When Water Learns to Queue: Layer-by-Layer Filling of Water in Nanoscale Capillaries

Forum topic · 二一 · 2026-05-01

Summary

A Chinese-language forum post on zhichai.net explains a new study (arXiv:2604.07946, Chen et al., University of Manchester) showing how water fills molecular-scale capillaries layer by layer. Building on Andre Geim's 2020 Nature paper confirming that the 150-year-old Kelvin equation still holds in 1-nanometer channels, the new work uses van der Waals-assembled capillaries with graphene spacers and atomic force microscopy (0.2 Å precision) to watch water enter in real time. In flexible capillaries (~20 nm-thick top walls), wall height rises in discrete steps of about 3 Å—exactly one molecular layer of water—revealing sequential, layer-by-layer filling up to roughly four layers before filling becomes continuous. In rigid capillaries (≥30 nm walls), filling occurs abruptly at a critical humidity with no intermediate states. The post explains this contrast through a competition between elastic deformation energy of the walls and oscillatory solvation forces (disjoining pressure), corroborated by free-energy calculations and molecular dynamics simulations. Applications span friction, adhesion, lubrication, corrosion, semiconductor lithography, drug delivery, and nanofluidic device design.

When Water Learns to Queue: Water Molecules' "Layer-by-Layer Climb" in Nanoscale Gaps

> *Original paper: Layer-by-layer water filling in molecular-scale capillaries (arXiv:2604.07946, Chen et al., Manchester)*

---

1. It Starts with a Sandcastle

Anyone who built sandcastles on the beach as a child knows the trick: dry sand is loose and useless, slipping through your fingers; add a little water, and the same sand can hold towers, tunnels, and walls.

This magical transformation comes from a phenomenon called capillary condensation. Water molecules spontaneously condense in the extremely narrow gaps between sand grains, forming tiny liquid bridges that glue the grains together. In 1871, the British physicist William Thomson—later Lord Kelvin—proposed an elegant equation describing this process: the narrower the capillary, the more easily water condenses inside it. This is the famous Kelvin equation, which predicts how thin a capillary must be before it fills with water at a given humidity.

For over 150 years, this equation has been textbook material. It works for millimeter-scale glass tubes, micron-scale soil pores, and even artificial channels tens of nanometers wide. But physicists have long asked: if a capillary is only a few water molecules wide—say 1 nanometer or less—does the Kelvin equation still hold?

Intuitively, the answer seems to be "no." At that scale, water is no longer a continuous fluid but a pile of discrete molecules. The very notion of a curved liquid surface loses meaning—you cannot draw a meniscus when there is room for only two or three molecules. Classical thermodynamics should fail here.

But Nature loves to defy expectations.

---

2. Geim's "Friday Night Experiments"

To tell this story, we must first meet an unconventional physicist: Andre Geim.

Geim is a professor at the University of Manchester and winner of the 2010 Nobel Prize in Physics—for isolating single-atom-thick graphene from graphite using adhesive tape, together with his student Novoselov. Less well known: Geim holds a unique record as the only person to win both a Nobel Prize and an Ig Nobel Prize. In 2000, he levitated a live frog with a strong magnet, earning the prize dedicated to research that "first makes people laugh, then makes them think."

Geim's lab has a tradition called "Friday Night Experiments." Every Friday evening, when others head home, the team stays in the lab doing "unserious" experiments—not for papers, not for grants, purely out of curiosity. The levitating frog was born this way. So was graphene.

In 2020, Geim and his postdoc Qian Yang turned that curiosity toward the 150-year-old question: does the Kelvin equation really work at the molecular scale?

---

3. Building the World's Smallest "Water Pipes"

Answering the question required making capillaries thin enough. Conventional lithography cannot do it—like carving flowers with a chisel, the tip has limits. The Geim team used their signature technique: van der Waals assembly.

In essence, they stacked atomically flat crystals of mica or graphite like a sandwich, using strips of graphene as "spacers." Graphene thickness can be controlled precisely: a monolayer is one carbon atom thick, about 0.34 nm. By choosing different numbers of graphene spacer layers, they fabricated flat capillaries ranging from 1 nm to tens of nanometers in height. The thinnest channel could hold only a single layer of water molecules—the smallest capillary physically possible.

They placed these nanocapillaries in a humidity-controlled chamber and monitored the deformation of the top wall with an atomic force microscope (AFM). The AFM tip acts like an exquisitely fine finger, sensing height changes with 0.2 Å precision—about one hundredth of a single atom's diameter.

The 2020 results surprised everyone: the Kelvin equation remained valid at the 1 nm scale, within acceptable error. Kelvin would likely have been both pleased and puzzled—he himself hinted in his paper that the equation could not extend to atomic scales, yet it did.

But that experiment told us only about the equilibrium state: whether the capillary ends up filled. It said nothing about how the water gets in.

---

4. Water's "Climb"

This is the question answered by the new (2026) paper.

The team designed nanocapillaries that can "breathe": the top wall is made of graphite crystal 20–40 nm thick—thin enough to bend measurably when water enters. They raised the humidity stepwise from 10% to 90%, waiting about two hours after each small increment for equilibrium, then recorded the top wall's height with AFM.

The results showed two distinct personalities.

The first is the flexible capillary (top wall ~20 nm thick). As humidity slowly rises, the wall height does not increase smoothly but in a series of distinct steps—roughly every 3 Å (0.3 nm), the wall suddenly rises, then holds steady at a plateau until humidity rises further and it jumps to the next step.

3 Å. This number is no coincidence—it is exactly the thickness of one molecular layer of water.

In other words, water molecules do not rush in all at once; like passengers queuing to board, they enter one layer at a time. The first layer goes in, pushing the top wall up by 3 Å; once the first layer is "settled," the second follows; then the third, the fourth... Only beyond about four layers do the steps blur and filling become continuous.

The second is the rigid capillary (top wall ≥30 nm thick). Here the picture is completely different: the wall barely deforms, and at some critical humidity the channel suddenly snaps full. No steps, no queue—water is either absent or entirely present.

---

5. A Tug-of-War Between Elastic Energy and Solvation Forces

Why do flexible walls make water "queue" while rigid walls make it "cut in line"?

The answer lies in a competition between two kinds of energy.

The first is the elastic deformation energy of the wall. Picture a thin plank spanning two walls, suspended in the middle. Inflate it from below and the plank bends upward—but you must overcome its stiffness. The thicker the plank, the harder it is to bend (elastic energy scales with the cube of thickness).

The second is the interaction between wall and water, technically called the solvation force or disjoining pressure. When water molecules are confined between two walls, they are forced into layered arrangements. This ordering produces an oscillatory pressure: at some separations the molecules sit comfortably and the pressure is attractive; at others they are squeezed too tightly and the pressure is repulsive. The oscillation's period corresponds exactly to one molecular layer.

With flexible walls, the wall is thin enough that its elastic restoring force is weak. The oscillatory solvation pressure dominates, pinning the wall at a series of discrete stable positions—corresponding to an integer number of water layers. Each added layer moves the system from one energy minimum to the next, like a ball rolling from one valley to the next in a wavy landscape.

With rigid walls, the wall is so thick that its elastic restoring force suppresses the solvation oscillations. The total energy curve has only one deep well: the system is either outside it (empty channel) or at its bottom (full channel), with no intermediate states. Water floods in at once, completing the transition.

The team verified this picture with theoretical models and molecular dynamics simulations. They computed total free-energy curves for different wall thicknesses: flexible walls show multiple clear minima corresponding to 1, 2, 3, and 4 water layers; rigid walls show a single minimum at the filled state. In simulations, water in flexible capillaries also enters layer by layer, matching the experiment perfectly.

---

6. From Sandcastles to Microchips

Why does this matter?

First, it reminds us that even in the most everyday phenomena, nanoscale physics can overturn macroscopic intuition. Water's behavior in narrow gaps governs friction, adhesion, lubrication, and corrosion—industrial problems costing hundreds of billions of dollars annually. In chip manufacturing, photoresist behavior in nanoscale trenches is governed by capillary condensation. In pharmaceuticals, drug release from porous carriers depends closely on how water fills the pores.

Second, it offers a precise lever for controlling water at the nanoscale: by changing wall flexibility, you can make water "queue" (layer-by-layer filling) or "cut in line" (abrupt filling). This could prove very useful in designing nanofluidic devices, bioinspired materials, or novel separation membranes.

Finally, it extends the legend of the Kelvin equation. 150 years ago, a Victorian physicist wrote an equation for millimeter-scale capillaries. 150 years later, we find it applies even to 1-nanometer channels—and the details of filling, layer-by-layer or abrupt, can be understood through classical energy competition. Good theories often outlive their supposed limits.

---

7. Epilogue: The Continuity of Science

Return to the sandcastle.

When you pour a bucket of water onto dry sand, you are—without realizing it—replaying a physical drama spanning 15 orders of magnitude. From micron-scale gaps between sand grains to atomically thin artificial capillaries, water molecules follow the same script: condense, arrange, form bridges in confined spaces. The only difference is that on the smallest stage we can finally see the actors' expressions—molecular discreteness is no longer averaged away but written directly into the data, in steps of 3 Å.

Geim once said something thought-provoking: "Good theories often transcend their limits of applicability." Perhaps the truer statement is that nature is more patient than we expect, willing to let old theories travel a little further—until, around the next corner, we find where they must be reinterpreted.

And science moves forward, step by step, at exactly these corners.

---

*References:*

Chen et al. (2026), arXiv:2604.07946
Yang et al. (2020), Nature, 588, 250-253 — validity of the Kelvin equation at the atomic scale
Geim & Novoselov (2004), Science, 306, 666-669 — mechanical exfoliation of graphene
Israelachvili (2011), Intermolecular and Surface Forces — classic text on solvation forces
Geim & Berry (2000), Eur. J. Phys., 18, 307-313 — "Of flying frogs and levitrons"