Slow water in engineered nano-channels revealed by color-center-enabled sensing

Abraham Wolcott; Ankit Bhardwaj; Boya Radha; Carlos A. Meriles; Daniela Pagliero; Gustavo L\'opez; Kang Xu; Kapila Elkaduwe; Nicolas Giovambattista; Rohma Khan

REVIEW 4 major objections 3 minor 1 references

Slow water in engineered nano-channels revealed by color-center-enabled sensing

T0 review · 4 major / 3 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Using shallow nitrogen-vacancy color centers as proton NMR sensors, this paper measures water self-diffusivity in 5.6-nm channels that is orders of magnitude below bulk water and attributes the slowdown to photogenerated charge…

desk verdict First NV-based proton NMR of water in engineered nanochannels, but the headline diffusivities are compromised because the paper ignores the hyperfine decoherence it itself invokes. read the letter →

arxiv 2412.18969 v1 pith:2K7DCIB4 submitted 2024-12-25 cond-mat.mes-hall

Daniela Pagliero , Rohma Khan , Kapila Elkaduwe , Ankit Bhardwaj , Kang Xu , Abraham Wolcott , Gustavo López , Boya Radha

show 2 more authors

Nicolas Giovambattista Carlos A. Meriles

This is my paper · ORCID

classification cond-mat.mes-hall

keywords nitrogen-vacancycentersnanoconfinedwaternuclearmagneticresonanceself-diffusionnanofluidicshexagonalboronnitridechannelsinterfacialopticallydetected

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

Water squeezed into 5.6-nm-tall channels between a diamond crystal and a hexagonal boron nitride flake moves far more slowly than bulk water. The paper measures this by using a roughly 8-nm-deep layer of nitrogen-vacancy (NV) color centers in the diamond as optically detected magnetic sensors: proton NMR correlation signals from the confined water decay with a fast time constant of about 0.5 µs and a slow one of about 30 µs, which translate to room-temperature self-diffusivities of roughly $2\times10^{-11}$ m²/s and $4\times10^{-13}$ m²/s, compared with about $2.3\times10^{-9}$ m²/s for bulk water. The authors argue that spatial confinement alone cannot explain the slowdown, because their molecular dynamics simulations of pure water in the same geometry show bulk-like diffusion. Instead, they attribute it to photogenerated carriers accumulating at the diamond-water and hBN-water interfaces, forming an electric double layer that pins down the first water layer; simulations with added anions reproduce a nearly immobile first layer. The work matters because it demonstrates a non-invasive, local probe of molecular dynamics inside sealed nanofluidic channels at ambient conditions.

What carries the argument

The experiment rests on nitrogen-vacancy (NV) centers—atomic-scale paramagnetic defects in diamond whose spin state can be initialized, manipulated, and read out with light and microwaves—working as optically detected magnetometers in an ~8-nm-deep layer. Proton NMR is acquired with dynamical decoupling pulse trains (XY8-N) that make the NV sensitive to the ~1.9 MHz Larmor precession of 1H spins in the channel. To measure diffusion, the paper uses correlation spectroscopy: two encoding and decoding blocks separated by a variable time $\tilde{\tau}$, where the nuclear spin correlation signal $C(\tilde{\tau})$ picks up a decay envelope $F(\tilde{\tau}/T_D)$ set by molecular self-diffusion out of the sensing volume. The diffusivity follows from $D = d^2/(6T_D)$ with $d \approx 8$ nm, and a two-layer envelope with $T_D = 0.5$ µs and 30 µs yields the fast and slow components. On the modeling side, classical molecular dynamics with a rigid water model and added fluoride anions as stand-ins for injected charge shows a pinned first water layer, providing the mechanistic explanation.

What would settle it

Re-run the correlation measurement with the laser off during the variable delay interval, or with the diamond electrically grounded so photogenerated carriers can drain away; if the slow decay component and the ~2 MHz hyperfine shoulders persist unchanged, the charge-accumulation explanation is wrong, whereas if the diffusivity rises toward the bulk value, the carrier mechanism is confirmed. A second check: vary the NV depth $d$; the relation $D = d^2/(6T_D)$ predicts $T_D$ should scale as $d^2$, and a mismatch would signal that decoherence other than diffusion shapes the envelope.

Watch

Extended reading notes

Core claim

The paper's central claim is that nitrogen-vacancy (NV) color centers, engineered about 8 nm below a diamond surface and read out optically, can detect proton NMR from room-temperature water confined in 5.6-nm-tall channels, and that the measured nuclear spin correlation decay implies water self-diffusivity orders of magnitude below bulk: a fast population with $D \approx 2\times10^{-11}$ m²/s and a slow population with $D \approx 4\times10^{-13}$ m²/s, and in some runs $D \leq 1\times10^{-13}$ m²/s. The authors further claim that pure confinement is not enough to produce such slow motion, since their classical molecular dynamics simulations of pure water in the same channel show bulk-like diffusion, and they propose instead that illumination photogenerates carriers that accumulate at the interfaces, forming an electric double layer that immobilizes the first water monolayer. Simulations replacing some water with anions reproduce the immobile first layer, supporting the charge-based mechanism. The paper also reports an asymmetric proton spectrum with roughly 2 MHz broadening that it attributes to hyperfine coupling between protons and carrier spins, the same interaction that could complicate the diffusion readout.

Load-bearing premise

The load-bearing premise is that the measured proton correlation decay is caused purely by molecular self-diffusion, converted to a diffusivity through $D = d^2/(6T_D)$ with $d \approx 8$ nm; if other decoherence (for example, hyperfine coupling to the photogenerated carriers the paper itself invokes) shortens the envelope, the reported D values are lower bounds, not the true self-diffusivities.

Editorial extensions

If this is right

NV-based NMR can measure molecular diffusion in sealed, room-temperature nanofluidic channels without invasive probes.
Water in ~6-nm confinement can be orders of magnitude slower than bulk, so fluidic transport models that neglect interfacial charge will underestimate residence times.
The observed slowdown is tied to illumination, implying optical control of carrier injection could tune water mobility in nanoscale devices.
Proton NMR with color centers can be extended to other confined fluids, ions, and possibly biomolecular hydration layers.
The ~2 MHz spectral broadening suggests hyperfine coupling to carriers that may itself be used to sense interfacial charge density.

Reading between the lines

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

If the decay envelope is partially set by hyperfine decoherence rather than pure diffusion, the true diffusivities could be even lower than reported, strengthening the 'slow water' conclusion while weakening the quantitative D values.
The illumination dependence of the mechanism is testable: sweeping laser power or grounding the diamond should modulate the slow component if photogenerated carriers are the cause.
The two-layer model implies the slow population sits near the walls; varying channel height should change the fast/slow weight ratio, mapping the spatial extent of the arrested layer.
The use of fluoride anions as electron surrogates leaves open the microscopic identity of the charge carrier; an ab initio treatment of a solvated electron in the channel would be the decisive modeling step.

Share X Bluesky LinkedIn Reddit HN

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

First, the headline: this is the first NV-based proton NMR of water in engineered nanochannels, and the observation of long-lived correlation signals does indicate strongly suppressed water mobility. But the specific diffusivities (D_fast ~ 2e-11, D_slow ~ 4e-13 m2/s) are not as clean as the abstract implies. The correlation envelope is fit assuming purely diffusive decay, yet the paper itself argues that hyperfine coupling to photogenerated carriers broadens the proton spectrum by ~2 MHz. That coupling will also shorten the correlation envelope, so the fast component (T_D = 0.5 us) may be contaminated. The paper never puts the hyperfine term into the correlation analysis, nor shows a control varying carrier density. This is a real internal inconsistency, and it means the D values should be read as lower bounds at best (or as mixed diffusion plus decoherence times), not as clean self-diffusion coefficients.

What is genuinely good: the device fabrication and the use of shallow NVs in a sealed channel is a credible extension of existing NV-NMR techniques. The long tail at tens of microseconds is not something you would see with bulk-like water, so the qualitative slowdown is probably real. The MD simulations are thorough: four surface terminations, fluoride concentration and size checks, and a pure-water baseline showing bulk-like diffusion. That gives a plausible physical story for interfacial charge immobilizing the first water layer. The paper also honestly flags run-to-run variability and the crude nature of the hyperfine model.

Soft spots, in order: (1) no error bars or repeated fits for the two-layer model; (2) the detection radius d = 8 nm is used as a hard number in D = d^2/(6T_D) though the sensing volume is not a sharp sphere; (3) the fluoride electron analog is classical and the authors know it; (4) the hyperfine-diffusion entanglement is the biggest problem and is not addressed quantitatively. A referee should ask for a joint fit including a distribution of hyperfine couplings, or a control measurement at different laser intensities, to separate the two decay sources.

Who this is for: people working in NV magnetometry, nanofluidics, and interfacial water. It deserves a serious referee round: the idea is novel and the artifact is, in principle, addressable. I would send it to review, but with a strong request to fix the correlation analysis and provide statistics.

Referee Report

4 major / 3 minor

Summary. The manuscript reports nitrogen-vacancy (NV) center-enabled 1H NMR of water confined to ~5.6-nm-tall, 185-nm-wide channels formed between a diamond substrate and a suspended hBN flake. Using XY8 dynamical-decoupling sequences and correlation spectroscopy, the authors detect proton spin signals whose correlation envelope they fit to a two-component decay with times T_D^(1)=0.5 us and T_D^(3)=30 us (weights 0.8 and 0.2); converting via D=d^2/(6T_D) with d=8 nm gives D_fast ~ 2.1x10^-11 m2/s and D_slow ~ 0.4x10^-12 m2/s, orders of magnitude below bulk water. They propose that photogenerated carriers injected into the water slow its dynamics and support this with classical MD simulations in which fluoride anions proxy for solvated electrons. A later correlation measurement shows coherences exceeding 25-100 us, implying D <= 0.1x10^-12 m2/s.

Significance. The paper demonstrates, seemingly for the first time, NV-enabled 1H NMR of water in an engineered nanofluidic channel, and the qualitative observation of proton coherence persisting for tens of microseconds in room-temperature confined water is technically novel and impressive: it requires shallow NV ensembles in a 13C-depleted layer, nanofabricated hBN spacers, and correlation protocols with stringent stability requirements. The MD study across four surface terminations and several charge configurations is also a useful contribution, and the result that uncharged confined water retains bulk-like diffusivity sharpens the case for an electrostatic slowdown mechanism. However, the central quantitative claim, namely the bimodal diffusivity with D_fast ~ 2x10^-11 m2/s and D_slow ~ 4x10^-13 m2/s, is not yet established to the standard the conclusions require: the correlation-envelope analysis omits the hyperfine interactions that the same paper invokes to explain the spectral broadening of the same data, and the fitted parameters are presented without any uncertainty quantification.

major comments (4)

[SM Section II.1; main text Fig. 3c] There is an internal inconsistency between the diffusion analysis and the hyperfine analysis of the same correlation data. The envelope fit in SM Section II.1 assumes that 'molecular self-diffusion governs the overall decay' (main text) and that the envelope is a function of tau~/T_D only, as written in SM Eq. (1). Yet SM Section II.2 interprets the Fourier transform of the very same correlation signal as dominated by hyperfine couplings spanning roughly 2 MHz between water protons and photogenerated carrier electron spins. A distribution of hyperfine couplings of this magnitude necessarily contributes an additional, diffusion-independent decay channel to the tau~ envelope, both through the inhomogeneous spread of resonance frequencies and through fluctuating hyperfine fields set by carrier-spin dynamics; the corresponding dephasing rate is comparable to the inverse of the fast fitted time constant (T_D^(1)=0.5 us corresponds to about 0.3 MHz). The bimodal fit therefore cannot uniquely determine T_D, and D=d^2/(6T_D) would overestimate the true self-diffusivity if hyperfine dephasing is active. The authors' own caveat in SM Section II.2 (hydrated-electron clusters 'diffuse rapidly' and conflict with the observations unless long-lived structures are imposed) concedes the tension. To make the headline D values load-bearing, the correlation model should incorporate the hyperfine-shifted components already used to reproduce the spectrum, and a control measurement such as a green-laser-power series should demonstrate that the envelope is insensitive to carrier density. Without this, the quantitative claim D_fast ~ 2.1x10^-11 m2/s and D_slow ~ 0.4x10^-12 m2/s is not established, though the qualitative existence of a very slowly decaying component (tens of microseconds) is more robust.
[SM Section II.1] The reported diffusivities are extracted from a single representative correlation trace (Fig. S5) with a bimodal fit (alpha=0.8, beta=0.2, T_D^(1)=0.5 us, T_D^(3)=30 us) presented without error bars, confidence intervals, run-to-run statistics, or a discussion of parameter degeneracy. The main text acknowledges 'substantial variability in the evolution of the spectra after channel filling' (Fig. 3c), and the later measurement yields T_D >= 100 us, i.e., a third, even slower time constant. Given this variability, the precision and representativeness of the two quoted D values across measurement rounds, fill cycles, and channel sites are not established. The manuscript should report statistics over multiple correlation traces and, if possible, over multiple sites within the channel, and should state how the reported values relate to the observed spread.
[SM Section II.1] The conversion from correlation time to diffusivity uses d=8 nm (assumed equal to the NV depth) as the detection radius, and D scales as d^2, so a plausible uncertainty of +/-2 nm in d changes D by roughly +/-50%. In addition, the functional form F_{3/2}(z) in SM Eq. (2) is derived for a half-space geometry, whereas the actual sensing volume is truncated by the 5.6-nm channel ceiling and the NV layer has a finite depth distribution. The paper should justify the single value d=8 nm, provide the NV depth distribution, and estimate the systematic error in D arising from the geometry mismatch and from the choice of d.
[SM Section III] The MD support for the charge-injection mechanism replaces roughly 2700 to 5400 of about 30,000 water molecules with fluoride anions, which corresponds to a concentrated electrolyte (about 9-18 mol%), and the mapping between this simulated charge concentration and the actual photogenerated carrier density in the experiment is not established. The simulations therefore demonstrate the qualitative viability of an electrostatic slowdown mechanism, but they cannot currently anchor the magnitude of the slowing to the experimental conditions. The authors acknowledge the crudeness of the proxy, but the conclusions should state explicitly that the MD provides qualitative, not quantitative, support for the proposed mechanism, and the paper should either estimate the experimental carrier density or refrain from implying a quantitative comparison.

minor comments (3)

[SM Section II.1] The bulk self-diffusion constant of water is quoted as 2.3x10^-4 m2/s; the standard value (used implicitly in the main text) is 2.3x10^-9 m2/s, so the SM value is off by five orders of magnitude and should be corrected.
[SM Section I.5] The correlation protocol employs two XY8 blocks separated by a free interval tau~, but the manuscript does not state whether NV spin coherence during tau~ is refocused (e.g., by a central pi pulse) or how the measured NV coherence (XY8-5: 19.8 us, SM Section I.4) constrains the accessible tau~ range. Since the claim of 'no apparent decay' over 25 us in Fig. 3c (left panel) is important for excluding instrumental decay, the sequence and the relevant NV-coherence characterization should be described explicitly.
[SM throughout] The Supplementary Material contains several typographical and formatting issues that will hinder readers: the section heading in Section III.5 is numbered 'I.5' instead of 'III.5', Fig. S17's caption reads 'bollom' for 'bottom', and several equations (SM Eq. (1) in Section II.1 and Eqs. (1)-(4) in Section II.2) are garbled in the rendering. A careful proofreading pass is needed.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the diffusivity is extracted by fitting an independently derived correlation envelope, and the MD simulation is an independent check of a hypothesized mechanism.

full rationale

The paper's central quantitative claim is an experimental measurement, not a prediction from a fitted model. The correlation envelope is fit to the independently derived F_{3/2}(τ̃/T_D) from Ref. 8 (Staudenmaier et al.) with two diffusion times, and D is obtained from D = d^2/(6T_D) with d = 8 nm; this is a standard parameter extraction, not a reduction of the output to the input. The MD simulations test a hypothesized charge-injection mechanism by adding F− anions and observing slowed first-layer water; that is a separate, falsifiable computation, not an echo of the experimental fit. Self-citations (Refs. 19, 21, 22, 27) supply methodology and prior demonstrations of shallow-NV NMR and charge injection; the charge-injection premise is also supported by non-self references 26, 32, and 33, and the slow-water measurement does not depend on that mechanism. The concern that hyperfine coupling to photogenerated carriers could contaminate the correlation decay is a validity or modeling-omission issue; the paper itself cautions in SM Section II.2 that the hyperfine assignments are only a 'crude guide' with ambiguity. This is a legitimate scientific caveat, but it is not circularity: no equation in the paper defines T_D in terms of the hyperfine parameters or makes the reported D equal to a fitted input by construction.

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

All quantitative conclusions rest on fitted diffusion times, an assumed detection radius, a classical proxy for electrons, and the pre-existing NV sensing framework. No new fundamental entity is postulated; the main unverified premise is the photo-carrier mechanism and its electrostatic effect.

free parameters (6)

Fast diffusion time T_D^(1) = 0.5 microseconds
Fit to the initial decay of the 1H correlation envelope in SM Section II.1; central to deriving D_fast approximately 2x10^-11 m2/s.
Slow diffusion time T_D^(3) = 30 microseconds
Fit to the long-lived tail of the correlation signal; central to deriving D_slow approximately 4x10^-13 m2/s.
Two-layer weights alpha and beta = 0.8 and 0.2
Chosen to reproduce the bimodal envelope in SM Fig. S5b.
Detection radius d = 8 nm
Assumed equal to NV depth; converts T_D to D via d^2/(6 T_D). Uncertain by a factor that could shift D substantially.
Hyperfine coupling constants A_i = A_i^2 approximately 2 MHz^2, linewidth 0.1 MHz
Hand-selected five couplings to reproduce the asymmetric 1H spectrum in SM Section II.2.
Number of fluoride anions N_F in MD = 2704 (and 5408 in sensitivity test)
Ad hoc choice to model injected electrons in MD; actual carrier density in the experiment is unknown.

assumptions (4)

domain assumption The 1H correlation decay is governed by molecular self-diffusion, with negligible internuclear dipolar and hyperfine contributions to the envelope.
SM Eq. (1) and main text 'molecular self-diffusion governs the overall decay'; the paper later invokes hyperfine broadening, creating tension.
ad hoc to paper Classical MD with fluoride anions can represent the electrostatic effect of solvated electrons on water dynamics.
SM Section III and main text 'we ignore chemical transformations and use instead classical MD simulations'; electron quantum effects are acknowledged as missing.
domain assumption Green illumination photo-generates carriers that accumulate at the diamond-water and hBN-water interfaces and inject into the fluid.
Main text 'Charge injection from shallow color centers' is based on prior citations, not measured directly in this device.
standard math Standard force fields (TIP4P/2005, Madrid2019, CHARMM) and NV sensing models are valid for this system.
Used throughout SM Section III; standard in the field.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Slow water in engineered nano-channels revealed by color-center-enabled sensing." pith.science (2026). https://pith.science/paper/2K7DCIB4

@misc{pith2026241218969,
  author       = {Pith},
  title        = {Pith review of: Slow water in engineered nano-channels revealed by color-center-enabled sensing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2K7DCIB4}},
  note         = {Machine review of arXiv:2412.18969}
}

read the original abstract

Nanoscale confinement of molecules in a fluid can result in enhanced viscosity, local fluidic order, or collective motion. Confinement also affects ion transport and/or the rate and equilibrium concentration in a chemical reaction, all of which makes it the subject of broad interest. Studying these effects, however, is notoriously difficult, mainly due to the lack of experimental methods with the required sensitivity and spatial or time resolution. Here we leverage shallow nitrogen-vacancy (NV) centers in diamond to probe the dynamics of room-temperature water molecules entrapped within ~6-nm-tall channels formed between the diamond crystal and a suspended hexagonal boron nitride (hBN) flake. NV-enabled nuclear magnetic resonance measurements of confined water protons reveal a much reduced H2O self-diffusivity, orders of magnitude lower than in bulk water. We posit the slow dynamics stem from the accumulation of photogenerated carriers at the interface and trapped fluid, a notion we support with the help of molecular dynamics modeling. Our results provide feedback for theories describing interfacial water, and lay out a route for investigating other fluids under confinement.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

1 extracted references · 1 canonical work pages

[1]

Fabrication of angstrom-scale two-dimensional channels for mass transport

1 A. Bhardwaj, M.V.S. Martins, Y. You, R. Sajja, M. Rimmer , S. Goutham, R. Qi, S.A. Dar, B. Radha, A. Keerthi, “Fabrication of angstrom-scale two-dimensional channels for mass transport”, Nat. Prot. 19, 240 (2024). 2 R. Frisenda, E. Navarro -Moratalla, P. Gant, D. Pérez De Lara, P. Jarillo -Herrero, R.V. Gorbachev, A. Castellanos - Gomez, “Recent progres...

work page 2024

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [1]

Fabrication of angstrom-scale two-dimensional channels for mass transport

1 A. Bhardwaj, M.V.S. Martins, Y. You, R. Sajja, M. Rimmer , S. Goutham, R. Qi, S.A. Dar, B. Radha, A. Keerthi, “Fabrication of angstrom-scale two-dimensional channels for mass transport”, Nat. Prot. 19, 240 (2024). 2 R. Frisenda, E. Navarro -Moratalla, P. Gant, D. Pérez De Lara, P. Jarillo -Herrero, R.V. Gorbachev, A. Castellanos - Gomez, “Recent progres...

work page 2024