Water site headerMasthead Island, Great Barrier Reef Print-me keygo to Water Visitor Book contributions
Go to my page Water Structure and Science


Surface generating osmotic

pressure, repelling solutes


Surface generating osmotic pressure repelling solutes

Self-generation of colligative properties at interfaces a

Surfaces can generate osmotic pressure without the need for solutes.


V Osmotic pressure of particles and membranes

V Exclusion zone (EZ) water
V Proposal for the generation of osmotic pressure at aqueous interfaces

V Overview of conventional colligative properties
V Conventional osmotic pressure
V Nanobubbles

V Osmotic pressure stabilizes nanobubbles

V The Nernst and Knudsen layers

V Nafion®

Osmotic pressure of particles and membranes

Although there is some contention amongst physical chemists over how to describe osmotic pressure, it is generally agreed that it is a reversible thermodynamic colligative property and correlates well with other colligative properties such as freezing point depression. As such, the osmotic pressure is thought ideally to depend on the number of dissolved ‘particles’ (e.g., molecules) present in a volume of liquid. It is, perhaps, unsurprising that ion exchange surfaces can generate very high osmotic pressures of over 100 MPa in water [1669], as they create high surface concentrations of counter-ions (100 MPa is the equivalent of 10 km head of water). However, the contained electrolytes produce stronger osmotic pressure than if they were free (uncontained). Thus, nano-particles surfaced with immobilized polyelectrolytes can produce two hundred times osmotic pressure than the same quantities of free electrolytes [2404], clearly establishing that osmotic pressure can be generated under non-colligative conditions. Poly-ionic nanoparticles, with high surface area, produce such a great osmotic pressure (far greater than expected as due to counterions) f that they can be used in practical desalination processes [1768a,d,f,h]. Likewise; polyelectrolytes may also produce high osmotic pressures between molecules (~100 MPa) [2006]. However, hydrophilic uncharged molecules, without any counterions, also produce high osmotic pressures similar to polyelectrolytes between molecules [2006]. Also, it has been experimentally verified that uncharged hydrophilic particle surfaces generate high osmotic pressure, without the presence of counter-ions or free solutes [1768b,c,g,h,i]. Thus, neutral uncharged triethylene glycolated particles [2404] and carbon particles coated in neutral N-isopropylacrylamide [1768g] both generate considerable osmotic pressure if less than similarly coated poly-acrylic acid particles. Anomalously high osmotic pressures may occur within low-permeability shale [2795]. It is clear, therefore, that relatively small numbers of particles, far less than required by the ‘conventional’ colligative law, can generate high osmotic pressures so long as they possess extensive hydrophilic surfaces [1772].


The lowered water activity within particles and membranes needs some explanation. The water within behaves as though it is under hydrophilic confinement and will be held more strongly due to capillarity. Both these effects increase the grip that the particles and membranes have on their water, so lowering their activity. Additionally, where polyelectrolytes are concerned, the high concentration of fixed binding sites for the counterions prevent their easy loss from the particles or membranes with any counter-ion loss causing an energy consuming large charge separation between the enclosed solution and the bulk.


Where the osmotic pressure is generated within a confined volume, such that insufficient water flow away from the surface may occur, then the question arises as to what happens next. Clearly, tension arises in the neighboring water or else there will be an interface of neighboring water molecules with very different chemical potentials. Maybe this will be slightly relieved by the release of some ions, if present, but with an energetically unfavorable charge separation. Alternatively, the low potential water molecules must transfer their reduced entropy to their neighbors away from the surface until it can be balanced by the averaged energy of 'bulk' water. This will cause flow out of a capillary tube [2669].

Exclusion zone (EZ) water

Microphotograph, from [1740]

based on microphotograph from [1740]
Much work has been done in the laboratory of Gerald Pollack (and confirmed by many other independent workers [2058]) concerning the mesoscopic properties of aqueous solutions next to surfaces [1328, 1740, 2274, 2387]. c In essence, it has been found that the interfacial water next to ionic charged (e.g. polyacrylate see right from [1740], Nafion e, charged Pt [2792]) or neutral uncharged (e.g. polyvinyl alcohol, cellulose acetate, cellulose [2702], quartz [2715], contact lens, rabbit cornea [2790]) hydrophilic surfaces expel solutes to the bulk of the solution that may be several hundred microns away. These exclusion zones (named as EZ-water) can be visualized when low relative molecular mass (molecular weight) dyes, proteins, micron-sized beads, salts [2387] or other solutes are used. With a laser tweezers system, the existence of force fields inside the solute-free EZ have been found to reduce as a function of distance from the surface [1784]. Microspheres introduced into the EZ move away from the interface to the outer extent of the EZ, so showing the force remains even in the exclusion zone and no new phase had formed there [2169] (which might have been expected to prevent such movement). The particles also develop their own EZ [2704]. The movement of the EZ boundary follows a power law close to that expected for free diffusion [2169]. Also the EZ-water seems to possess other physical properties such as absorption at 270 nm, greater density, greater viscosity and negative charge compared with the bulk water. Further, the efficient operation of uncharged hydrophilic reverse osmosis fibrous membranes (that function in spite of not possessing useful molecular-sized pores) show that they must present thick multi-molecular layers of water that are relatively impenetrable by low molecular mass salts and molecules. It has also been independently experimentally verified that uncharged but highly hydrophilic nanoparticles, with high surface area, produce such a great osmotic pressure that they can be used in practical desalination processes [1768a,b]. Also, uncharged hydrophilic polymers have been shown to generate osmotic pressures of greater than MPa at nanoscopic distances [2006]. The 1-3 mm thick interfacial zone of water found at quartz surfaces in 60 ˣ 20 µm2 quartz cavities has its major infrared peak at 3200 cm-1 indicative of more strongly linked hydrogen bonding by about +1 kJ ˣ mol-1 throughout this zone of vicinal water [2715].


Contrasting to these experimental studies, theoretical studies using water molecular models show no extension of the interface beyond the ultra-thin external surface [2198]. However, such studies do not explain the experimental results. It is interesting and relevant to note that, where solid surfaces interact strongly with contacting liquids (including water but not exclusively so), there are non-negligible interfacial effects. These can be revealed by shear flow infrared experiments to extend up to several millimeters into the bulk [2913]. Liquid flowing slowly over the wetting surfaces produces cooling due to stretching, while at fast rates heating is produced by frictional interactions.


There is no generally-accepted explanation for these experimental phenomena or properties. However, below I give a new explanation that is very simple, easily understood and potentially very important in a number of related fields.


It is interesting to hypothesize the effect of removing ions (including H3O+ and OH-) from water within the exclusion zone due to the interfacial osmotic pressure. Ions would terminate hydrogen bonding 'wires' within the aqueous milieu. The absence of such terminations allows more extensive hydrogen bonding and consequent structuring of the low-density form of water. This does not necessarily lead to a higher fraction of the low-density form, just that such clusters will be more extensive. As such, the water will have somewhat different properties, so possibly explaining the experimentally found properties of EZ water, such as changed UV absorption [1328a], refractive index [2793], diffusion [2792c], spin-lattice relaxation times (T1 ) [2792c] and permittivity [2792b].

Proposal for the generation of osmotic pressure at aqueous interfaces

Wherever water is present in solution it may be considered as being either 'bound' or 'free' (bulk), although there will be a transitional water between these states. When considering the colligative properties, 'water' is considered bound to any solute when it has a very low entropy compared with pure liquid water. Such water is considered part of the solute and not part of the dissolving 'free' (bulk) water [1064]. As pure liquid water consists of a mixture containing low-density water, made up of extensively hydrogen-bonded structures, and higher density water, consisting of much smaller less extensive clusters, the proportions of 'bound' or 'free' water in pure liquid water can vary; the strongly-bound larger clusters behaving like 'bound' water. In bulk liquid water, the relative concentrations of the two aqueous forms are of no consequence as all the water behaves the same throughout. If volumes of the solution contain different proportions of strongly and weakly hydrogen-bonded water molecules (or even more simply that there is more extensive clustering present), then these different volumes will show a difference with respect to their water activity and chemical potential.


Affect of surface on water activity


Affect of surface on water activity

Normally, any such instantaneous differences in water activity, and chemical potential between different volumes within the same mass of liquid, would rapidly cause liquid movement from one to the other in order to equalize these states and so remove the potential differences. Where surfaces interact with the liquid water, the concentration of the more extensive hydrogen-bonded clusters differ from the bulk values with the surface interactions preventing the potential equalization between bulk and surface volumes [3216]. When this occurs, the surface water has a lower water activity and lower chemical potential than the bulk, leading to differences in osmotic pressure and other colligative properties such as freezing point and 'opening' pressures. In contrast, volumes of water within hydrophobic clefts and pores (see left) have higher activity than bulk water whereby the water has a tendency to leave creating 'closing' pressures. This tendency has been shown in both hydrophobic [3130] and hydrophilic [3135] clefts.


Self-generation of osmotic pressure at interfaces


Self-generation of osmotic pressure at interfaces

The change in the chemical potential (μw) is –{RTLn(xws) -RTLn(xwb)} (that is, a negative energy term is added to the chemical potential when xws < xwb) where xws is the mole fraction (activity) of the ‘free’ water (0 < xws < 1) in the surface layer and xwb is the mole fraction (activity) of the ‘free’ (bulk) water (0 < xwb < 1) in the bulk liquid.


Small reductions in water activity give rise to large osmotic pressures

Small reductions in the interfacial water activity give rise to large osmotic pressures


The generated osmotic pressure (Π) is given by,

Π = -(RT/VM).Ln(xws/xwb)

Π = -(RT/VM).Ln(p/p0)

where VM is the molar volume of water, and p/p0 is the ratio of the partial pressures of solution (p)and pure water (p0) (relative humidity). Clearly, quite small reductions in the interfacial water activity give rise to large osmotic pressures (see right).


At hydrophilic surfaces, interactions between the surface and neighboring water molecules fix the localized hydrogen bonding and this, together with steric factors, increases the cluster extent and lifetime [2059].b As the ‘free’ water reduces as compared with its bulk value when the formation of longer-lived and more extensive hydrogen-bonded clusters increases, so the water activity reduces and the osmotic pressure increases. Solutes next to the surface will move to equalize the water activity throughout the liquid. In other words, the osmotic pressure generated next to the surface will displace solutes from the surface towards the bulk until its effect is equaled by the osmotic pressure of the surrounding solution or the system reaches a steady state.


A typical unstirred layer


a typical unstirred layer, Nernst layer depth delta


As the first effect of this solute expulsion is naturally the formation of an increased concentration band as expelled solute mixes with the prior solute concentration, the extent of the expulsion will affect the whole of the unstirred layer. The effect of the surface will decay exponentially out to the stirred bulk (the Nernst layer, δ ~ 1-100 µm or more if unstirred pure water, see diagram right) with a thickness dependent on the cube root of the diffusivity of the solutes [2699]. g The unstirred layer is not stationary but is a region of slow laminar flow parallel to the surface where transport is by diffusion. The thickness of the layer, δ, is less than the actual thickness but rather an operational thickness (see right) defined by the equivalent uniform concentration gradient. As the diffusion of hydrogen ions is greater than hydroxide ions, the unstirred layer will be negatively charged. Where hydrophilic microparticles or nanoparticles are suspended in aqueous solutions, their surfaces will necessarily cause mutually repulsive osmotic pressure effects that may result in the ordering of the particles within small volumes of the liquid [272a, 2060]


It should be noted that osmotic drive does not require a membrane to separate the two solutions [1744] provided there are two phases (e.g. [1669]). Here the two phases consist of the unstirred and stirred layers. In this context [1739], the affected aqueous layer behaves similarly to that described for exclusion zone (EZ) water by Pollack and forms a simple explanation of his experimental data [1740]. It also shows similarities with the experiments on autothixotropy [509, 1898, 1975].


Other experimental properties of the EZ layers, such as the reduction in its thickness with the size of the particle causing the effect [2062] and with increased bulk ionic strength [1328b], are easily explained by the presented hypothesis, as the unstirred layer around particles is known to depend on the particle size and high bulk osmotic pressures will oppose the osmotic pressure at the hydrophilic surface. An interesting alternative explanation of the EZ layers, using Nafion,e involving diffusiophoresis has been proposed [2808] but does not seem to be applicable for all cases of EZ formation.


A standing electromagnetic wave


A standing electromagnetic wave gives rise to  standing hydrogen-bonded water clustering at interfaces

Another effect of interfaces is the formation of evanescent waves due to the internal reflection of electromagnetic radiation. The standing electromagnetic evanescent wave within the interfacial water is caused by impinging electromagnetic radiation and the angle at which total internal reflection occurs (θ). The standing electromagnetic wave produced will interact with water molecules to stabilize a standing wave of hydrogen-bonded clusters that will alter the local concentration and extent of hydrogen-bonded clusters so increasing the above osmotic effect, in agreement with the experimental data [1173, 1589, 1741, 2704]. It has also been shown that infrared radiation increases the hydrogen-bonded interactions of water molecules in protein hydration shells [2489].


It appears that a similar effect on solutes to the one described for water may occur in other polar solvents that can form hydrogen bonds [1742]; thus reinforcing the likelihood that a mechanism is acting that does not depend on the specific properties of water, such as the here-described colligative thermodynamics. Increasing ethanol to water gives a maximum exclusion zone width at about 12 % v/v followed by a decrease to about half the pure water width at 95 % v/v ethanol [3145].


Once the osmotic origin of the EZ water has been accepted, other properties of the EZ water can be rationalized. The increase in density at the interface, as found in EZ water, has been explained previously by the increase in clustering causing the water to behave as though it is at a lower temperature, which also explains the ease with which this surface layer freezes. The presence of 270 nm absorption in the interfacial water, as described for EZ water [1328], has been ascribed to the delocalization of electrons within the extended clustering. d As the cluster lowest unoccupied molecular-orbital (LUMOs) are both huge and relatively low energy, these electron delocalizations are stabilized by the addition of an electron but not by protonation, so causing the charge separation seen at these interfaces [1744, 2061].

Osmotic pressure stabilizes nanobubbles

Typical surface density, as described by molecular models


Typical surface density as described by molecular models. There is often a high density peak just underneath the surface.


There is no a priori reason why the same phenomenon as described above should not also apply to gas/liquid interfaces. The surface of water differs from the bulk as the density decreases to zero in a complex manner, see left for the typical density at the surface as shown by modeling studies. The presence of an extensive EZ layer (up to 4 mm at a flat surface) at the gas liquid interface has been experimentally confirmed [2389].


The generation of the osmotic pressure at the surface is due partly to the necessarily under-coordinated water molecules at the gas-liquid interface forming an ice-like, low-density phase [2004] that has lower water activity than the bulk water, and partly to the higher solute concentration formed just under the interface. Hydrogen bonding in the surface is stronger than in the bulk [1261], due to the reduced competition from neighboring water molecules, lower anti-cooperativity, and compensation for the increased chemical potential on the loss of some bonding.


There is now much evidence that sub-micron-sized gas-filled cavities (often called nanobubbles) can exist for significant periods of time both in bulk aqueous solution and at submerged aqueous-hydrophobe interfaces. Their contained gas is in constant flux with gas molecules both leaving and entering continuously. The cavities are under excess pressure given by the Laplace equation (2γ/r, where γ is the surface tension and r is the cavity radius) as the surface tension causes a tendency to minimize their surface area and, hence, volume. Nanobubbles grow or shrink by diffusion according to external pressure and the degree of over- or under-saturation in the surrounding solution, with the dissolved gas relative to the raised cavity pressure. As the solubility of gas is proportional to the gas pressure, and this pressure is exerted by the surface tension in inverse proportion to the size of the bubbles, there is increasing tendency for gases to dissolve as the bubbles reduce in size. This accelerates the bubble-dissolution process. Such size-reduction is increased by the bubble's movement and contraction during this activity, which aids the removal of any gas-saturated solution from around the cavities.


Calculations show that nanobubbles should only persist for a few microseconds, in contrast to their long lifetimes (hours to days) as detected experimentally by light scattering or resonant mass measurement (bulk nanobubbles [1972]) or tapping mode atomic force microscopy (surface nanobubbles). Interestingly, these cavities (bubbles) are subject to Brownian motion, so behaving as though they have solid shells similar to solid nanoparticles. Surface and bulk-phase nanobubbles can both give rise to the otherwise difficult to explain long range attraction between hydrophobic surfaces due to bridging nanoscopic gas cavities or the osmotic effect due to local nanobubble depletion.


A further reason for at least some of the stability of nanobubbles is that the nanobubble gas/liquid interface is charged [1591], expanding the surface and introducing an opposing force to the surface tension, so slowing or preventing the nanobubble dissipation. It is clear that the presence of like charges at the interface will reduce the effect of the surface tension, with charge repulsion acting in the opposite direction to the surface minimization due to surface tension. Any effect may be increased by the presence of additional charged materials that favor the gas-liquid interface, such as OH- ions at neutral or basic pH [1591]. Recently, it has been proposed that the surface tension is reduced by the degree of supersaturation at the surface [2013]. As this supersaturation is prevented from equilibrating away from the bubbles by the imposed osmotic pressure, this introduces a further stabilization effect on the nanobubbles.


Osmotic pressure may be generated at nanobubble surface


Osmotic pressure generated at nanobubble surface is responsible for nanobubble stability

Contributing to the stability of nanobubbles is the slow rate of gas diffusion to the bulk liquid surface from both surface and bulk-phase nanobubbles [1973, 1987]; in particular, nanobubbles in a cluster of bulk nanobubbles protect each other from diffusion by a shielding effect [2074]. Apparently, there is a thick interfacial layer; a phenomenon experimentally supported by the higher forces required to penetrate greater depths of the nanobubbles' interface [1986, 1987]. It should be noted that the interface has extensive volumes on both sides where there is laminar flow orthogonal to the surface only. Here we assume that the cause of this layer is the osmotic pressure produced by solutes [2406] and water structuring beneath the gas liquid interface [2057], also the cause of the experimentally visible EZ-water layer [2389], so both preventing the gas dissolving towards the bulk and driving any dissolved gas near the interface back into the nanobubble (see above left). The depth of the ‘unstirred’ layer (δ) is approximately proportional to the size of the nanobubbles due to their Brownian motion. This thickness will be reduced by the mixing caused by any surface changes in the bubbles as they rise or change size. A higher osmotic pressure around bubbles has been suggested to explain why bubbles fail to coalesce, although its origin was unproven [2536].


[Back to Top to top of page]


a This theory [2057] was first presented at the international conference 'Water and Nanomedicine', Academy of Sciences and Arts of Republic of Srpska, Banja Luka, Aug. 30, 2011 [1772]. [Back]


b This agrees with the earlier proposition for long-range water ordering involving the partial alignment of water molecules induced by the surface [1328]. [Back]


c These ideas and experiments have been the subject of a critical analysis giving an alternative view involving concentration gradients [2036a] . This, in turn, has been mostly answered by Pollack [2036b]. However, there does not seem to be any good reason to resuscitate the planar hexagonal water structuring hypothesis as proposed in the polywater debacle to explain EZ-water. [Back]


d An alternative view of the 270 nm absorption has been given in terms of an (improbable) highly charged extensive planar hexagonal ice-like model [2077]. [Back]


Partial Nafion® structure



e Nafion®


Nafion® (manufactured by DuPont) has good chemical and electrochemical stability as well as high proton conductivity, but with limited operating temperature (< ~80 °C). It.is a sulfonated tetrafluoroethylene based fluoropolymer-copolymer [1880] in which the hydrophilic side chains are organized to form a network of inter-connected channels and nanometric (2–6 nm) cavities that fill with water upon hydration and partially empty on dehydration, so enabling the water transport through the membrane and supporting excellent proton conduction. Nafion has been extensively used in EZ (exclusion zone) experiments. Nafion strongly absorbs water [3242].


Although proposed to be stable in earlier experiments, recently it has been proposed to produce a slowly separating colloidal crystal at its surface [2388] and release a large number of protons [2388c]. At higher temperatures, a proportion will dissolve in proportion to its swelling [3243]. If correct, this may call into question much of the EZ work using this material and cause confusion over the real properties of the EZ; particularly as Nafion has been the material of choice in many studies as it shows a maximal easily obtained EZ. Work on EZ using other hydrophilic materials are not affected by this finding.


Investigations into the water content of Nafion and other sulfonated polymers showed strong relationships among the different types of water (strongly bound, weakly bound and free) dependent on the charge density and distribution

[2888]. A model of the Nafion® membrane, incorporating water content, activity-dependent diffusivity and volume expansion has been used to resolve Nafion's water activity profiles [3205].


The closely related disulfonated poly(arylene ether) and polyimide copolymer proton exchange membranes have been discussed [3163]. [Back]


f As the length of a linear polyelectrolyte increases, the chain folds up into a ball with consequential osmotic pressure increase [2705]. [Back]


g The Nernst layer. Walther Hermann Nernst, (1864 – 1941) was a German chemist who won the 1920 Nobel Prize in chemistry. The Nernst layer is a theoretical layer corresponding to the green line in the above Figure which shows the concentration profile along the direction perpendicular to a catalytic surface coinciding with the tangent to the true concentration profile at the interface. The effective thickness δ of this diffusion layer is that it would have if the concentration profile were a straight line coinciding with the tangent to the true concentration profile at the interface and that straight line is extended up to the point where the bulk concentration is reached.


It is impossible to stir any solution right up to an interface. there is always an unstirred layer adjacent the surface, usually extending to between 20 μ to 0.5 mm thick, even with stirring. Strong stirring will reduce this but still leave unstirred layers several µm thick, dependent on the diffusivity. Extreme stirring has no further effect due to the bulk turbulence as there is always laminar flow at the surface.


The Knudsen layer. The Knudsen layer is a similar non-equilibrium unstirred layer (named after Martin Knudsen, 1871 - 1949) on the gas side of the gas-liquid interface. The thickness of this layer is several times the mean free path of the mixture. The mean free path is proportional to the viscosity and the square-root of the temperature, and inversely proportional to the pressure and the square-root of the molecular mass of the mixture. At 20 °C and atmospheric pressure, the mean-free-path of water vapor and dry air are 2.16 µm and 0.07 µm. [2714]. [Back]



Home | Site Index | Conventional colligative properties | Water activity | The anomalies of water | Water: Introduction | The icosahedral water clusters | LSBU | Top


This page was established in 2014 and last updated by Martin Chaplin on 5 April, 2018

Creative Commons License
This work is licensed under a Creative Commons Attribution
-Noncommercial-No Derivative Works 2.0 UK: England & Wales License


The preferred Web Browser is Mozilla Firefox