Surface Tension and Surface Free Energy

Blow a soap bubble and watch it snap into a perfect sphere. Something is pulling that film taut. That same something lets water striders walk on ponds. That "something" is surface tension. It is one of the oldest quantities in physical chemistry measured precisely. It is also one of the most useful in modern surface science. Surface tension matters because interfaces are everywhere. Raindrops, detergents, lung alveoli, inkjet printing, oil recovery, and protein folding all depend on it. Each hinges on the energetics of molecules sitting at a boundary. Understanding it opens the door to curved-interface pressure, wetting, and nucleation phenomena. This article builds surface tension from molecular cohesion up to the thermodynamic definition. You will see why force per length and energy per area are the same quantity. Then you will derive the reversible work of creating surface. Finally you will apply it to two realistic numerical problems.

Key takeaways

Surface tension $\gamma$ is the force per unit length along an interface, or equivalently the Gibbs energy per unit area needed to create it.
It arises because molecules at a surface have fewer neighbors, hence higher potential energy, than molecules in the bulk.
Units: N m⁻¹ = J m⁻² ; water at 20 °C has $\gamma \approx 72.8\ \text{mN m}^{-1}$ .
Formally, $\gamma = (\partial G/\partial A)_{T,p,n}$ , linking surface tension directly to Gibbs energy.
Reversible surface creation costs work $dw = \gamma\, dA$ , the basis of the Wilhelmy plate and capillary-rise methods.
Key idea: Surface tension is the thermodynamic price, in energy per area, of exposing molecules to an unbalanced, attractive-force-poor environment at an interface.

The core insight

Surface tension is not a separate force so much as the macroscopic bookkeeping of a molecular energy deficit: molecules at an interface lack neighbors on one side, so pulling more molecules to the surface (increasing area) costs reversible work $\gamma\,dA$ , which is why the same number $\gamma$ can be measured either as a mechanical force per length or as a Gibbs energy per area.

The physical picture

Consider a liquid in equilibrium with its vapor. A molecule deep in the bulk is surrounded on all sides by neighbors, and the attractive intermolecular forces — van der Waals, dipolar, or hydrogen-bonding — pull equally in every direction. The net force on that molecule averages to zero, and its potential energy is at a comfortable minimum set by the full coordination shell of neighbors. A molecule sitting at the liquid surface is missing roughly half of its neighbors, since the vapor phase above provides far weaker attractive interactions than the dense liquid below. This molecule experiences a net inward pull and, more importantly, has a higher potential energy than its bulk counterpart, precisely because fewer favorable attractive contacts are satisfied. The system therefore always tends to minimize the number of molecules exposed at the surface, which is exactly why liquid droplets adopt spherical shapes: a sphere minimizes surface area for a given volume. To increase the surface area — say, by stretching a soap film or blowing a bubble — molecules must be moved from the energetically comfortable bulk to the energetically costly surface. This costs work. The work required per unit area of new surface created, at constant temperature and pressure, defines the surface tension $\gamma$ . Because work per area has the same dimensions as force per length (J m⁻² = N m⁻¹), $\gamma$ can equivalently be pictured as a tension pulling along any line drawn in the surface, trying to contract it, like an infinitesimally thin elastic skin under uniform stress. This dual identity — mechanical tension and thermodynamic excess energy — is what makes surface tension such a versatile quantity, measurable by pulling a wire frame, by watching a drop detach, or by measuring how high liquid climbs a capillary tube.

Figure

Molecules at a liquid surface have fewer attractive neighbors than those in the bulk

Figure

Force-per-length and energy-per-area pictures of γ are numerically identical

The core relations

\gamma = \left(\frac{\partial G}{\partial A}\right)_{T,p,n}

where: $\gamma$ is surface tension (J m⁻²), $G$ is Gibbs energy, $A$ is interfacial area, held at constant temperature, pressure, and composition.

dw_{\text{rev}} = \gamma\, dA

where: $dw_{\text{rev}}$ is the reversible non-expansion work to create area $dA$ of new surface.

F = \gamma\, l

where: $F$ is the force needed to hold a movable boundary of length $l$ on a stretched film (a soap film has two surfaces, so $F = 2\gamma l$ ).

\gamma(T) \approx \gamma_0\left(1 - \frac{T}{T_c}\right)^{n}

where: $\gamma_0$ is a substance-specific constant, $T_c$ is the critical temperature, and $n \approx 1.2$ (an empirical Eötvös/Guggenheim-type relation showing $\gamma \to 0$ at the critical point).

Derivation AFrom molecular cohesion to

\gamma = (\partial G/\partial A)_{T,p}

Goal: show rigorously that the reversible work of creating new surface area equals $\gamma\,dA$ , and that this identifies $\gamma$ as a Gibbs energy per unit area.

Step 1. Set up the system. Consider a liquid film held in a rectangular wire frame, one side of which is a frictionless movable bar of length $l$ . The film has two faces (front and back), each an interface with the vapor. Pulling the bar outward by $dx$ increases the total film area by $dA = 2l\,dx$ , exposing new liquid surface that must be populated by molecules drawn from the bulk.

Step 2. Identify the force. Experimentally, a well-defined force $F$ is needed to hold the bar in equilibrium against the film's tendency to contract. This force does not depend on how far the bar has already moved, only on the length $l$ and the nature of the two liquid–vapor interfaces, so we write

F = \gamma_{\text{surface}}\cdot 2l ,

where $\gamma_{\text{surface}}$ is a proportionality constant with units of force per length — this is the operational definition of surface tension.

Step 3. Compute the reversible work. Moving the bar quasi-statically through $dx$ against this force does work on the film equal to

dw_{\text{rev}} = F\,dx = \gamma_{\text{surface}}\,(2l\,dx) = \gamma_{\text{surface}}\,dA .

Because the process is carried out reversibly at constant $T$ and $p$ , and no other non-expansion work is done, this mechanical work is exactly the increase in the system's Gibbs energy:

dG = dw_{\text{rev}} = \gamma_{\text{surface}}\,dA .

Step 4. Take the thermodynamic derivative. Rearranging the last line at constant temperature, pressure, and amount of substance,

\left(\frac{\partial G}{\partial A}\right)_{T,p,n} = \gamma_{\text{surface}} .

This shows the mechanical constant measured by pulling the bar is identical to the thermodynamic surface excess Gibbs energy per unit area. There is only one $\gamma$ , not two coincidentally equal numbers.

Step 5. Connect back to molecular cohesion. Because surface molecules have fewer attractive neighbors (Step in the physical picture above), the molar Gibbs energy of a surface molecule exceeds that of a bulk molecule by an amount that, integrated over the surface excess of molecules per unit area, gives precisely this $\gamma$ . Stronger cohesive forces (e.g., hydrogen bonding in water, metallic bonding in liquid mercury) produce a larger energy deficit at the surface and hence a larger $\gamma$ .

Boxed result:

\boxed{\;\gamma = \left(\frac{\partial G}{\partial A}\right)_{T,p,n} = \frac{F}{2l}\;}

Worked examples

Example 1 — Wilhelmy plate: measuring $\gamma$ directly (foundational). A thin glass plate of width $w = 2.00\ \text{cm}$ and negligible thickness is suspended from a balance so its bottom edge just touches the surface of water at 20 °C. The liquid wets the glass completely (contact angle $\theta = 0°$ ), forming a meniscus around the plate's perimeter. The balance records an extra downward force of $F_{\text{extra}} = 2.91\times10^{-3}\ \text{N}$ due to the surface tension pulling the plate down, beyond the plate's own weight (already tared out). The wetted perimeter is the two long faces of the plate, so

p = 2w = 2(0.0200\ \text{m}) = 0.0400\ \text{m}.

Surface tension pulls along this whole perimeter, and since $\theta = 0$ , the vertical component equals the full tension:

F_{\text{extra}} = \gamma\, p \cos\theta = \gamma\, p .

Solving,

\gamma = \frac{F_{\text{extra}}}{p} = \frac{2.91\times10^{-3}\ \text{N}}{0.0400\ \text{m}} = 7.28\times10^{-2}\ \text{N m}^{-1} = 72.8\ \text{mN m}^{-1}.

This matches the accepted literature value for pure water at 20 °C (72.8 mN m⁻¹) to three significant figures — a good sanity check that the plate was clean and fully wetted, since any surfactant contamination would have lowered this number substantially (soap solutions run 25–40 mN m⁻¹).

Example 2 — Work to atomize a raindrop into a fog of micro-droplets (applied, harder). A single spherical raindrop of radius $R = 1.00\ \text{mm}$ falls and shatters into $N = 10^{6}$ identical smaller droplets. Assume no mass is lost and the water stays at 20 °C, where $\gamma = 72.8\times10^{-3}\ \text{N m}^{-1}$ . Find the minimum (reversible, isothermal) work needed to create this fine mist, and the resulting temperature-equivalent thought experiment of where that energy comes from.

Step 1 — volume conservation gives the small-droplet radius. Total volume is conserved:

\frac{4}{3}\pi R^3 = N\cdot \frac{4}{3}\pi r^3 \quad\Rightarrow\quad r = \frac{R}{N^{1/3}} = \frac{1.00\times10^{-3}\ \text{m}}{(10^{6})^{1/3}} = \frac{1.00\times10^{-3}}{100} = 1.00\times10^{-5}\ \text{m}.

Step 2 — initial and final surface areas.

A_i = 4\pi R^2 = 4\pi (1.00\times10^{-3}\ \text{m})^2 = 1.2566\times10^{-5}\ \text{m}^2.

A_f = N\cdot 4\pi r^2 = 10^{6}\times 4\pi (1.00\times10^{-5}\ \text{m})^2 = 10^{6}\times 1.2566\times10^{-9}\ \text{m}^2 = 1.2566\times10^{-3}\ \text{m}^2.

Step 3 — area created and the reversible work.

\Delta A = A_f - A_i = 1.2566\times10^{-3} - 1.2566\times10^{-5} = 1.2441\times10^{-3}\ \text{m}^2.

w_{\text{rev}} = \gamma\,\Delta A = (72.8\times10^{-3}\ \text{N m}^{-1})(1.2441\times10^{-3}\ \text{m}^2) = 9.06\times10^{-5}\ \text{J}.

Sanity check. This is a tiny but nonzero energy (~0.09 mJ), consistent with the fact that atomization multiplies surface area 100-fold ( $A_f/A_i = N^{1/3} = 100$ ) while the total mass and volume stay fixed — exactly the physics behind why fog and aerosols are thermodynamically "expensive" states of water compared to a single drop, and why they spontaneously coalesce back given the chance (Ostwald ripening, foreshadowing the Kelvin equation). For comparison, this energy is far smaller than the drop's gravitational potential energy on falling even a few centimeters, showing that surface effects, while measurable, are a minor energy player at the millimeter scale — but dominate for micrometer and nanometer particles.

Limiting cases & common traps

Situation	What happens to $\gamma$	Common trap
$T \to T_c$ (critical point)	$\gamma \to 0$ (no distinction between phases)	Forgetting $\gamma$ is temperature-dependent; using a single "room temperature" value at all $T$
Surfactant added	$\gamma$ drops sharply (e.g., water 72.8 → soap solution ~25–40 mN m⁻¹)	Assuming tap water and "clean" water have the same $\gamma$ — trace contamination changes it significantly
Small vs. large droplets	$\gamma$ itself is a bulk-interface property, independent of curvature to first order	Confusing surface tension with Laplace pressure ( $\Delta p = 2\gamma/r$ ), which does depend on radius
Two-sided films (soap films)	Total restoring force doubles: $F = 2\gamma l$	Using $F = \gamma l$ for a soap film and getting half the correct force
Solids	Surface stress and surface energy are not generally equal (unlike for liquids)	Treating a solid surface energy exactly like a liquid's mechanically measured $\gamma$

The chemist behind it

Agnes Pockels, working without formal university training or laboratory access in 1880s Germany, built a homemade trough from a kitchen sink and metal strips to measure how surface tension changed as she added tiny, controlled amounts of contaminants (oils, soaps) to water. Her sliding-barrier trough — later refined by Irving Langmuir into the modern Langmuir trough — let her show that surface tension depends sensitively on the amount of surface-active material present, laying the experimental groundwork for the whole field of monolayer and surfactant science. She sent her results to Lord Rayleigh in 1891, who recognized their importance and arranged for their publication in Nature, launching a scientific career she pursued largely from home for decades afterward.

Origins · 1 minAgnes Pockels and how this was discovered

Agnes Pockels

1862–1935 · Pioneered quantitative surface-tension measurement with a homemade trough, founding the experimental study of surfactant monolayers

Ask the tutor: How does surface tension combine with contact angle to predict whether a liquid will wet a solid surface?

Why does adding soap lower water's surface tension so much?

Soap molecules are amphiphilic: a polar or ionic head prefers water, while a nonpolar hydrocarbon tail prefers to escape it. At the air–water interface, soap molecules orient with tails pointing into the air and heads in the water, replacing water–water contacts at the surface with weaker water–tail contacts. This lowers the energy penalty for maintaining that surface, so less work is needed to create new area, and γ drops from about 72.8 mN/m for pure water to roughly 25–40 mN/m for typical soap solutions.

How is surface tension related to the Gibbs energy of the whole system, not just the interface?

The total Gibbs energy of a two-phase system can be split into a bulk contribution (extensive, proportional to volume of each phase) plus a surface excess contribution, γA, that only depends on the interfacial area. This partition is precisely how Gibbs originally formalized interfacial thermodynamics: γ is defined so that all the "extra" energy not explained by bulk phase behavior is attributed to the interface, making γ the natural bridge between molecular cohesion and macroscopic Gibbs energy.

Does surface tension depend on the curvature of the interface?

To an excellent approximation for droplets larger than a few nanometers, no — γ itself is treated as a bulk material property independent of the radius of curvature. What does depend on curvature is the pressure difference across the interface, given by the Laplace equation, Δp = 2γ/r for a sphere. Only at extremely small radii, comparable to a few molecular diameters, does γ itself start to shift measurably (the Tolman correction), a refinement beyond the introductory picture.

Why do water striders and needles float on water despite being denser than water?

The object's weight is supported not by buoyancy but by the vertical component of surface tension acting along the contact line where the object depresses the water surface. As long as the object's weight does not exceed the maximum restoring force the deformed surface can supply, roughly γ times the contact-line length, the surface behaves like a taut, slightly stretched membrane and holds the object up, a striking everyday demonstration of γ as force per length.