes (1) McTaggart, H. A. Philosophical Magazine 1914, 27, 297. (2) Graciaa, A.; Morel, G.; Saulner, P.; Lachaise, J.; Schechter, R. S. J. Colloid Interface Sci. 1995, 172, 131. (3) Takahashi, M. J. Phys. Chem. B 2005, 109, 21858. (4) Graciaa, A.; Creux, P.; Lachaise, J. Surf. Sci. Ser. 2002, 106, 825. (5) Beattie, J. K., The Intrinsic Charge at the Hydrophobe/Water Interface. In Colloid Stability - The Role of Surface Forces - Part II; Tadros, T., Ed. Wiley-VCH: Weinheim, 2006; Vol. 2, p 153. (6) Beattie, J. K.; Djerdjev, A. M. Angew. Chem., Int. Ed. 2004, 43, 3568. (7) Franks, G. V.; Djerdjev, A. M.; Beattie, J. K. Langmuir. 2005, 21, 8670. (8) Graciaa, A.; Creux, P.; Dicharry, C.; Lachaise, J. J. Disper. Sci. Technol. 2002, 23, 301. (9) Jungwirth, P.; Tobias, D. J. Chem. ReV. 2006, 106, 1259. (10) Exerowa, D. Kolloid Z. Z. Polym. 1969, 232, 703. (11) Dang, L. X. J. Phys. Chem. B 2002, 106, 10388. (12) Bostro¨m, M.; Williams, D. E.; Ninham, B. W. Phys. ReV. Lett. 2001, 87, 168103. (13) Geissler, P. L.; Dellago, C.; Chandler, D.; Hutter, J.; Parrinello, M. Science 2001, 291, 2121. (14) Dang, L. X.; Chang, T.-M. J. Phys. Chem. B 2002, 106, 235. (15) Mamatkulov, S. I.; Khabibullaev, P. K.; Netz, R. R. Langmuir 2004, 20, 4756.