Hofmeister effects in membrane biology: The role of ionic dispersion potentials M. Bostro ¨ m* Department of Physics and Measurement Technology, Linko ¨ping University, SE-581 83 Linko ¨ping, Sweden D. R. M. Williams, P. R. Stewart, and B. W. Ninham ² Department of Applied Mathematics, Research School of Physical Sciences and Engineering, Institute of Advanced Studies, Canberra, ACT 0200, Australia ~Received 16 June 2003; published 3 October 2003! Membrane biology is notorious for its remarkable, and often strong dependence on the supposedly irrelevant choice of ion pair of background salt solution. While experimentally well known, there has been no progress towards any real theoretical understanding until very recently. We have demonstrated that an important source behind these Hofmeister effects is the ionic excess polarizabilities of ions in solution. Near an interface an ion experiences not only an electrostatic potential, but also a highly specific ionic dispersion potential. At biologi- cal concentrations ~around 0.1M and higher! when the electrostatic contribution is highly screened this ionic dispersion potential has a dominating influence. We present the result of model calculations for the interfacial tension and surface potential that demonstrates that inclusion of ionic dispersion potentials is an essential step towards predictive theories. Our results are compared with experimental surface and zeta potential measure- ments on phospholipid bilayers, zirconia, and cationic micelles. DOI: 10.1103/PhysRevE.68.041902 PACS number~s!: 87.16.Dg, 73.30.1y I. INTRODUCTION Hofmeister, or specific ion, effects are as ubiquitous in biology and colloid science as they are ignored @1,2#. Ex- amples abound, surface tension of electrolytes @3#, interfacial work of adhesion at electrolyte-oil interfaces @4,5#; force measurements @6,7#, zeta and surface potentials @8–14#, p H measurements @15–17#, ryanodine binding to calcium release channels @18#; cutting-efficiency and stability of DNA @19,20#; and formation of silicates @21#. We will demonstrate that a model originally proposed by Ninham and Yaminsky @1# offers an explanation for the ion specific surface and zeta potential of membranes, cationic micelles, and zirconia. We will also discuss how this model can accommodate the ex- perimentally observed ion and alcohol specific oil-water in- terfacial tension, as well as leading to different perspectives on the origin of membrane folding. The standard Gouy-Chapman mean-field theory @22–24#, commonly and often successfully used in theoretical model- ing in membrane biology and colloid science, rely on elec- trostatics, which in turn rely on the ionic charge. According to this theory all salt solutions with the same valency should be equivalent. Deviations from this theory ~which occur commonly and are often very large! have been attributed to binding of unknown origin @8#, sometimes associated with ion specific water structure effects ~due to ions supposedly being either ‘‘structure breaking’’ or ‘‘structure creating’’ @25#!. An important step towards a solution of the problem and predictive theories is to realize that the original double- layer theory of charged interfaces in salt solutions is thermo- dynamically inconsistent @1,26#. For consistency ionic dis- persion potentials acting between ions and interface must be included in the theory. When this is done ion specific results emerge naturally @1,27,28#. An important question that re- mains is to what extent the inclusion of ionic dispersion po- tentials is sufficient to explain specific Hofmeister effects. Other ion specific properties like water structure @29#, ion size, dielectric constant variation near the interface, solva- tion, electronegativity, counterion and co-ion exclusion @30,31# may clearly also be important, as may be the role of dissolved gas @32#. Ionic dispersion potentials have an im- portant role in the ion specificity of surface tension @27,33#, double-layer forces @28#, ion binding to micelles @34#, poly- electrolytes @35#, and p H measurements with glass electrodes @17#. The ion specific double-layer theory is rehearsed in Sec. II. Ion specific oil-water interfacial tensions and other ion specific alcohol effects are considered in Sec. III. We discuss why the ionic dispersion potential near an oil-water interface, similar to the chemical potential of different n-alkanes, de- pend both for sign and magnitude on the chain length of different hydrocarbons. We show that large attractive ionic dispersion potentials acting on anions at biological concen- trations result in negative interfacial tension changes. We briefly consider how these ion specific alcohol effects influ- ence the self-assembly of silicates. We demonstrate in Sec. IV how a few experimentally measured ion specific surface and z potentials of cationic micelles, zirconia, and phospho- lipid membranes can be understood once ionic dispersion potentials are included in the theory. In Sec. V we summarize our results and discuss a mechanism in which the folding of membranes and proteins is a natural consequence of alcohol and Hofmeister effects. II. ION SPECIFIC DOUBLE-LAYER THEORY The model system that we consider is an aqueous solution of negatively charged monovalent anions and positively *Electronic address: mabos@ifm.liu.se ² Present address: Department of Chemistry and CSGI, University of Florence, 50019 Sesto Fiorentino, Italy. PHYSICAL REVIEW E 68, 041902 ~2003! 1063-651X/2003/68~4!/041902~6!/$20.00 ©2003 The American Physical Society 68 041902-1