To the editors:

Water is the chemical most familiar to us. Oxygen surrounds us, but it cannot be seen or touched. Water is visible, in its three states. It can be tasted, held, and appreciated for its freshness on a hot day. It is used to wash, cook, heat, and cool. Water is so fundamental for humankind that almost all cultures and religions have endowed it with at least some form of special significance: baptismal and purification water, holy water, the Fountain of Youth, the Great Flood, and the great rivers. Zarathustra named water one of the four basic elements, and Thales considered it the primary principle of life. No other material has ever received such veneration.¹

Water’s uniqueness is apparent in its pressure-temperature phase diagram, which, particularly in the low temperature region, is not only complicated, but also contains several ice phases. Most boundaries between the ice phases, and particularly those in the liquid phase, are parallel to the temperature axis, reflecting density-driven phase changes. There are also entropy-driven transitions, with phase boundaries parallel to the pressure axis. Marc Henry’s essay on the properties of supercooled water offers an intriguing top-level argument about one of the most fascinating behaviors of H₂O.²

Despite being essential for daily life, water remains mysterious, exhibiting almost 40 anomalous properties. A partial list of its unexpected behaviors includes a high boiling point and surface tension; a high enthalpy of fusion—almost 41 kJ/mol, compared to 19 kJ/mol for H₂S—which significantly lowers the cryoscopic constant; a maximum density at 3.98°C; a high melting point due to its low entropy in the liquid state; a lower melting point as pressure increases, and hence a negative value in the slope of the ∂P/∂T curve in the phase diagram, until about –22°C at 210 MPa; 9% volume contraction during melting, similar to other tetrahedral solids such as Si or SiO₂; high viscosity and Arrhenius activation energy of the viscous flow; anomalous changes in its viscosity upon cooling, in high temperatures, and at high pressure; a low value of isothermal compressibility,

$\kappa =-\frac{1}{V}{\left(\frac{\partial V}{\partial P}\right)}_T$,

and of isobaric expansion,

$\alpha =-\frac{1}{V}{\left(\frac{\partial V}{\partial P}\right)}_P$;

high isobaric heat capacity in the liquid state, which drops 50% in the solid and vapor states; more neighbors in the liquid state with increasing temperature; a minimum in the solubility of gases and of scarcely soluble materials as temperature increases; and a large dielectric constant.

Some of these anomalies are responsible for remarkable natural phenomena: icebergs float, energy from the sun is redistributed across the oceans and between the oceans and the atmosphere, capillary forces allow trees to lift enormous amounts of water to their leaves, climates in humid areas remain moderate, and natural convection occurs in water basins.

In October 1611, Galileo Galilei engaged Lodovico delle Colombe in a public debate about the reason ice floats.³ More than 400 years later, a topic in this debate—the structure of liquid water—is still vigorously discussed. Investigations into short- versus long-range water structure have spawned a multitude of published studies. The existence of high-density versus low-density water has been invoked.⁴ Hydrogen-bond clusters in expanded or collapsed states seem to quickly change from one state to the other. The changes appear to depend on temperature, pressure, and the presence of solutes, proteins, and polysaccharides.⁵ If this is true, several anomalous features of water would be explained.

Two other phenomena are challenging current scientific certainties in relation to water, and other liquids too: Hofmeister effects and dissolved gases.

Hofmeister Phenomena

Hofmeister effects are so common in chemistry, one cannot help but wonder why they are not addressed in fundamental university chemistry courses or the subject of a chapter in the main general and inorganic chemistry textbooks.

Around 1888, Franz Hofmeister and his collaborators in Prague performed a simple experiment precipitating egg-white albumin in the presence of different salts, at the same concentration. They observed that solutions of some salts precipitated albumin more effectively than others. A year later in Saint Petersburg, Ivan Mikhaylovich Sechenov realized that scarcely soluble compounds such as gases possess different solubilities in salt solutions, depending on the nature of the electrolyte. He was able to quantify the observation and introduce the Sechenov constant, K_s:

$\mathrm{l}\mathrm{n}\frac{S_{salt}}{S_{water}}=-K_S{c}_S$,

in which s_salt, s_water, and c_s denote the solubility in the salt solution, the solubility in pure water, and the concentration of the salt, respectively. K_s depends on the specific nature of the ion pair in the salt.

After Hofmeister’s pioneering experiments and his first six papers on the topic, the phenomenon that took his name developed rapidly.⁶ Research confirmed that specific ion effects occur in bulk solution at any interface in chemistry, in biology, and elsewhere.⁷ The effects appear not only in water, but also in polar organic solvents.⁸ One of the first studies devoted to this issue, long before Hofmeister’s observations, was undertaken in 1847 by Jean Léonard Marie Poiseuille and concerned the viscosity of salt solutions as a function of their concentration and composition.⁹

Specific ion, or specific salt, effects depend on the solvent composition, the properties of the interfaces—polarity and hydrophobicity¹⁰—the nature of the dissolved particles, and their concentration. Hofmeister phenomena usually emerge at moderate concentrations of salts, although there are some exceptions. At extremely low concentrations, below 0.1 mM, all salts behave similarly. In this dilute regime, electrostatics dominate. The system is so dilute that only charge and interionic distance matter. Under these conditions, electrostatic models, namely the Debye–Hückel theory, predict the average ionic activity coefficient γ_± with great accuracy:

$\mathrm{L}\mathrm{o}\mathrm{g} {\gamma }_{\pm }=-A\sqrt{I}$,

where I is the ionic strength of the solution, and A depends on temperature and on the dielectric constant and density of the solvent.

When the ionic strength increases above 10 mM, the behavior of a salt is determined by its character. Here the Debye–Hückel equation needs to be modified. The following expression reflects ion specificity and the linear contribution of I:¹¹

$\mathrm{L}\mathrm{o}\mathrm{g} {\gamma }_{\pm }=-\frac{A\sqrt{I}}{1+Ba\sqrt{I}}+bI$.

The ion-specific parameter is b. For I → 0, the equation reverts to the previous formula.

Some salts, for example MgSO₄, make water more viscous, while others, such as CsNO₃, make it less viscous at the same temperature. Grinnell Jones and Malcolm Dole quantified the effect in a series of equations.¹² The most well-known is

$\frac{{\eta }_{solution}}{{\eta }_{water}}=1+A\sqrt{c}+B_{JD}c$,

where B_JD is the Jones–Dole coefficient, positive for salts such as MgSO₄ and negative for solutes such as CsClO₄. The former are called kosmotropes, and the latter are known as chaotropes. η_solution is the viscosity of the salt solution at concentration c; η_water is the viscosity of pure water at the same temperature. A derives from the Debye–Falkenhagen theory. For very dilute solutions, the term in $\sqrt{c}$ prevails. For moderately concentrated solutions, the third term in the equation dominates.

Kosmotropes increase the structure of water, binding water molecules more effectively. They also possess high absolute values in Gibbs free energy of hydration (Δ_hydrG). Salts such as MgSO₄ or CaCl₂ are often used in chemical laboratories as desiccants to strip water from wet chemicals. Chaotropes, in contrast, are hydrophobic. They show low absolute values of Δ_hydrG and move closer to interfaces.

When a solute is added to water, it perturbs the water’s structure. As the solute scrambles the orientation and distribution of water molecules, the cohesive energy that is responsible for water’s high surface tension is at least partly overcome. If the solute—for example, Li⁺, Al³⁺, F^–, or CO₃^2–—interacts strongly with the water, then a neat solvation layer with water molecules of reduced dielectric constant and mobility will form around the kosmotrope. The solution then settles into a more stable state. But if a chaotrope such as Cs⁺ or I^– is added, after disturbing the water’s structure, the ion will not reconstruct any firm solvation shell.¹³

The reasons that different ions produce distinct effects are unclear. Despite many experiments and articles examining these issues, a complete understanding remains elusive. It is not known why the Hofmeister series exists, and the behaviors of some systems and reactions remain unpredictable. In some cases, depending on the polarity and hydrophobicity of the interfaces, the direct Hofmeister effect observed for cations and anions is reversed.¹⁴

A high concentration of salt is usually needed to observe the Hofmeister effect, about 10 mM or more.¹⁵ And at very low concentrations, electrostatics dominate. This means that Coulomb forces, which are nonspecific, cannot explain specific ion effects. Other forces are at work. Dispersion forces depend on ionic polarizability, a parameter reflecting the softness of the electron cloud. In 2001, Barry Ninham et al. proposed that dispersion forces should be taken into account when describing ion–ion and ion–solvent interactions.¹⁶ More recently, investigations have targeted different kinds of ions, such as Ph₄B^– and Ph₄As⁺, and superchaotropes such as polyoxometalates and borate clusters, where multipolar interactions, various hydration sites, and charge delocalization play a major role.¹⁷

Deep at the core of chemistry, Hofmeister phenomena stand apart as one of the most intriguing and intricate topics where specificity emerges. Important issues remain to be investigated, including the different interactions that water molecules establish with ions. In the case of complexes such as [Al(H₂O)₆]³⁺, water acts as a coordinating ligand that uses electron lone pairs located on the oxygen atom to fill the d orbitals of the central metal ion. In hydrated Na⁺ or K⁺, water interacts through ion–dipole forces. Another issue is the electronic configuration of the intervening atoms. For halides, the kosmo-to-chaotropic ranking is F^– > Cl^– > Br^– > I^–. Surprisingly, the series is reversed for halates: IO₃^– > BrO₃^– > ClO₃^–.¹⁸ This may be related to the electron configuration of Cl, Br, and I and the energy of their d orbitals.

Dissolved Gases

To date, fewer studies have been devoted to dissolved gases than Hofmeister effects. The common understanding seems to be that dissolved gases are nothing more than negligible, inert impurities in water. Dissolved oxygen and CO₂ are essential to aquatic life. Dissolved CO₂ participates in the oceanic buffer system. Some chemists have noted negative effects of dissolved oxygen, due to its role in radical reactions, and CO₂, due to its acidic properties. Otherwise, dissolved gases have been disregarded, including in computational models.

Degassed water is known to be a better cleaning agent than water that has not been degassed. Its dispersion with an alkane takes much longer to phase separate than in normal conditions.¹⁹

Bubble–bubble coalescence in an aqueous salt solution has been shown to depend on the nature and concentration of the solute.²⁰ Seawater is somewhat foamy and retards the coalescence of bubbles. In fresh water, as in a waterfall, bubbles almost immediately coalescence and burst. By filling chromatography columns with either pure water or different salt solutions and then bubbling air through the frits, researchers have shown that some salts inhibit bubble coalescence at more or less the same concentration, 0.15 M. Vincent Craig checked the effect of several salts and identified two kinds of ions, which he named α and β. When the ion pair in the salt is of the type αα or ββ, bubble coalescence is inhibited. When it is αβ or βα, the electrolyte has no effect.²¹ Various mechanisms have been proposed to explain the results, but studies are thus far inconclusive.

Ninham and I recently observed that removing gases from salt solutions reproducibly changes electrical conductivity, depending on the nature of the salts resulting from their α and β ions.²² Dissolved gases are associated with other strange phenomena. One example is the formation kinetics of pseudopolyrotaxanes from cyclodextrin and polyethylene-glycol–like linear polymers.²³ The cloud point of aqueous dispersions of dioctanoyl-phosphatidylcholine can change significantly if the water is completely degassed.²⁴

These phenomena are not exclusive to water. Ninham and I recently demonstrated that in a perfluorooctane and n-hexane system, degassing the two liquids remarkably stabilizes the monophasic state, even below the upper critical solution temperature.²⁵

In Closing

The structure of water is elusive. The array and strength of interactions that keep water molecules together with such high cohesive energy determines their anomalous properties and, in particular, their interactions with solutes, whether neutral or ionic.

It is not only life that relies on water. In the natural inorganic world, the most abundant minerals are silicates, carbonates, sulfates, aluminum, magnesium, calcium, sodium, and potassium. In the Earth’s mantle, the most abundant elements are magnesium 23%, silicon 22%, iron 5.8%, calcium 2.3%, aluminum 2.2%, sodium 0.3%, and potassium 0.3%. These ions are all strong or mild kosmotropes with important hydration layers. In contrast, the poorly hydrated chaotropes are not common in the natural inorganic world or in living systems. The large quantities of water found on earth and the origins of life may depend on these features.

Pierandrea Lo Nostro

Marc Henry replies:

In my article, I have stressed that when exposed to electromagnetic excitations in a frequency range far below 1,000 cm^–1, water responds in such a way that indicates it is not merely an ensemble of H₂O molecules interacting through van der Waals forces. Something else is needed to explain water’s static and dynamic properties. This additional force, which has been named hydrogen bonding, is fundamental to water structure. Unlike van der Waals forces, hydrogen bonding does not find an explanation in classical physics or conventional quantum mechanics.²⁶ Pierandrea Lo Nostro has provided a pertinent and concise summary of the consequences of not knowing exactly what hydrogen bonding is; he describes its many unresolved anomalies. As I stress in another article, anomalies may also be viewed as mysteries.²⁷

Lo Nostro and his colleague Barry Ninham are top-level colloid scientists who have long been interested in Hofmeister effects and in the importance of dissolved gases in biology. They were the first scientists to point out and demonstrate that the behavior of non-diluted aqueous solutions cannot be explained with DLVO concepts. As a consequence, electrostatics cannot be the sole explanation for the behavior of colloidal forces in water. In fact, there is good evidence that water should be considered as a polymer, {H₂O}_n, where n is an integer that may range from n = 2 in the case of the water dimer, up to n = 10⁷ in the coherence domain in quantum electrodynamics. The crucial point is that, as evidenced by vibrational spectroscopy and neutron scattering, such polymers are dynamic entities with timescales ranging from 100 femtoseconds to 10 picoseconds; they are not a static clustering of water molecules. Water structure then should not be considered only as a structure in space, such in water clusters, but also as a structure in time—composed of incredibly minute frequencies. In another article, I describe how Hofmeister effects may be interpreted as being in harmony or disharmony with the frequencies coming from water polymers and dissolved species, whether kosmotropes, chaotropes, or gases.²⁸

For highly diluted solutions or degassed solutions, only diffusion leading to bubble coalescence for gases and electrostatics for ionic species must be considered, as water polymers impose the frequency. Above a critical threshold, specific effects are to be expected where the overall harmony is influenced by the frequencies of both water polymers and solute species. I believe that the anomalies pointed out by Lo Nostro will soon be explained, if researchers move from thinking in terms of space-oriented static colloidal chemistry to thinking in terms of a frequency- or time-oriented dynamic colloidal symphony. The challenge for a new generation of chemists and physicists will be to root biology more in electromagnetism and quantum physics than in chemistry. Much more remains to be learned, but this paradigm provides at least a new direction for experiments in liquid water structure, Hofmeister phenomena, and the role of dissolved gases in hydrophobic interaction.

To conclude, I will mention two books related to the issues Lo Nostro discusses in his letter. The Fourth Phase of Water, by Gerald Pollack, is about exclusion zone water. Pollack explains how water can be used to build simple devices allowing the transduction of infrared light, coming from the sun or the earth, into movement.²⁹ Infrared energy is everywhere and is constantly renewed by sunlight, and water is one of the most abundant substances on the earth’s crust. The dream of having a quasi-perpetual and inexpensive machine based on water is becoming a reality.

The second book, Sur les traces de René Quinton, by Jean-François Dray et al., relates to the importance of seawater. This wonderfully illustrated book is devoted to the life and works of René Quinton, the so-called French Darwin at the dawn of the twentieth century.³⁰ The authors describe how seawater was used in medicine as a substitute for blood and for healing a wide range of diseases. Spanish and English versions of the book are in preparation. Finally, for readers interested in learning more about the many fascinating aspects of water, a special issue will be soon published by the journal Substantia under the initiative of Lo Nostro as editor-in-chief, with the help of the University of Florence.³¹