Desalination due to Electrical Image Forces

J. B. Sokoloff

REVIEW 4 major objections 5 minor 18 references

Desalination due to Electrical Image Forces

T0 review · 4 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Electrical image forces can push salt ions past their solubility limit at a clean metal wall, suggesting a membrane-free desalination method.

desk verdict A clear, honest theory paper that applies standard image-charge electrostatics to a new desalination idea, but the quantitative claim rests on a fragile transfer of the near-wall permittivity from confined graphene-slit water to a single metal plate; worth serious peer review, not a desk reject. read the letter →

arxiv 1908.05359 v1 pith:UTRNUQBW submitted 2019-08-14 cond-mat.soft cond-mat.mes-hall

J. B. Sokoloff This is my paper · ORCID

classification cond-mat.softcond-mat.mes-hall

keywords desalinationimagechargeforceselectricaldoublelayerBjerrumlengthinterfacialwaterdielectricconstantionprecipitationmetallicsurfacessaltsolubility

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper argues that electrical image forces can push dissolved salt ions toward a metallic wall strongly enough that their local concentration exceeds the salt's solubility limit, causing salt to precipitate out. The effect relies on water near the wall having a much lower dielectric constant than bulk water, which lengthens the local Bjerrum length to tens of nanometers and makes the image attraction reach far into the solution. A second attractive force is predicted at the plane where the dielectric constant jumps back to its bulk value, creating another precipitation zone free of a physical wall. If correct, flowing salt water between closely spaced conducting plates could desalinate it without membranes, electrodes, or periodic fouling.

What carries the argument

The central object is the electrical image potential of an ion near a conducting wall, combined with a local Bjerrum length $\ell_B = q^2/(4\pi\epsilon k_BT)$ evaluated with the near-wall permittivity ($\epsilon/\epsilon_0 = 2.1$), which raises $\ell_B$ to about 28 nm. The argument balances the image force against osmotic pressure and integrates the resulting differential equation to obtain an ion density profile $\rho_i(h) = \rho_i^0 \exp[\ell_B/(4h)]$ near the wall. The second mechanism is an image potential produced by the dielectric jump from 2.1 to 81 at about 1 nm from the wall, represented by terms of the form $\ell'_B k_BT/[4(h_0-h)+\Delta]$ on the wall side and $\ell''_B k_BT/[4(h-h_0)+\Delta]$ beyond it. These ingredients turn an equilibrium force balance into a prediction of localized supersaturation and precipitation.

What would settle it

Measure the equilibrium ion concentration profile within a few nanometers of a clean, oxide-free metal surface in contact with a sodium chloride solution below its bulk solubility limit; if no crystalline salt layer appears against the wall, or if the measured interfacial permittivity is close to bulk water's, the proposed precipitation mechanism is not operating.

Watch

Extended reading notes

Core claim

The paper's central claim is that, at equilibrium, the image-charge force pulling an ion toward a metal wall is balanced by osmotic pressure, and because the Bjerrum length evaluated with the near-wall permittivity ($\ell_B = q^2/(4\pi\epsilon_1 k_BT)$) is about 28 nm rather than 0.7 nm, the Boltzmann factor $\exp[\ell_B/(4h)]$ makes the ion density near the wall exceed its bulk value by orders of magnitude. For sodium chloride, the predicted concentration exceeds the solubility limit within a few nanometers of the wall, so salt precipitates out. In addition, the paper predicts an attractive image potential near the plane where water's dielectric constant changes from 2.1 to 81, with a minimum roughly one molecular diameter from that plane, producing a second precipitation zone that is not attached to a solid surface. The treatment is worked out for salts with ions of equal charge magnitude, and it notes that ions carrying larger charge are enhanced more strongly.

Load-bearing premise

The argument stands on the claim that water within about 1 nm of a clean metal wall has a dielectric constant of about 2.1, as measured for nanoconfined water; if the interfacial low-permittivity layer is thinner or more polar, the near-wall Bjerrum length drops from 28 nm toward 0.7 nm and the predicted supersaturation largely disappears.

Editorial extensions

If this is right

If the predicted supersaturation is real, a vertical stack of clean conducting plates with sub-micron gaps should precipitate salt from flowing seawater and let the salt fall out under gravity, leaving desalinated water behind.
Because the second precipitation zone sits at the dielectric-permittivity jump plane rather than on a solid surface, the mechanism should also work for nonmetallic walls, broadening the range of candidate materials.
The proposed device would need no membranes or electrodes, avoiding the periodic fouling and cleaning that reverse osmosis and capacitive desalination require.
The enhancement is stronger for ions with larger charge magnitude, so the mechanism should preferentially remove multivalent ions such as calcium and magnesium.
Thick oxide coatings and surface roughness suppress the near-wall peak, so practical implementations would likely need clean, smooth conductors such as doped graphene.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A natural extension is that the same near-wall enhancement should appear in other interfacial processes, such as corrosion, electrodeposition, and heterogeneous nucleation, because the underlying image-force balance is not specific to desalination; this would make precipitation one instance of a broader interfacial-ion-concentration effect.
A direct test would be surface-specific spectroscopy or X-ray reflectivity on a clean metal in contact with an undersaturated salt solution; observing a crystalline salt layer within a nanometer of the wall would confirm the mechanism.
The calculation uses a sharp dielectric transition; using a smoother permittivity profile or explicit hydration structure could shift the predicted peak position but would likely keep a near-wall enhancement if the low-permittivity layer exists.
Throughput would be limited by ion diffusion into the near-wall region and by the osmotic work needed to maintain flow; the paper's estimates of millisecond-scale replenishment for micron-spaced plates could be checked with a small prototype.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Joe,

Quick take: this is a clear, short theory paper that transfers standard image-charge electrostatics to a new application—precipitating salt near a clean metal wall and near the dielectric jump in water—and the central quantitative prediction is only as good as the assumption that water within ~1 nm of a metal plate has ε≈2.1. That assumption comes from Ref. [14], a measurement on water confined between graphene sheets in a slit, not from water next to a single clean metal electrode. The geometry mismatch is real and load-bearing. But the author does honest thermodynamics, flags his own limitation about clean smooth metals, and proposes a concrete, falsifiable device. It deserves a serious referee, not a desk reject.

What is new: the specific proposal to use image-force-enhanced concentration at a metallic wall and at the plane where water's permittivity jumps from 2.1 to 81 as a desalination mechanism. Previous image-force work cited deals with surface tension, capacitive desalination, and confined ionic liquids. The paper gives an explicit equilibrium Boltzmann profile with two screening options, and a rough time-scale argument for how fast the near-wall population forms.

Where it holds up: the osmotic-pressure balance and the derivation of Eq. (5) are standard and clear. The parameter values are mostly taken from independent work. The author openly states the oxide and roughness caveats, and the roughness calculation shows why a smooth conductor matters. The dielectric-jump peak at h0−Δ is a nice touch because it is not blocked by a physical wall, so precipitate can fall away.

Soft spots, in proportion: the near-wall permittivity is the load-bearing input. The paper first says 'it is known' that ε/ε0=2.1 within 1 nm of a solid surface [14], and later softens to 'believed'; Ref. [14] measured confined water between graphene sheets. If ε near a single metal wall is 5–10, the predicted supersaturation shrinks or disappears. That is a genuine external-evidence risk, not an internal inconsistency. Second, the two screening models give different precipitation scales—correlation-hole gives h≤4.35 nm, Debye–Hückel gives Angstroms. The author notes this but does not resolve it. Third, supersaturation is not precipitation. The paper estimates ion diffusion times but does not address nucleation barriers, critical nucleus size, or whether the flow residence time is long enough. That is a bigger gap than the author acknowledges.

Citation pattern is fine; the only self-citation is to an unpublished previous desalination proposal, which is a minor stylistic point.

Who this is for: anyone working on desalination or interfacial electrostatics. It is a hypothesis-generating paper with a clean experimental target: clean graphene-coated plates. My recommendation: send it to peer review, with referees asked to scrutinize the transferability of the dielectric data and the nucleation step. If those hold, the mechanism is plausible; if not, the desalination claim collapses, but the paper still contains a useful discussion of image forces near dielectric jumps.

Referee Report

4 major / 5 minor

Summary. The manuscript proposes that electrical image forces can drive salt ions out of solution near a metallic wall. It derives a Boltzmann-type concentration profile from an osmotic-pressure balance, using a Bjerrum length based on a near-wall relative permittivity of about 2.1, and argues that the ion density within a few angstroms of the wall can exceed the salt solubility limit, causing precipitation. A second image-force mechanism is proposed at the plane where the water permittivity is assumed to jump from 2.1 to 81, creating another preferential precipitation zone. The paper suggests that an array of closely spaced, clean, smooth conducting plates could desalinate water without membranes or electrodes. It also gives order-of-magnitude estimates for the equilibration time and discusses the effects of oxide coatings, surface roughness, and unequal ion valences.

Significance. If the central mechanism is correct, this would be a conceptually new desalination route that avoids membranes, electrodes, and the associated fouling, and it would extend the well-known Onsager-Samaras image-charge picture from surface tension to ion precipitation. The paper has genuine strengths: the osmotic-pressure balance and the resulting concentration profile follow standard electrostatics and equilibrium thermodynamics; the input parameters are mostly taken from independent experiments; and the prediction is falsifiable in the sense that a vertical plate array either precipitates salt or it does not. However, the quantitative claim hinges on a single, weakly supported input—the persistence of a low-permittivity (ε/ε0 ≈ 2.1) water layer about 1 nm from a single metal wall—and on a screening model for which the two offered choices give very different precipitation zones. The significance is therefore conditional on closing that evidential gap.

major comments (4)

[after Eq. (6), paragraph beginning 'Although in bulk water...'] The central prediction of a supersaturated ion layer rests on the assertion that within about 1 nm of a solid surface ε/ε0 = 2.1, cited to Ref. [14]. Reference [14] (Fumagalli et al.) reports the effective out-of-plane dielectric constant of water confined between two graphene sheets in a slit of ~1 nm; it does not measure a 1 nm thick low-permittivity layer adjacent to a single clean metal plate in contact with bulk water. The manuscript itself uses the word 'believed' for this input, and later narrows the applicability to clean, smooth metals. The mechanism is quantitatively fragile: with ε/ε0 = 5 the near-wall Bjerrum length drops from 28 nm to about 12 nm, and with ε/ε0 = 30 it drops to about 2 nm, at which point even an ion at the distance of closest approach (h = a ≈ 0.3 nm) has an enhancement of only exp(2/(4×0.3)) ≈ 5, below the roughly tenfold enhancement needed to exceed NaCl solubility in seawater. Because Eqs. (5), (6), (9), and (10) all inherit this input, the desalination claim is not robust to the most plausible uncertainty in the near-wall permittivity.
[paragraph following Eq. (6), sentence beginning 'Using ρ_i^{-1/3}...'] The two screening models are not merely alternative refinements; they give qualitatively different predictions for the precipitation zone. The correlation-hole model gives precipitation for h ≤ 4.35 nm, while the Debye-Hückel model gives precipitation only for h on the order of angstroms. This is a factor of ten or more in the width of the proposed precipitation layer, and it directly affects the amount of salt that could be removed by a plate array. The manuscript does not explain which model is more appropriate for the concentrated, strongly correlated layer near the wall, nor does it discuss that the correlation-hole screening length itself depends on the local ion density ρ_i(h), creating a self-consistency problem that is not addressed in Eqs. (4) and (5).
[Eqs. (7)-(10), second desalination mechanism] The second proposed precipitation zone at the plane where the permittivity jumps from 2.1 to 81 relies on the same unverified near-wall permittivity profile, but in addition the interpolation formula with Δ of molecular size produces a potential minimum whose depth is ℓ_B' kT / (8Δ). With ℓ_B' ≈ 28 nm and Δ ≈ 0.3 nm this is many kT, implying essentially complete pinning of ions to the plane h = h0 − Δ. Such a strong predicted accumulation is a striking, quantitative consequence, not a small correction, and it deserves a consistency check: for example, the paper should show that the local ion density at the minimum is not so high that the dielectric-constant profile itself or the point-charge image formula breaks down. As written, the strength of the second mechanism is itself a red flag that the assumed step-like permittivity profile and the linear-response image treatment are being pushed beyond their range of validity.
[full text, paragraph 'Although the increase of the dielectric constant...'] The desalination scenario requires that precipitated salt be transported out of the gap by gravity while the solution continues to flow. The manuscript estimates the diffusion time for ions but does not estimate the growth and sedimentation time of salt crystallites, nor does it compare the gravitational body force on a precipitate with the drag and adhesive forces it will experience in a sub-micron gap. Without such an estimate, the statement that 'the salt can just fall out from between the plates under the force of gravity' is not quantitatively supported. This is a load-bearing step for the proposed device, even though it is not load-bearing for the purely equilibrium image-force calculation.

minor comments (5)

[References, Ref. [1]] The author name is misspelled as 'Onsager and Samara'; the correct spelling is Onsager and Samaras.
[References, Refs. [9] and [10]] Ref. [9] appears with a period in '9.10' in the introduction ('freezing of a confined room temperature ionic liquid[9.10]'); the punctuation should be '[9,10]'.
[References, Ref. [7]] 'Phs. Rev. E' should be 'Phys. Rev. E'.
[Section 'Let us consider how rapidly equilibrium is established...'] 'If the plates were 1 m apart' should read '1 μm apart'; the following diffusion-time estimate uses 1 μm.
[Throughout] There are several typographical issues in the equations, such as missing subscripts and inconsistent notation (e.g., ℓ_s vs. s), and the second occurrence of 'salt water' in the abstract is followed by a double period. A careful proofread would improve readability.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the predicted ion enhancement is derived from standard electrostatics plus an externally measured near-wall permittivity, not from the desalination claim.

full rationale

The derivation chain is not circular. The concentration enhancement in Eqs. (5)-(6) follows from the standard Boltzmann balance between osmotic pressure and the image-charge force, Eqs. (1)-(4), with the near-wall Bjerrum length l_B = 28 nm obtained from the externally measured value epsilon/epsilon0 = 2.1 attributed to Ref. [14]. That permittivity is an independent experimental input, not a parameter fitted to desalination outcomes, and the predicted supersaturation threshold is not used to define any term in the model. The second precipitation zone near h0 is likewise a direct application of the classical image-charge potential at a dielectric step, Eqs. (7)-(10), not a renaming of a known salt-removal result. The paper's own caveats, including oxide layers, roughness, nonzero electronic screening length, possible repulsive dielectric image contribution, and the 'believed' thickness of the low-permittivity layer, are acknowledged uncertainties about the applicability of the input epsilon = 2.1, not circular imports of the conclusion. There is one minor self-citation, Ref. [11], an unpublished proposal of the plate geometry, but it is not load-bearing for the physics. The correlation-hole screening length s = rho_i^{-1/3} does introduce a self-consistency between the screening cutoff and the local density being solved for; this is a modeling approximation and a possible correctness risk, but it does not make the predicted concentration profile equivalent to an input by construction. The main vulnerability is external-evidence transferability: Ref. [14] measured confined water between graphene sheets, whereas the present mechanism assumes a roughly 1 nm low-permittivity layer at a single metal wall; this is a correctness risk, not a circularity.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

All inputs are either standard electrostatics or measured quantities from prior work; no new physical entities are introduced. The fragile inputs are the near-wall permittivity and the assumption that supersaturation leads to useful precipitation.

free parameters (4)

near-wall relative permittivity ε1/ε0 = 2.1
Taken from nanoconfined water measurements (Ref. 14) and used to set the near-wall Bjerrum length to 28 nm. The concentration enhancement scales strongly with this value.
thickness of low-permittivity layer h0 = 1 nm
Chosen from Ref. 14 as the location where the dielectric constant reverts to bulk water; it sets the position of the second image-potential peak.
distance of closest approach a = 0.3 nm for clean metal, 5 nm for oxide-coated metal
Used as the cutoff for the image potential; directly controls the maximum ion concentration near the wall.
dielectric transition width Δ = comparable to a water molecule, about 0.3 nm
Introduced ad hoc to regularize the dielectric jump and to set the peak positions h0 ± Δ for the second image potential.

assumptions (6)

standard math Standard image-charge electrostatics for a charge near a conducting wall and a dielectric interface, as in Jackson.
Equations (1), (3), (4), and (7)-(10) all rely on the classical image-charge method.
domain assumption Equilibrium balance between osmotic pressure and the deterministic image force gives a Boltzmann density profile.
Equation (2) equates the force from osmotic pressure to the image force, and integration yields Eqs. (5) and (6).
domain assumption Image forces between an ion and the image charges of other ions cancel because ion positions are random.
Stated in the text after Eq. (1): the attractive forces between an ion and image charges due to other ions cancel and can be neglected.
domain assumption Screening can be represented either by Debye-Hückel or by the Nordholm correlation-hole radius s = ρ_i^(-1/3).
Equations (3) and (4) present two alternative screening models, and both are used for quantitative claims.
domain assumption The low near-wall permittivity measured in nanoconfined water applies to water adjacent to a clean metal plate.
The value ε/ε0 = 2.1 from Ref. 14 is carried over to a metal-water interface without sensitivity analysis.
domain assumption Exceeding the solubility limit leads to precipitation fast enough and reliably enough for desalination.
The paper moves from 'concentration exceeds solubility' to 'salt will precipitate' without discussing nucleation barriers or induction times.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Desalination due to Electrical Image Forces." pith.science (2026). https://pith.science/paper/UTRNUQBW

@misc{pith2026190805359,
  author       = {Pith},
  title        = {Pith review of: Desalination due to Electrical Image Forces},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UTRNUQBW}},
  note         = {Machine review of arXiv:1908.05359}
}

read the original abstract

It will be shown that for a solution of salt dissolved in water in contact with a metallic wall, the concentration of salt ions (both positive and negative) within a few Angstroms of the wall can be large enough to exceed the solubility limit of the salt, as a result of electrical image charge forces. In addition, since the dielectric constant of water increases from 2.1 at the wall to 81 at about a nanometer from a solid wall, there will be an attractive image potential near the plane on which this increase of the dielectric constant occurs. The possible existence of these image potentials suggests that the salt can be removed from the water by making salt water flow between an array of parallel solid plates..

Figures

Figures reproduced from arXiv: 1908.05359 by the authors.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

18 extracted references · 18 canonical work pages

[14]

Fumagalli, A

L. Fumagalli, A. Esfandiar, R. Fabregas, S. Hu, P. Ares, A. Janardannan, Q. Yang, B. Radha, T. Taniguchi, K. Watanabe, G. Gomila, K. S. Novoselov and A. K. Geim, Science 360, 1339-1342 (2018)

work page 2018
[1]

Onsager and N

L. Onsager and N. N. T. Samarasa, J. Chem. Phys. 2, 528-536 (1933)

work page 1933
[2]

Markovich, D

T. Markovich, D. Andelman and R. Podgornick, Langmuir 33, 34-44 (2017)

work page 2017
[3]

Markovich, D

T. Markovich, D. Andelman and R. Podgornick, J. Chem. Phys. 142, 044702 (2015)

work page 2015
[4]

Wang and Z.-G

R. Wang and Z.-G. Wang, J. Chem. Phys. 142, 104705 (2015)

work page 2015
[5]

Kanduc and R

M. Kanduc and R. Podgornik, European Physical Journal E 23, 265-274 (2007)

work page 2007
[6]

Bakhshandeh, A

A. Bakhshandeh, A. P. dos Santos and Y. Levin, Phys. Rev. Lett. 107, 107801 (2011)

work page 2011
[7]

Buyukdagli, M

S. Buyukdagli, M. Manghi and J. Palmeri, Phs. Rev. E 81, 041601 (2010)

work page 2010

Show all 18 references

[8]

P. M. Biesheuvel, S. Porada, M. Levi and M. Z. Bazant, J. Solid State Electrochemistry 18, 1365-1376 (2014)

2014
[9]

Comtet, A

J. Comtet, A. Nigues, V. Kaiser, B. Coasne, L. Bocquet and A. Siria, Nature Materials 16, 634-639 (2017). 7

2017
[10]

A. A. Kornyshev, A. I. Rubinshtein and M. A. Vorotyntsev, Phys. Stat. Solidi (b) 84, 125 (1977); A. A. Kornyshev, M. A. Vorotyntsev, J. Phys. C:Solid State Phys. 11, L691 (1978); V. Kaiser, J. Comtet, A. Nigues, A. Siria, B. Coasne and L. Bocquet, Faraday Discussions 199, 129-...

1977
[11]

J. B. Sokoloff, unpublished
[12]

Classical Electrodynamics,

J. D. Jackson, “Classical Electrodynamics,” second edition (John Wiley and Sons, New York, 1975), pp55-56

1975
[13]

Nordholm, Chem

S. Nordholm, Chem. Phys. Lett. 105, 302-307 (1984)

1984
[15]

Hummer, J

G. Hummer, J. C. Rasaiah and J. P. Noworyta, Nature 414, 188-190 (2001)

2001
[16]

Majumder, N

M. Majumder, N. Chopra, N. Andrews, R. Andrews, B. J. Hinds, Nature 438,44 (2005)

2005
[17]

J. K. Holt, H. G. Park, Y. Wang, M. Stadermann, A. B. Artyukhin, C. P. Grigorop, Science 312, 1034-1037 (2006)

2006
[18]

J. M. Nielson, A. W. Adamson and J. W. Cobble, J. Am. Chem. Soc. 74, 446 (1952)

1952

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [14]

Fumagalli, A

L. Fumagalli, A. Esfandiar, R. Fabregas, S. Hu, P. Ares, A. Janardannan, Q. Yang, B. Radha, T. Taniguchi, K. Watanabe, G. Gomila, K. S. Novoselov and A. K. Geim, Science 360, 1339-1342 (2018)

work page 2018

[2] [1]

Onsager and N

L. Onsager and N. N. T. Samarasa, J. Chem. Phys. 2, 528-536 (1933)

work page 1933

[3] [2]

Markovich, D

T. Markovich, D. Andelman and R. Podgornick, Langmuir 33, 34-44 (2017)

work page 2017

[4] [3]

Markovich, D

T. Markovich, D. Andelman and R. Podgornick, J. Chem. Phys. 142, 044702 (2015)

work page 2015

[5] [4]

Wang and Z.-G

R. Wang and Z.-G. Wang, J. Chem. Phys. 142, 104705 (2015)

work page 2015

[6] [5]

Kanduc and R

M. Kanduc and R. Podgornik, European Physical Journal E 23, 265-274 (2007)

work page 2007

[7] [6]

Bakhshandeh, A

A. Bakhshandeh, A. P. dos Santos and Y. Levin, Phys. Rev. Lett. 107, 107801 (2011)

work page 2011

[8] [7]

Buyukdagli, M

S. Buyukdagli, M. Manghi and J. Palmeri, Phs. Rev. E 81, 041601 (2010)

work page 2010