I think the best answer at the moment is that we don't know. There is a large effort on understanding this at the moment and a good deal of disagreement in the literature. Some even claim that the surface has a positive charge. I suggest you look at articles by James Beattie, Richard Saykally and Pavel Jungwirth to follow the debate. I think one of the complications is that when you are talking about the interface you need to understand the region of the interface that you are referring to. (what depth into the interface). Different experimental techniques will probe different depths and for most techniques this depth is not known. There are also challenges with the theoretical interpretation of simple measurements such as surface tension and how ions alter the surface tension, especially when you consider that concentrations near the interface are not going to be monotonic but oscillatory. Great question, but I think it will be some years yet before we have a clear answer.

Dear John,

thank you for the links. You are right - the excess of hydroxide ions explains the negative charge of air/water interface. The presence of the ions decreases the surface energy due to Coulomb repulsion. If the surface adsorbs the hydroxide ions along with the more rapid protons from water volume then the net result will be negative only in the case of OH-hydrogen bonding with the surface H2O molecules. These processes are accompanied by diffusion of positive impurity counter ions and formation of a double electric layer. Counter ions partly screens the repulsion between the hydroxide ions that increases the surface tension. On the other hand, the repulsion between the counter ions lowers the surface tension. Obviously salinity and pH of water control the rates of these processes. Do you know any computer simulation of the picture? And the question stays what is the molecular structure of the surface?

Best regards

There is a paper published in the 1980s: Farrell and McTigue J. Electroanal. Chem. 139; 37 (1982) who used some beautiful electrochemical methods to measure the surface potential of water. This is ascribed to an oriented dipole layer at the air/water interface a few molecular diameters thick. This produces a dipole moment distribution near the interface.

I think the best answer at the moment is that we don't know. There is a large effort on understanding this at the moment and a good deal of disagreement in the literature. Some even claim that the surface has a positive charge. I suggest you look at articles by James Beattie, Richard Saykally and Pavel Jungwirth to follow the debate. I think one of the complications is that when you are talking about the interface you need to understand the region of the interface that you are referring to. (what depth into the interface). Different experimental techniques will probe different depths and for most techniques this depth is not known. There are also challenges with the theoretical interpretation of simple measurements such as surface tension and how ions alter the surface tension, especially when you consider that concentrations near the interface are not going to be monotonic but oscillatory. Great question, but I think it will be some years yet before we have a clear answer.

Vincent is correct. I've attached a very noteworthy paper that discusses interactions at the interface. This paper, by Jungwirth and Tobias, highlights interactions of ions and the impact on surface tension through a molecular dynamics standpoint. Jungwirth and Tobias have investigated this question very heavily over the past decade. I encourage you to continue researching beyond this article as this remains a very important topic in atmospheric science.

I agree with Vincent and Michael. The truth is, we don't really know. When a phenomenon is sensitive to how one measures it, that's a very strong indication we don't even know enough about the phenomenon to measure it properly. True, lots of theoretical analyses seem to indicate a negative charge on the surface, but so are there theoretical analyses that indicate the reverse. Theoretical models of water are notoriously inaccurate and any result they might yield should be considered with very careful deliberation.

I don't want to dis the many excellent researchers in this area, but many of these experimental studies are done "open dish" with the surface exposed to the ambient atmosphere. Carbon dioxide absorption is famous for lowering the pH of pure water after only a few minutes. If the sample is small, and the surface area is great, this process can occur very quickly. And, of course temperature also plays a critical roll. Considering how easy it is to push the pH of water a few tenths above or below the theoretical ideal of 7, these experiments are very difficult to execute and sometimes even reproduce between any two researchers.

And, to make the result even more muddled, let me add the following: I have used Sum Frequency Generation Vibrational Spectroscopy (SFG for short) to study the surface of aqueous systems, and in particular single crystal ice, for a number of years. SFG is extremely selective to the top molecular layer at an interface. In the air/water system the spectrum is extremely sensitive to the pH of the surface. This is because the presence of any charge distorts the molecular bonding motifs and average molecular orientation at the surface. Because SFG is a non-linear process, it generates a complex (ie: real and imaginary) signal that is also polarization and orientation sensitive to only the top molecular layer at the interface. All these characteristics make SFG a unique and powerful tool for studying interface (ie: surface) chemistry.

Along with single crystal ice surfaces, I have studied pure water, aqueous NaOH and aqueous HCl systems using SFG. There is ample and clear experimental evidence in the literature that basic solutions have a net negative charge at the surface, and acidic solutions the reverse. And indeed, there are characteristic spectral signatures of each type of interface when studied with SFG. Finally, when looking at a pure water (18M Ohm, in a close container, at zero degrees C) using SFG, it's clear that the interface appears to be neutral, having no net negative or positive charge. If there is a net charge at the surface of pure water, the concentration is well below the detection limit of SFG. That is to say, a small fraction of a monolayer.

Kak govorila odna iz prepodovatelnec - "a gde tam zaryad?" ....

The effect of double charged layer on the air/liquid interface has been described quantitatively by Kharkats and Ulstrup, see http://dx.doi.org/10.1016/0022-0728(91)85055-T and http://dx.doi.org/10.1021/jp026229d. The negative charge of a clean water surface is caused by finite-size stratification of hydroxide ions, and presence of any impurities should increase the effect due to hydration.

I couldn't find any works on the molecular dynamics calculations for that cases, though.

In my opinion a possible explanation for the negative charge is determined by the CO2 molecules from air, which dissolve in water at the interface. In the same way is explained also the pH decrease of water when is expose to air a longer time.

Dipoles. But I haven't yet published it.

I also estimated the number of H-bonds per unit area - it can be done through various surface tension data (surface tension of pure water, surface tension of various inorganic acids and bases in aqueous solution). It is very low - one per 1-3 sq. nm. Water minimizes that at the expense of entropy.

Let's look at the facts:

1. Experiment on drainage on foam film from ultra-pure water shows absence of charge; - Very high level of purity of the system.

2. Experiment on measuring zeta potential of bubbles in pure water shows negative potential - decreased level of purity of the system;

3. Electro-chemical experiments shows dipole moments orientated with oxygen toward air's phase; - I don't know the level of purity, but I suppose it is quite high.

4. SFG on Air/Water interface shows absence of charge on Air/Water interface - I don't know the level of purity, but I suppose it is quite high.

An answer, see: http://www.waterjournal.org/uploads/vol1/chaplin/WATER-Vol1-Chaplin.pdf

Dear Vladimir,

I suppose that it is also noteworthy that dipole-dipole interactions also govern to a much extent the situation at the water-solid interface, as is is clearly seen from the experiments related to the hydrophobic recovery of plasma-treated polymers.

And this quite surprising, because London forces are few orders of magnitude stronger than dipole-dipole interactions. However, dipole-dipole interactions will prevail on London forces when dipoles are at least partially fixed, see:

Applied Surface Science 273 (2013) 549– 553

Ed. Bormashenko

I am not in the field but I may suggest a very qualitative idea: hydrogen bonding. The molecules in the bulk have statistically the same probability to donate and receive H-bonds. On the surface there may be a larger quantity of them donating toward the bulk but not receiving (obviously) from the air, causing a preferential orientation of the dipole. If this is approximately correct, the negative charge at the surface should be dependent on temperature, as the degree of total H-bonding is. Cooler water should be more negatively charged.

All fascinating stuff! Since the question was frame as a "why," I'd like to go back to the core of the question, just for fun, and answer more generally (why) rather than specifically (i.e., how). I know all of you previous responders are aware of these, but maybe someone else reading our string could benefit... or maybe not. :o If not, sorry to be so basic in my reply:

If under any given system condition an air-water interface is negative (or positive), it is because (1) there is a positive charge distribution (or negative) below or above the interface (however you wish to define the interface; perhaps "interphase" is better as many have suggested); and/or (2) there is a fugacity distribution in the vicinity of the interphase. (Let's not resort to chemical potentials just yet, which are not nearly as convenient thermodynamically... and I just like using the word "fugacity" as much as possible!) I often find it helps to go back to first principles to understand how and why many seemingly different explanations can produce the same type of observables (viz., the so-called law of electroneutrality and better yet, the third set of criteria for chemical equilibria--the fugacities of a given species in solution must be equal and uniform everywhere to be in equilibrium, including all contacting phases, so both air and water).

I think that water interfae can be charged negatively like mica in water. Negative charge at mica surface in water is proved by experimental way. Similarly, there is additional experimental evidence of negative charge at water interface in the field of Langmuir monolayer films. Indeed, it is well known that ordered monomolecular film of fatty acid cannot be transferred to solid surface without distortion due to the negative charge of the carboxyl groups of fatty acid. These carboxyl groups are located exactly at water interface and its degree of dissociation is directly related to the properties of the surface layer of water and it dissociation would not occur in case of excess of positive charge at the interface.

Negative charge at water surface can be stipulated by minimization of surface free energy. Possibly, the water molecules dipoles are rotated outwardly to hydrophobic air due to the "hydrophobicity" of the oxygen atoms relative to protons.

Anyone wants to comment on my hypothesis? I would like to know if it is totally farfetched. Thank you.

Dear colleagues,

thank you for your replies and links. But, obviously, the question needs to be clarified in order to narrow the range of possible answers. Let's agree that:

1. We are interested in the theoretical interpretation of the negative charge of the water surface.

2. The aim of our discussion is to identify the most commonly accepted model of the surface and the subsequent formation of the electric double layer.

3. Air/water interface is a layer of several angstroms in width (a monolayer, if one may say so on the surface of the liquid). This layer is sometimes called “ice-like clathrate shell”.

4. Since all known experimental data are interpreted in terms of different models , they should not be absolute. Let us trust in the common sense!

Best regards

As mentioned in earlier post by Pivovarenko, I would have a look at Martin Chaplin's mini review on this issue, in particular the information from vibrational spectroscopy and X-ray spectroscopy from liquid jets that discusses the amount and geometry of dangling OH, as well as arguments inferred from macroscopic data such as surface tension

http://www.waterjournal.org/uploads/vol1/chaplin/WATER-Vol1-Chaplin.pdf

Dear Vladimir,

I would suggest you'd have a look at the following paper:

Katarzyna Hänni-Ciunel_Natascha Schelero_Regine von Klitzing_Water-air interface charge_FaradayDiscuss_v141_pp41-53_2009

These colleagues have performed a careful experimental work to study the problem in detail. Now, to answer all of your questions in detail ought to require some additional theoretical work ...

Dear Giacomo,

You are right concluding: “On the surface … a preferential orientation of the dipole (may exists)”. And that is under my question (look above).

You point on a possible temperature dependence of the negative charge but it may be shielded by the temperature dependence of diffuse positive charge which forms the electric double layer.

Thanks Vladimir, for your clarification.

I feel awkward to talk about this subject since for the last 20 years I dealt mostly with conservation of Cultural Heritage problems. Nevertheless ...

Lets take a big, perfectly pure quartz synthetic crystal. In the bulk all atoms are saturated, receiving the proper number of bonds. On the faces, due to the termination of the crystal, oxygen atoms receive one positive charge only from Si and Si are one negative charge short from the missing O. The resulting charges are saturated by absorbing gases, humidity etc. so that the surface may result neutral. In water, things are more complex because in the bulk not all molecules are making 4 H-bond (this may happen at 4 degrees centigrade, when liquid water reached the maximum density and certainly in ice, more similar to the quartz example) and are relatively free to move. What may count is to compare the probability that a water molecule at the interface may receive or donate H-bonds. If it is statistically more likely that it donates, that the negative residual charges will be directed toward the surface. I leave it here to reduce the chance of making a fool out of myself.

The discussion is highly interesting and I agree with the necessity of considering hydrogen bonding (e.g. Giacomo Chiari). The original question focused on the air-water interface even if a part of the discussion touched also other situations. It is well known that in water (but not only there, of course) the water molecules are interbonded via hydrogen bonding. Hydrogen-bonded water molecules are necessarily polarized and their oxygen atoms are negatively charged. Consequently, one must expect that any water surface is negatively charged, unless the negative charges have been saturated by “something else”; however, in this case at the surface we have “something else”.

To those interested in the critical analysis of the possible explanations, I would recommend our article "Charging of Oil-Water Interfaces Due to Spontaneous Adsorption of Hydroxyl Ions", Langmuir 12 (1996) 2045. In this article we succeeded to demonstrate unambiguously the following:

1. To show that all procedures for purification of the aqueous and oily phases from possible polar contaminants (e.g., fatty acids and anionic surfactants) did not remove the negative potential of the oily droplets, as determined from their electrophoretic mobility.

2. To show that the negative zeta potential of oily drops is practically independent of the type of oil, if the latter is nonpolar. At that time we studied aliphatic, aromatic and silicone oils and the results for the electrophoretic mobility were very similar for all of them. In parallel, French scientists measured very similar data with air bubbles and later we obtained very similar results with perfluorinated oils.

These results clearly demonstrate that the negative potential could not be due to contamination in the oily phase (these contaminations could not be the same for all different oils studied).

Most importantly, these results demonstrate that the potential is governed exclusively by the aqueous phase which is in contact with a nonpolar phase (air or oil) - the nonpolar phase itself is just creating the boundary where the charging occurs.

3. To demonstrate the strong pH dependence of the negative potential which means that the ions involved in water dissociation (H+ and OH-) are the only possible source of charging - these are the potential-determining ions.

4. To show experimentally with foam and emulsion films that there is a clear electrostatic repulsion between two air-water or oil-water interfaces, placed in contact with each other. Therefore, the negative surface potential was registered by independent method, beside the electrophoretic mobility.

4. We used a series of experiments to discard several possible hypothesis for the origin of the electrical potential (incl. some of the hypotheses still discussed today). The final conclusion of the critical analysis of all data was that the only possible explanation is the preferential adsorption of OH- ions at the interface.

5. Based on the above assumption, we calculated the area per adsorbed OH- ion and the specific adsorption energy. This adsorption energy turned out to be in the range of the hydrogen bond energies.

6. Finally, we proposed a mechanistic explanation which includes the following key elements: (1) a layer of boundary water molecules is formed in contact with the nonpolar phase. (2) This layer of water molecules is oriented with oxygen atoms towards the nonpolar phase. (3) The hydrogen atoms oriented toward the aqueous phase serve as a substrate for adsorption of OH- ions via hydrogen bonding.

For the years that passed from these experiments and their interpretation, I have not seen any experiment or explanation that discards any of our conclusions.

Therefore, I have no doubt that, when "detected" from the aqueous phase (electrophoresis, in foam or emulsion films), these surfaces are negatively charged and this is not due to contaminations or oriented dipoles (the dipoles cannot explain the strong pH dependence).

The real question for me is why the other methods (e.g. spectroscopy of the boundary layer) give different results?! I do not know the answer to this question.

Here is the paper

Article Charging of Oil−Water Interfaces Due to Spontaneous Adsorpti...

I agree with Vincent Craig

Experimental results by Nikolai Denkov are confirmed for the interesting water/ porous graphitic carbon interface. I used a chromatographic technique and I found that both the proton and the hydroxide are adsorbophilic (they do adsorb at the interface) but hydroxide is more prone to do so. I've measured the adsorption isotherms of both HCl and NaOH and I compared their adsorption to that of NaCl (that is lower).

dx.doi.org/10.1021/jp409480t | J. Phys. Chem. C 2013, 117, 25579−25585

I agree that a solid surface cannot perfectly mimic air/water interface but they are both dielectric boundaries, hence their behavior has to be compared for a full understanding of why the Onsanger−Samaras theory view (ion free interface) seem to be completely wrong.

Also because the solid hydrophobic interface interface may be easier to study

Vol 141 of Faraday Discussion is very inspiring:

https://www.mpibpc.mpg.de/277769/Link_Faraday_2009.pdf

An explanation of the negative charge of surface water is also here:

Phys. Chem. Chem. Phys., 2009, 11, 10994–11005

Dear Dr. Cecchi,

Thank you for the useful links. I’d be appreciate if you send me a copy of the last one.

Concerning the fact you pointed to “…I found that both the proton and the hydroxide are adsorbophilic (they do adsorb at the interface) but hydroxide is more prone to do so...” – please look at my second post in the discussion where I wrote “ If the surface adsorbs the hydroxide ions along with the more rapid protons from water volume then the net result will be negative only in the case of OH-hydrogen bonding with the surface H2O molecules.” That is what I’d wander to simulate on PC by solving a set of differential equations. But such calculations need some experimental parameters. For example, could you estimate how much “hydroxide is more prone…” from your data?

Best regards

Dear Dr. Shatalov, thank for underlying that sentence, I've just sent you the paper on adsorption of HCl and NaOH onto the water/porous graphitic carbon interface

and I have added another concerning the adsorption of the kosmotropic LiF onto the same interface

Yes, from my results the standard free energy of adsorption of both analytes can be calculated, please feel free to ask me, the technique is not very common among physical chemists....

There's also another interesting paper "An explanation for the charge on water’s surface": Phys. Chem. Chem. Phys., 2009, 11, 10994–11005

that should be taken into account when considering water surface pH

H(+) ions have smaller radius than OH(-) ions. Due to their higher hydration energy, they are rejected from non-polar air, leaving excess oh OH(-) ions near interface.

PDF file on the Hydration Effects is attached.

Dear Alexander,

Thank you for your comment and attached file. I've just read the papers pointed by Dr. T. Cecchi and some others and I have to say the question has no indisputably a definite answer. That is the reason why I collect a variety of opinions. My futher aim is to investigate how air/water double layer would response to RF or IR irradiation. I hope the discussion will help me.

Sure, Vladimir, I am glad to help. I tried to send you a messege, but it dissappiered from the screen.

Luchshe na Russkom, ne poymut. Ya nadeyus', chto u vas utresetsya, zarplatu budet chem platit'. A tak, sidit sebe tixiy uchenyi, i nate vam. Stalin, v svoe vremya, dal prikaz:

sperva Banderovtsa ubit', a potom oformit' yuridicheski. Chelovek on byl lyutyi, no inogda eto shlo na pol'zu.

Sure, Alexander, you may contact me by E-mail: [email protected]

I really agree with John in his opinion

In the case of sea water, Na(+) has higher hydration energy than Cl (-). Thus, Na(+) will have stronger repulsion from Water - Air Interface, leaving Interface negatively charged. This effect will be dominant in comparison with contribution of H and OH ions.

Dear Alexander, I am not an expert in this area and, please correct me if I´m wrong, but I thought that Cl- has higher energy of hydration that Na+ (at least looking at the tables of e.g. Schmid, R.; Miah, A. M.; Sapunov, V. N. A new table of the thermodynamic quantities of ionic hydration: values and some applications (enthalpy-entropy compensation and Born radii). Phys. Chem. Chem. Phys. 2000, 2 (1), 97−102.). Is there any additional effects that makes this change in the case of sea water?

Svetlozar, please see Table 1 in the paper attached. The results are correct. Many sources on Hydration effects are incorrect, as people don't have a clear picture in their minds.

Thanks a lot Alexander, I´ll definitely read it!

I completely agree with Alexander that the issue of hydration of anions are often interpreted incorrectly, which in many cases is the result of the use of erroneous data on the hydration energy of anions.

Without going into the wider interpretation, can be said that the significantly lower hydration energy of the anions compared to the corresponding energy for most of the cations, is a clear consequence of the presence of repulsive forces between the anions and the free electron pairs on the oxygen atom of the water molecule.

Thanks, Andrzej, for support.

I did not read all of the answers but to me the dangling bonds at the interface (water molecules does not establish all of the three or four bonds as other bulk molecules do) leaves a total negative equivalent charge. We believe to have shown that in an article (see my list) Caer et al.

Have a good day !

Thank you, Parsel, you are the first who prove a scenario from my second post!

Pascale, water molecule can exist seperately in the vapor phase, that is neutral. Moleculas are neutral, ions are charged.

Of course Alexander but water has a strong permanent dipole . When molecules are disposed erratically next to each other this effect is cancelled but if there is a preferred orientation there could be a non zero charge available for molecules at the interface to come in contact. Hence the negative charge "effect"

To clear up, which variant is more appropriated - Alexander's or Pascal’s - that was the main goal of my question. There exist many evidences in favor to the both. It seems to me to make a final choice one needs to measure dynamical surface tension. If the surface charge is originated by hydroxyls then it takes some time to state a steady surface tension. It consists of two independent phases including hydroxyls and counter ions diffusion from volume to the surface. If the surface charge belongs to the molecular structure of the surface layer (the normally oriented dipoles) then the steady state formation includes only one phase - counter ions diffusion. The question stays – can any experiment distingue the cases?

Possibly surface probing by infrared. By varrying the angle one can favour the surface response. Hydroxyl gives a diffrent OH stretching absorption ( higher frequency). But unless the OH is unambiguous (nice narrow band) it may just appear as a shoulder on the water OH. Moreover, you would need super pure water...

Yes, Pascal, a lot of unexpected difficulties. And it is unclear whether the solution of this problem of high costs?

indeed this behavior is related to the polarity of water molecules, and its configuration. The packing of these molecules occurs in the direction of lowest energy for the system, that corresponds to the oxygen atoms at the interface.

but the question stays how does the interface form itself. Does the ice-like monomolecular film with co-oriented dipoles is generated at the same time when the density junction go rise? If this is an ethropy effect then it'll take some time to change dipoles orientation and to form a clatrate net of H-bonds. Otherwords, it's interesting how the surface born, is not it?

Most of colloid particles are negatively charged at neutral pH. Titrating the water phase by acid, zero point charge can be reached, at pH < 3, the interface becomes positively charged.

It is mostly argued that the oil-water interface is negatively charged because there will be strong specific adsorption of OH- ions at the oil-water interface. As many of the above mentioned there are even studies where people have argued that the surface has positive charge. However it is widely accepted that the surface is negatively charged.

In case of oil-water it is suggested to be due to contamination by carboxylic acids formed by oxidation-

Dear Hari and Titus,

if the oil-water and air-water interfaces are negatively charged than the surface structure is the same in the both cases. That may be OH- ions H-bonded to the surface, or unidirected H2O dipoles. What is right? - that is the question.

For oil-water it can have a different or better said additional mechanism, namely oxidized oil. Of course contamination can also contribute to negative charge at the interface. This does not mean automatically that there is no reasonable explanation for its existence in pure systems.