Renormalized Defect Free Energy

• Renormalized defect free energy is the finite excess free energy attributed to localized inhomogeneities after subtraction of divergent bulk, UV, or singular contributions.
• It is constructed via methods such as thermodynamic limits in atomistic chains, finite-size entropy corrections in crystallites, and partition-function ratios in conformal defect theories.
• This renormalized energy determines equilibrium defect stability, interaction strengths, and plays a key role in effective temperature-dependent behavior in materials and field theoretic contexts.

Renormalized defect free energy is a finite defect-associated thermodynamic potential obtained after removing the extensive background contribution, the ultraviolet self-energy, or other singular terms that make a raw defect energy non-universal or divergent. Across the cited literature, the phrase appears in several technically distinct but structurally related settings: the thermodynamic-limit defect-formation free energy $G_\infty(A)$ of an atomistic chain (Dobson et al., 2016), a size-dependent effective vacancy formation free energy in finite crystallites (Hossein-Babaei et al., 2016), the universal part of $-\log(Z_{\mathcal D}/Z_{\mathrm{CFT}})$ for conformal defects (Giombi et al., 2023, Yuan et al., 2022), and the finite remainder that survives after subtracting logarithmic core energies for dislocations and vortices (Wu, 2016, Badal et al., 2022). In each case, the operative question is the same: what is the physically meaningful excess free energy carried by a localized or lower-dimensional inhomogeneity once bulk and singular pieces have been removed?

1. Terminological scope and canonical definitions

The literature does not use a single normalization for renormalized defect free energy. Instead, the term labels a family of constructions in which a defect contribution is isolated from a larger free-energy functional. In atomistic statistical mechanics this isolation is often literal: the defect-formation free energy is defined as a free-energy difference,

$G_N(A):=F_N^P(A)-F_N(A),$

between defective and perfect systems under the same macroscopic constraint. In defect CFT it is the logarithm of a partition-function ratio,

$$F_D=-\log\!\left(\frac{Z^{\mathrm{DCFT}}}{Z^{\mathrm{CFT}}}\right),$$

whose universal part is either a finite term or a logarithmic coefficient, depending on defect dimension. In continuum theories of singular defects it is the finite remainder left after subtracting the divergent core contribution from the elastic or Ginzburg–Landau energy (Dobson et al., 2016, Giombi et al., 2023, Yuan et al., 2022, Wu, 2016, Badal et al., 2022).

Setting	Representative object	Renormalization mechanism
Atomistic chain	$G_N(A)$, $G_\infty(A)$	thermodynamic limit and coarse-grained reduction to a single core bond
Finite crystal vacancies	$\Delta G_v^{\mathrm{eff}}(n_0,T,c_v)$	finite-size entropy correction to vacancy formation free energy
Defect CFT	$F_D=-\log(Z^{\mathrm{DCFT}}/Z^{\mathrm{CFT}})$	cancellation of bulk partition function and extraction of universal finite or logarithmic term
Dislocations / vortices	$F$, $W(u)$, $-\log(Z_{\mathcal D}/Z_{\mathrm{CFT}})$0	subtraction of logarithmic core energies or bulk Coulomb terms

A common misconception is that renormalization here always means perturbative UV renormalization in the field-theoretic sense. The cited work shows a broader usage. In some settings the renormalization is combinatorial and finite-size entropic; in others it is a rigorous subtraction of core singularities; in defect CFT it is the extraction of the universal defect contribution after bulk cancellation. The unifying feature is not a single scheme but the construction of a finite excess quantity with direct thermodynamic or RG meaning.

2. Atomistic defect thermodynamics: thermodynamic limits and finite-size entropy corrections

A rigorous atomistic realization appears in a one-dimensional constrained chain with a localized defect potential $-\log(Z_{\mathcal D}/Z_{\mathrm{CFT}})$1, where the defect-formation free energy is defined by

$-\log(Z_{\mathcal D}/Z_{\mathrm{CFT}})$2

and shown to admit a thermodynamic limit

$-\log(Z_{\mathcal D}/Z_{\mathrm{CFT}})$3

Here $-\log(Z_{\mathcal D}/Z_{\mathrm{CFT}})$4 is the finite-temperature Cauchy–Born energy density, and the limit is reached with error $-\log(Z_{\mathcal D}/Z_{\mathrm{CFT}})$5. The same paper constructs a coarse-grained energy $-\log(Z_{\mathcal D}/Z_{\mathrm{CFT}})$6 by integrating out all degrees of freedom except the core bond $-\log(Z_{\mathcal D}/Z_{\mathrm{CFT}})$7, so that the thermodynamic defect free energy reduces to a one-dimensional canonical average over the core variable. In the no-force case,

$-\log(Z_{\mathcal D}/Z_{\mathrm{CFT}})$8

and the coarse-grained defect free energy has the same $-\log(Z_{\mathcal D}/Z_{\mathrm{CFT}})$9 accuracy as the full atomistic free energy (Dobson et al., 2016).

This construction is renormalized in a precise multiscale sense: the influence of the environment is absorbed into an effective energy landscape for the defect core. The defect is not described by a bare local energy alone; it is described by a local degree of freedom coupled to a coarse-grained background encoded by $G_N(A):=F_N^P(A)-F_N(A),$0 and $G_N(A):=F_N^P(A)-F_N(A),$1. A second common misconception is that such coarse-grained defect free energies are merely heuristic. In this one-dimensional setting the rate of convergence and the reduction to a single reaction coordinate are proved rigorously (Dobson et al., 2016).

A different renormalization mechanism governs vacancies in finite crystallites. Standard Stirling-based treatments predict a size-independent equilibrium vacancy concentration and therefore suggest that smaller crystals should have higher defect populations. A corrected combinatorial treatment shows instead that the equilibrium condition contains an explicit finite-size term,

$G_N(A):=F_N^P(A)-F_N(A),$2

with $G_N(A):=F_N^P(A)-F_N(A),$3 the number of atoms in the perfect crystallite and $G_N(A):=F_N^P(A)-F_N(A),$4. This can be rewritten as an effective vacancy formation free energy

$G_N(A):=F_N^P(A)-F_N(A),$5

so the bulk quantity $G_N(A):=F_N^P(A)-F_N(A),$6 is replaced by a size-dependent effective free-energy cost. The resulting theory predicts a critical size

$G_N(A):=F_N^P(A)-F_N(A),$7

below which vacancy-free crystallites are thermodynamically stable, and it reverses the standard inference that equilibrium defect populations necessarily increase as crystallites shrink (Hossein-Babaei et al., 2016).

This finite-size result is important because the renormalization is not produced by a modified microscopic Hamiltonian. It comes from a more accurate treatment of configurational entropy. In that sense, the defect free energy is renormalized by statistics rather than by chemistry.

3. Finite-temperature first-principles defect free energies

First-principles calculations in bcc Fe make explicit how phonon and electron excitations renormalize defect free energies away from their $G_N(A):=F_N^P(A)-F_N(A),$8 values. In this framework the total Helmholtz free energy of a defect-containing supercell is written as

$G_N(A):=F_N^P(A)-F_N(A),$9

and the vibrational and electronic contributions to formation or binding free energies are defined by the same defect-minus-reference combinations used for ground-state energies. The vibrational term is evaluated in the harmonic approximation,

$$F_D=-\log\!\left(\frac{Z^{\mathrm{DCFT}}}{Z^{\mathrm{CFT}}}\right),$$0

while $$F_D=-\log\!\left(\frac{Z^{\mathrm{DCFT}}}{Z^{\mathrm{CFT}}}\right),$$1 is obtained from the ground-state electronic density of states with Fermi–Dirac occupations (Posselt et al., 2016).

The resulting renormalization of defect energetics is quantitatively large. For vacancy–solute pairs in bcc Fe, the contribution of phonon and electron excitations to the free binding energy is generally not negligible, and at $$F_D=-\log\!\left(\frac{Z^{\mathrm{DCFT}}}{Z^{\mathrm{CFT}}}\right),$$2 the maximum decrease and increase of the absolute value of the ZP-based free binding energy with respect to the corresponding ground-state value are found for the vacancy–W pair and the vacancy–Mn pair, namely $$F_D=-\log\!\left(\frac{Z^{\mathrm{DCFT}}}{Z^{\mathrm{CFT}}}\right),$$3 and $$F_D=-\log\!\left(\frac{Z^{\mathrm{DCFT}}}{Z^{\mathrm{CFT}}}\right),$$4 (Posselt et al., 2016). The same study reports that the general behavior of the free binding energy of vacancy and Cu dimers, trimers and quadromers is similar to that of the vacancy–solute pairs, and that quasi-harmonic corrections to the ZP-based results do not yield significant changes in the temperature range relevant for applications (Posselt et al., 2016).

These calculations also clarify the role of mechanical boundary conditions. The ground-state atomic structures are relaxed under constant volume and zero pressure conditions, and the phonon contribution is then computed around those two reference structures. A simple transformation similar to the quasi-harmonic approach is found between CV- and ZP-based frequencies, but, in contrast to ground-state energetics, the CV- and ZP-based defect free energies do not become equal with increasing supercell size (Murali et al., 2015). Electron excitations can either produce an additional deviation of the total defect free energy from the ground-state value or compensate the deviation caused by the phonon contribution (Murali et al., 2015).

In practical terms, these results replace a bare defect energy by a temperature-dependent effective free energy. The renormalized quantity controls equilibrium concentrations, binding equilibria, and diffusion kinetics. For self-diffusion via the vacancy mechanism, the same first-principles framework shows that the near-cancellation between the ratio of CV- and ZP-based vacancy diffusivities and the reciprocal ratio for equilibrium concentrations leads to almost identical self-diffusion coefficients, but the paper explicitly remarks that this agreement is accidental (Murali et al., 2015).

4. Conformal defects, universal defect free energy, and monotonicity

In defect conformal field theory the central object is the defect free energy on a sphere,

$$F_D=-\log\!\left(\frac{Z^{\mathrm{DCFT}}}{Z^{\mathrm{CFT}}}\right),$$5

where bulk UV divergences cancel in the ratio and only defect-localized divergences remain. For a round $$F_D=-\log\!\left(\frac{Z^{\mathrm{DCFT}}}{Z^{\mathrm{CFT}}}\right),$$6 surface defect inside $$F_D=-\log\!\left(\frac{Z^{\mathrm{DCFT}}}{Z^{\mathrm{CFT}}}\right),$$7, the general structure is

$$F_D=-\log\!\left(\frac{Z^{\mathrm{DCFT}}}{Z^{\mathrm{CFT}}}\right),$$8

with $$F_D=-\log\!\left(\frac{Z^{\mathrm{DCFT}}}{Z^{\mathrm{CFT}}}\right),$$9 and $G_N(A)$0 scheme-dependent and $G_N(A)$1 universal. In the $G_N(A)$2 model with a quadratic surface perturbation, the free theory flows to a nontrivial defect CFT with exact infrared anomaly coefficient

$G_N(A)$3

while the interacting $G_N(A)$4 Wilson–Fisher case gives

$G_N(A)$5

In both cases $G_N(A)$6, so $G_N(A)$7, in agreement with the $G_N(A)$8-theorem (Giombi et al., 2023).

The universal part of the defect free energy depends on defect dimension. For even-dimensional defects, the universal datum is the logarithmic coefficient, not the finite part. For odd-dimensional defects, the universal datum is the finite part after subtracting power-law divergences. This parity dependence is essential: one should not identify the renormalized defect free energy with a raw finite term in all dimensions (Giombi et al., 2023, Yuan et al., 2022).

Free scalar and free fermion DCFTs provide exact boundary-condition realizations of this structure on $G_N(A)$9. In the scalar case, Dirichlet-type boundary conditions always exist, while Neumann-type boundary conditions are allowed only for defects of lower codimensions; the quantity

$G_\infty(A)$0

is then used as an interpolating defect free energy, and explicit Neumann-to-Dirichlet flows satisfy the conjectured defect $G_\infty(A)$1-theorem (Nishioka et al., 2021). For free Dirac fermions, Dirichlet-type boundary conditions always exist, Neumann-type boundary conditions exist only for a two-codimensional defect, and the double-trace flow from Neumann to Dirichlet obeys

$G_\infty(A)$2

so the defect free energy decreases along the defect RG flow (Sato, 2021).

A complementary entropic formulation identifies the universal defect free energy with defect localized entropy at a conformal fixed point. For a $G_\infty(A)$3-dimensional defect,

$G_\infty(A)$4

and the universal pieces satisfy

$G_\infty(A)$5

This yields monotone defect $G_\infty(A)$6-functions,

$G_\infty(A)$7

which decrease along defect RG flows by strong subadditivity and reduce at fixed points to the universal part of the defect free energy (Yuan et al., 2022).

5. Singular defects: dislocations, fractional vortices, and Coulomb gases

For singular continuum defects, renormalized defect free energy usually means the finite part that remains after subtracting the logarithmic core divergence. In anti-plane shear of a hexagonal quasi-crystal with multiple screw dislocations, the core-regularized energy satisfies

$G_\infty(A)$8

where $G_\infty(A)$9 is the universal core-energy coefficient and

$\Delta G_v^{\mathrm{eff}}(n_0,T,c_v)$0

is the renormalized energy. It governs self-energy, pairwise logarithmic interactions, and coupling to boundary-induced elastic fields, and the Peach–Köhler force is

$\Delta G_v^{\mathrm{eff}}(n_0,T,c_v)$1

Here the renormalized defect free energy is a mesoscale effective energy for the dislocation configuration after the singular core contribution has been removed (Wu, 2016).

An analogous construction appears for fractional vortices with topologically induced free discontinuities on a closed two-dimensional Riemannian manifold. The basic functional is

$\Delta G_v^{\mathrm{eff}}(n_0,T,c_v)$2

with both point singularities and string defects. After subtracting the leading core term $\Delta G_v^{\mathrm{eff}}(n_0,T,c_v)$3, the $\Delta G_v^{\mathrm{eff}}(n_0,T,c_v)$4-limit is

$\Delta G_v^{\mathrm{eff}}(n_0,T,c_v)$5

The three pieces are, respectively, vortex interaction energy, string length, and universal microscopic core energy. Geometry enters through the manifold metric, the Gaussian curvature in the vorticity, and the geodesic structure of the jump set (Badal et al., 2022).

For Coulomb-type point systems, the Sandier–Serfaty renormalized energy and its configuration-level analogue $\Delta G_v^{\mathrm{eff}}(n_0,T,c_v)$6 measure disorder relative to a constant average density by subtracting the two leading terms of the Coulomb energy. In one dimension, for the $\Delta G_v^{\mathrm{eff}}(n_0,T,c_v)$7-sine processes,

$\Delta G_v^{\mathrm{eff}}(n_0,T,c_v)$8

while in the plane

$\Delta G_v^{\mathrm{eff}}(n_0,T,c_v)$9

For these processes the variance of the renormalized energy vanishes, so the energy concentrates near its expected value, and the $F_D=-\log(Z^{\mathrm{DCFT}}/Z^{\mathrm{CFT}})$0 sine process minimizes the renormalized energy in the class of translation-invariant determinantal point processes (Borodin et al., 2012).

Process	$F_D=-\log(Z^{\mathrm{DCFT}}/Z^{\mathrm{CFT}})$1
$F_D=-\log(Z^{\mathrm{DCFT}}/Z^{\mathrm{CFT}})$2 sine	$F_D=-\log(Z^{\mathrm{DCFT}}/Z^{\mathrm{CFT}})$3
$F_D=-\log(Z^{\mathrm{DCFT}}/Z^{\mathrm{CFT}})$4 sine	$F_D=-\log(Z^{\mathrm{DCFT}}/Z^{\mathrm{CFT}})$5
$F_D=-\log(Z^{\mathrm{DCFT}}/Z^{\mathrm{CFT}})$6 sine	$F_D=-\log(Z^{\mathrm{DCFT}}/Z^{\mathrm{CFT}})$7
Ginibre	$F_D=-\log(Z^{\mathrm{DCFT}}/Z^{\mathrm{CFT}})$8
GAF zeros	$F_D=-\log(Z^{\mathrm{DCFT}}/Z^{\mathrm{CFT}})$9

These examples show that renormalized defect free energy can also be a disorder functional. The subtraction removes the trivial mean-field part, and the remainder quantifies how far a point configuration lies from the most ordered state compatible with the same average density.

6. Common structure, related renormalizations, and conceptual limits

Despite the diversity of settings, the constructions above share a stable pattern. One begins with a defect-sensitive quantity that is either divergent, overly microscopic, or contaminated by a bulk background: a free-energy difference in an atomistic chain, a vacancy formation free energy in a finite crystal, a partition-function ratio in a DCFT, or an elastic or Coulomb energy with logarithmic core singularities. One then subtracts the bulk or singular piece, or integrates out non-core degrees of freedom, and the remainder becomes the renormalized defect free energy. Depending on context, the resulting quantity governs equilibrium defect populations, effective defect interactions, RG monotonicity, or the entropy carried by localized degrees of freedom.

A related, but explicitly analogical, line of work comes from the functional renormalization group treatment of the field-independent term $F$0. That term is associated with free energy in flat-space statistical field theory, but its RG flow contains spurious UV pieces. The paper proposes subtraction schemes removing $F$1 and $F$2 terms in $F$3, restoring the Gaussian fixed point and leaving only logarithmic UV behavior. Its extension to defect free energy is presented as an analogy rather than a worked defect calculation: the defect contribution would be defined as a difference of effective actions with and without the defect, with bulk-like UV pieces subtracted first (Marian et al., 2021).

A second related construction appears in asymptotically Poincaré–Einstein geometry, where differences in renormalized volumes give rigorous meaning to the Hawking–Page difference of actions and describe the free energy liberated in the transition (Bahuaud et al., 2013). Although this is not a localized defect theory, it exhibits the same structural principle: a physically meaningful free-energy difference emerges only after a geometric renormalization has removed divergent background terms.

The most important conceptual limit is therefore not technical but terminological. Renormalized defect free energy is not a single object with a universal formula. It is a class of excess thermodynamic functionals adapted to different defect problems. What is universal is the logic of construction: isolate the defect sector, subtract bulk or singular contributions, and retain the finite quantity that controls stability, interaction, or RG flow.