1.1 What a liquid is

A gas is a population of molecules in nearly-free flight, encountering each other rarely enough that the intermolecular potential between any given pair is negligible for almost all of each molecule’s history. The kinetic theory of Sound 1.2 refresher → is built on this picture, and it gives the equation of state $p V = N k_B T$ that underwrites the linear acoustics of the sound book.

A solid is a population of molecules locked into a periodic lattice by an intermolecular potential whose well depth is large compared to $k_B T$. Each molecule oscillates within its potential minimum but rarely escapes; long-range order persists.

A liquid sits in the awkward middle. The intermolecular potential is the same as the solid’s — water in its liquid and solid phases is held together by the same hydrogen bonds, with the same length scale and the same well depth of a few tenths of an electron-volt — but the thermal energy is now comparable to the well depth and the molecules wander. There is no long-range order. The mean distance between molecular centres is only a few percent larger than in the solid (1.0 g/mL liquid water vs 0.917 g/mL ice). Molecules are in soft contact with their neighbours: too close to be free, too thermally agitated to be locked.

The interaction potential

The pairwise potential between two molecules at separation $r$ is well-approximated, for noble gases and simple polar liquids, by the Lennard-Jones 12-6 form:

$U(r) \;=\; 4 \varepsilon \left[\left(\frac{\sigma_0}{r}\right)^{12} - \left(\frac{\sigma_0}{r}\right)^{6}\right],$

with two parameters: the well depth $\varepsilon$ (water: about $0.65 k_B T_\text{room}$ per hydrogen-bond pair, summed over coordination gives a few $k_B T$ per molecule) and the equilibrium separation $\sigma_0$ (water: about 2.8 Å molecule-centre to molecule-centre). The $r^{-12}$ term is a phenomenological short-range repulsion; the $r^{-6}$ term is the long-range van der Waals attraction.

At the equilibrium separation $r_\text{eq} = 2^{1/6} \sigma_0$, $U(r) = -\varepsilon$ is at its minimum and the force on each molecule from its neighbour vanishes. A small displacement either way brings a restoring force — pull the molecules apart and they attract; push them together and they repel. The liquid state, viewed in this potential, is the regime where molecules sit near $r_\text{eq}$ on average but each molecule is randomly hopping between neighbouring potential minima created by its varying local environment.

Why a liquid resists compression

Push a liquid (raise its pressure) and the average separation $r$ decreases below $r_\text{eq}$. The $r^{-12}$ repulsion grows extremely steeply — a 1% reduction in $r$ raises $U_\text{rep}$ by a factor of $\sim 1.12$. The required pressure to achieve even a modest density increase is therefore large; water’s bulk modulus

$K = -V \left.\frac{\partial p}{\partial V}\right|_T$

is about 2.2 GPa = 22,000 atm at room temperature. To compress water by 1% requires roughly 220 atm. This is why ocean pressure increases mostly with depth (gravity-driven) rather than with compression: even at the 11 km depth of the Mariana Trench, the seawater is only about 5% denser than at the surface.

Why a liquid resists tension

The same potential, run in the other direction, tells us about tension. Reduce the pressure below atmospheric and the liquid expands; the average $r$ grows; the molecules feel each other’s attractive force pulling them back together. At small expansion, $U(r)$ rises symmetrically with the compression case — the bulk modulus is the same for small tension as for small pressure — so tension up to a few hundred atm produces only a few percent strain.

Push harder. As $r$ grows past $r_\text{eq}$ the force $-\partial U / \partial r$ increases, reaches a maximum, and then decreases as $r$ continues to grow toward the inflection point of $U$. The maximum attractive force corresponds to the maximum tension the liquid can resist before any pair of molecules in soft contact loses contact entirely. Beyond that strain, the liquid is unstable: any small fluctuation grows, the local separation widens past the inflection, and the molecules cease to interact.

This is the theoretical cohesive limit — the maximum tension a perfectly homogeneous liquid could withstand before the intermolecular potential’s attractive arm gives out. The next lesson estimates its size.

Why the continuum picture is not enough

The cohesive-limit argument assumes that the liquid is homogeneous all the way down — that every cubic micron looks the same as every other cubic micron, that the intermolecular potential acts identically everywhere. This is a fiction. A real liquid contains:

• Dissolved gas. Atmospheric water at room temperature holds about 17 mg of dissolved air per litre — millions of dissolved gas molecules per cubic micron.
• Particulate impurities. Dust, ions, surface-active molecules, fragments of organic matter.
• Gas microbubbles trapped in solid surfaces. Any wettable container has microscopic crevices in its walls whose geometry can host a long-lived gas pocket below atmospheric pressure.
• Cosmic-ray-induced events. A single fast neutron or muon traversing the liquid leaves a brief track of intense local heating — a thermodynamic disturbance that can nucleate a bubble in a metastable liquid.

Each of these is a nucleation site — a place where the liquid is preferentially weak. The measured tensile strength of any real sample reflects whichever site is weakest, not the homogeneous cohesive limit.

The chapter develops this gap. The next lesson estimates the cohesive limit from the Lennard-Jones potential. The lesson after that surveys what real experiments find, and shows the gap to be three to four orders of magnitude wide — which Chapter 2 then resolves.