Interfacial Work in Thermodynamics

• Interfacial work is the reversible work required to create new interface area, quantified by the interfacial tension (γ) that represents the free-energy change per unit area.
• It applies to both equilibrium and nonequilibrium systems, with analytical and computational approaches like thermodynamic integration, mechanical pressure routes, and nonequilibrium estimators.
• The concept extends to active interfaces and fluid-wall systems, guiding simulations and continuum theories to resolve interfacial phenomena and predict dynamic responses.

Searching arXiv for the cited papers and related work on interfacial work. Interfacial work is the reversible work associated with creating, stretching, or deforming an interface between phases. In equilibrium thermodynamics it is quantified by the interfacial tension or interfacial free energy $\gamma$, defined as the work required to create unit area of interface and equivalently as a derivative of the Helmholtz free energy with respect to interfacial area. In nonequilibrium settings the same operational idea is retained, but the quantity conjugate to area can depend on the deformation protocol, on frozen density fields, on active stresses, or on interfacial mass transport. Related literatures also use adjacent notions such as the work of adhesion and the electronic work function, which encode different interfacial energetics and should not be conflated with $\gamma$ (Li et al., 2023, Bier, 2015, Carlson et al., 12 Aug 2025).

1. Classical thermodynamic meaning

In classical thermodynamics, the interfacial free energy $\gamma$ is defined as the reversible work required to create unit area of interface at constant temperature $T$, volume $V$, and number of particles $N$, or equivalently as the Helmholtz free-energy derivative with respect to interfacial area. The two standard statements are

$\gamma=\frac{\partial F}{\partial A}\bigg|_{N,T}$

and

$$\gamma=\frac{\partial F}{\partial A}\bigg|_{N,V,T},$$

with the corresponding infinitesimal work relation

$\delta W=\gamma\,\delta A.$

For an equilibrium interface, the reversible cost of creating area is therefore $W_{\rm eq}=\gamma_{\rm eq}A$ (Li et al., 2023, Benjamin et al., 2013).

This definition is independent of microscopic realization. In molecular simulation, one commonly constructs a reversible path between a bulk reference state and an interfacial target state, computes the free-energy difference $\gamma$0, and divides by $\gamma$1, so that

$\gamma$2

In continuum sharp-interface thermodynamics, the same quantity appears as the scalar coefficient multiplying area change, and for interfaces with interfacial mass it is identified with the negative interfacial thermodynamic pressure, $\gamma$3 (Benjamin et al., 2013, Bothe, 2020).

The classical definition encodes a reversible area work, not merely a stress anisotropy. That distinction becomes important in fluid-wall systems, relaxing nonequilibrium interfaces, and active matter, where mechanical stress formulas can be ambiguous, incomplete, or even qualitatively misleading.

2. Fluid-wall interfaces and the dividing-surface problem

For a fluid in contact with a flat structureless wall, the fluid-wall interfacial tension is not trivially and even unequivocally determined if one insists on identifying a precise fluid-wall dividing surface. Classical mechanical routes write

$\gamma$4

but for a fluid-wall system $\gamma$5, so the result depends on the arbitrary choice of $\gamma$6 (Li et al., 2023).

A recent resolution derives $\gamma$7 by two independent routes that do not require choosing a dividing surface. The first route imposes an isothermal, quasistatic “minimal deformation” of an interfacial slab in which the local density profile remains exactly the same function of the coordinate normal to the wall, so all intensive profiles are invariant. The second route regards the fluid-wall system as the limit of a fluid under a finite external potential field with two well-defined bulk regions and a nonuniform interface, then computes the work required to create a differential interface area. Both routes yield the same result:

$\gamma$8

and in the wall limit

$\gamma$9

This expression depends only on the bulk pressure $\gamma$0, the tangential pressure profile $\gamma$1, and the density profile $\gamma$2, and it is manifestly independent of any arbitrary dividing surface (Li et al., 2023).

In the minimal-deformation route, the effective fluid-wall border is automatically fixed to the equimolar dividing surface,

$\gamma$3

but that construction does not enter the final formula. The significance is methodological as well as conceptual: fluid-wall interfacial work can be uniquely quantified without an arbitrary geometrical prescription.

3. Computational routes in molecular simulation

For simple model fluids, interfacial work is routinely evaluated by free-energy methods, stress-based methods, and nonequilibrium work relations. In the Lennard-Jones wall studies of Benjamin and Horbach, the principal routes are thermodynamic integration (TI), the mechanical pressure-anisotropy route, nonequilibrium work estimators based on Jarzynski’s equality and the Bennett acceptance ratio (BAR), and Gibbs-Cahn integration along coexistence (Benjamin et al., 2013).

In TI, a coupling parameter $\gamma$4 is introduced into the Hamiltonian so that $\gamma$5 corresponds to the bulk system and $\gamma$6 to the system with walls. The free-energy difference follows from

$\gamma$7

The cited Lennard-Jones study uses a two-step path: bulk Lennard-Jones $\gamma$8 Lennard-Jones plus flat walls, then flat wall $\gamma$9 structured fcc wall. Typical runs use $T$0–$T$1 configurations for each $T$2 (Benjamin et al., 2013).

The mechanical route uses the Kirkwood-Buff formula

$T$3

where the factor $T$4 accounts for two interfaces in slab geometry. The same study emphasizes that this method requires very high precision in $T$5, because small statistical errors are magnified in the integral, and that it is equal to $T$6 only if the wall is structureless and rigid. Non-equilibrium work methods instead estimate $T$7 from switching trajectories. Jarzynski’s equality reads

$T$8

while BAR solves a self-consistent equation combining forward and reverse work distributions. In the Lennard-Jones wall calculations, BAR estimates agree with TI to $T$9 at moderate switching times $V$0–$V$1, and BAR can be $V$2 faster than TI under those conditions (Benjamin et al., 2013).

Gibbs-Cahn integration is especially efficient when $V$3 is needed along an entire coexistence line. For a one-component fluid at an external wall, the Gibbs adsorption relation is

$V$4

and along coexistence

$V$5

Once a single reference value of $V$6 is known, integration yields $V$7 cheaply. Across properly converged methods, the flat-wall Lennard-Jones system gives mutually consistent interfacial free energies with differences $V$8. These values can then be inserted into Young’s equation,

$V$9

to infer wetting behavior. At $N$0, the reported example $N$1 and $N$2 gives $N$3, corresponding to partial wetting with $N$4 (Benjamin et al., 2013).

4. Nonequilibrium interfacial tension during relaxation

Outside equilibrium, interfacial work can be defined operationally as the reversible isochoric work per unit area needed to deform a system while holding the instantaneous density perturbation effectively frozen. In the theory of relaxation of a colloidal fluid, the density profile is written $N$5, and the container walls are deformed very slowly so as not to excite turbulence but fast compared with the colloidal relaxation time. Because the molecular solvent is incompressible, each small fluid element preserves its volume and colloid number, and the deformed colloid density satisfies

$N$6

The excess Helmholtz free energy is represented as

$N$7

with $N$8 the direct correlation function of the uniform fluid (Bier, 2015).

To linear order in the deformation, the interfacial tension is obtained from the work-area derivative at constant volume:

$N$9

Within dynamic density functional theory, the Fourier amplitudes relax as

$\gamma=\frac{\partial F}{\partial A}\bigg|_{N,T}$0

so the time-dependent nonequilibrium interfacial tension is

$\gamma=\frac{\partial F}{\partial A}\bigg|_{N,T}$1

For an initial step profile, $\gamma=\frac{\partial F}{\partial A}\bigg|_{N,T}$2 and the formula simplifies accordingly (Bier, 2015).

A central result is that $\gamma=\frac{\partial F}{\partial A}\bigg|_{N,T}$3 is not necessarily positive. In the long-time limit, the sign is determined by the coefficient $\gamma=\frac{\partial F}{\partial A}\bigg|_{N,T}$4 in the small-$\gamma=\frac{\partial F}{\partial A}\bigg|_{N,T}$5 expansion of $\gamma=\frac{\partial F}{\partial A}\bigg|_{N,T}$6: if $\gamma=\frac{\partial F}{\partial A}\bigg|_{N,T}$7, then $\gamma=\frac{\partial F}{\partial A}\bigg|_{N,T}$8 at late times; if $\gamma=\frac{\partial F}{\partial A}\bigg|_{N,T}$9, then $$\gamma=\frac{\partial F}{\partial A}\bigg|_{N,V,T},$$0 at late times. Polymer solutions with a Gaussian-core potential and charge-stabilized colloids yield $$\gamma=\frac{\partial F}{\partial A}\bigg|_{N,V,T},$$1 and negative $$\gamma=\frac{\partial F}{\partial A}\bigg|_{N,V,T},$$2, whereas colloid-polymer mixtures in the Asakura-Oosawa model can show positive, negative, or sign-changing $$\gamma=\frac{\partial F}{\partial A}\bigg|_{N,V,T},$$3 depending on polymer concentration. This establishes that negative nonequilibrium interfacial tensions are consistent with strictly positive equilibrium interfacial tensions, and that the sign of $$\gamma=\frac{\partial F}{\partial A}\bigg|_{N,V,T},$$4 can influence the morphology of density perturbations during relaxation (Bier, 2015).

5. Active interfaces and motility-induced phases

In motility-induced phase separation, interfacial work must be reformulated because the interface is driven by irreversible, nonconservative forces and the quantity resisting deformations is neither the Kirkwood-Buff mechanical tension nor a thermal capillary stiffness. Starting from overdamped Langevin dynamics for active Brownian particles,

$$\gamma=\frac{\partial F}{\partial A}\bigg|_{N,V,T},$$5

with rotational noise

$$\gamma=\frac{\partial F}{\partial A}\bigg|_{N,V,T},$$6

Langford and Omar derive fluctuating hydrodynamics for the coarse-grained density and orientational order, and from these the interfacial height dynamics of coexisting motility-induced phases (Langford et al., 2023).

The steady interface carries an active stress

$$\gamma=\frac{\partial F}{\partial A}\bigg|_{N,V,T},$$7

with coefficients $$\gamma=\frac{\partial F}{\partial A}\bigg|_{N,V,T},$$8 and $$\gamma=\frac{\partial F}{\partial A}\bigg|_{N,V,T},$$9 determined by $\delta W=\gamma\,\delta A.$0 and $\delta W=\gamma\,\delta A.$1. Projecting these bulk fluxes onto capillary-wave modes via the Bray-Ohta ansatz,

$\delta W=\gamma\,\delta A.$2

yields a Langevin equation for the interface height,

$\delta W=\gamma\,\delta A.$3

where the capillary-wave tension $\delta W=\gamma\,\delta A.$4 contains both a local active-stress term and a nonlocal nematic-flow term. In the long-wavelength limit,

$\delta W=\gamma\,\delta A.$5

and this quantity is distinct from the mechanical Kirkwood-Buff tension

$\delta W=\gamma\,\delta A.$6

The distinction is substantive: simulations find $\delta W=\gamma\,\delta A.$7 while $\delta W=\gamma\,\delta A.$8 (Langford et al., 2023).

The capillary fluctuations are driven not by a thermal bath but by an active noise whose strength defines an athermal energy scale,

$\delta W=\gamma\,\delta A.$9

At wavelengths much larger than the persistence length, the anisotropic noise becomes negligible and the interface obeys an equilibrium-like capillary-wave spectrum with an interfacial stiffness that scales linearly with the intrinsic persistence length $W_{\rm eq}=\gamma_{\rm eq}A$0. In the same limit, the stationary distribution of height configurations is

$W_{\rm eq}=\gamma_{\rm eq}A$1

which is identical in form to equilibrium capillary-wave theory but with active $W_{\rm eq}=\gamma_{\rm eq}A$2 and athermal $W_{\rm eq}=\gamma_{\rm eq}A$3 (Langford et al., 2023).

The corresponding nonequilibrium interfacial work for deforming a flat interface to a small-slope height field is

$W_{\rm eq}=\gamma_{\rm eq}A$4

so the leading cost is

$W_{\rm eq}=\gamma_{\rm eq}A$5

Formally this mirrors $W_{\rm eq}=\gamma_{\rm eq}A$6 in equilibrium, but microscopically $W_{\rm eq}=\gamma_{\rm eq}A$7 is built from active stresses and nonconservative nematic flows. A common misconception is therefore corrected: in active matter, mechanical surface tension need not be the quantity that governs the work of interfacial deformation (Langford et al., 2023).

6. Sharp-interface continuum thermodynamics with interfacial mass

In sharp-interface continuum theories of multicomponent fluids, interfacial work appears as a local mechanical power term in the momentum, energy, and entropy balances on the moving interface $W_{\rm eq}=\gamma_{\rm eq}A$8. The bulk momentum balance is

$W_{\rm eq}=\gamma_{\rm eq}A$9

while the interface momentum balance includes the interfacial stress tensor $\gamma$00 and jump terms from the two bulk phases. The corresponding interfacial Helmholtz free energy per unit area is introduced by

$\gamma$01

with $\gamma$02 and the interfacial Gibbs-Duhem relation

$\gamma$03

Equivalently,

$\gamma$04

so the surface tension is identified as $\gamma$05 (Bothe, 2020).

The entropy production reveals the reversible interfacial work term directly. Writing $\gamma$06, one obtains

$\gamma$07

For a clean interface with no tangential stretch, the surface transport theorem gives

$\gamma$08

and therefore

$\gamma$09

This identifies the interfacial work rate as

$\gamma$10

In this formulation, interfacial work is the mechanical power supplied by the surface stress and exactly matches the time-rate of change of interfacial free energy (Bothe, 2020).

The same framework couples interfacial work to adsorption, desorption, and interfacial mass transfer. The entropy production contains species-transfer contributions driven by differences between bulk and interfacial chemical potentials, and under isothermal conditions one may close these by nonlinear energy-barrier laws of activated-reaction type. Because $\gamma$11 depends on the local surface tension through the interfacial Gibbs-Duhem relation, adsorption of surface-active agents changes $\gamma$12 and thereby modifies interfacial mass fluxes. The first and second laws then separate the reversible interfacial work block $\gamma$13 from irreversible dissipation due to viscous flow, heat conduction, species diffusion, reaction, slip-friction, and sorption (Bothe, 2020).

7. Adjacent interfacial energetic quantities: adhesion and work function

A related but distinct quantity is the work of adhesion, defined thermodynamically through Young’s equation,

$\gamma$14

with $\gamma$15 the liquid-vapor interfacial tension and $\gamma$16 the macroscopic contact angle. In sub-nanometer hydrodynamics, Carlson and Netz compute $\gamma$17 from molecular dynamics of cylindrical water droplets on self-assembled monolayers, extrapolating a microscopic contact angle $\gamma$18 linearly to $\gamma$19. They then show that the Navier friction coefficient $\gamma$20, interfacial viscosity excess $\gamma$21, and depletion length $\gamma$22 each follow exponential laws in the dimensionless adhesion energy $\gamma$23:

$\gamma$24

with fitted parameters $\gamma$25, $\gamma$26, $\gamma$27, $\gamma$28, $\gamma$29, and $\gamma$30 (Carlson et al., 12 Aug 2025). These results concern adhesion-controlled transport rather than the area work $\gamma$31 itself, but they demonstrate how a single interfacial energetic descriptor can organize nanoscale flow data.

The electronic work function is another distinct notion. It is the minimum energy needed to remove an electron from the Fermi level into the vacuum just outside the surface. In epitaxial RuO$\gamma$32/TiO$\gamma$33 heterostructures, coherent strain at a buried RuO$\gamma$34/TiO$\gamma$35(110) interface stabilizes interfacial polarization and modulates the RuO$\gamma$36 surface work function by over $\gamma$37 eV, controlled by thickness in the $\gamma$38–$\gamma$39 nm range with a critical thickness of $\gamma$40 nm. The reported relation between polarization and work-function shift is

$\gamma$41

and the measured sample work function rises from $\gamma$42 eV at $\gamma$43 nm to $\gamma$44 eV at $\gamma$45 nm before dropping as strain relaxes between $\gamma$46 and $\gamma$47 nm (Jeong et al., 10 Jul 2025). This is an interfacial polarization effect in a metal, not an interfacial tension measurement.

In metallic magnetic trilayers, work-function differences also correlate with magnetic interfacial energetics. For Pt/Co/X stacks, Park et al. define $\gamma$48 and find a clear linear trend between the interfacial Dzyaloshinskii-Moriya interaction constant $\gamma$49 and $\gamma$50:

$\gamma$51

with $\gamma$52, $\gamma$53, and $\gamma$54. The interpretation given is that a work-function mismatch creates an interfacial electrostatic potential gradient that enhances spin-orbit scattering at the Co/X interface (Park et al., 2017). This usage again lies outside capillary interfacial work proper, but it illustrates the broader role of interfacial energetic discontinuities in determining measurable material responses.