Scale-Invariant Chiral Perturbation Theory

• Scale-Invariant Chiral Perturbation Theory is an effective field framework that enhances traditional ChPT by including a light scalar dilaton alongside pions to jointly realize chiral and scale symmetries.
• It employs nonlinear compensator fields to systematically incorporate explicit scale breaking from quark masses, trace anomalies, and deviations from infrared fixed points.
• The approach is applied to three-flavor QCD and near-conformal theories, providing insights into phenomena like the f0(500) resonance, dilaton–pion coupling, and lattice hyperscaling.

Searching arXiv for recent and foundational papers on scale-invariant/dilaton chiral perturbation theory. Scale-Invariant Chiral Perturbation Theory is a class of low-energy effective field theories in which ordinary chiral perturbation theory is enlarged by an explicit light scalar singlet interpreted as a dilaton, so that approximate scale symmetry is treated alongside chiral symmetry rather than being discarded from the outset. In this literature, “scale-invariant” does not denote exact dilatation symmetry in the physical theory; it denotes a nonlinear realization in which scale symmetry is implemented by a compensator field and then broken explicitly by quark or fermion masses, by the trace anomaly, or by distance from a conformal threshold. Two major realizations dominate the subject: chiral-scale perturbation theory for three-flavor QCD near a hypothesized infrared fixed point, and dilaton chiral perturbation theory for confining gauge theories close to the conformal window. In both cases, the decisive departure from ordinary ChPT is that the scalar is retained as a leading low-energy degree of freedom rather than integrated out (Crewther et al., 2015, Golterman et al., 2021).

1. Historical setting and motivation

The immediate motivation for chiral-scale perturbation theory in three-flavor QCD was the claim that lowest-order $SU(3)_L\times SU(3)_R$ ChPT fails in channels dominated by the scalar-isoscalar resonance $f_0(500)$, whose quoted pole parameters were

$$m_{f_0}=441-i\,272\ \text{MeV},\qquad |g_{f_0\pi\pi}|=3.31\ \text{GeV}.$$

Because $f_0(500)$ lies inside the nominal low-energy region and couples strongly to $\pi\pi$, the usual separation between the pseudoscalar Nambu–Goldstone sector and heavier non-NG resonances is said to break down; the proposed remedy is to promote the scalar to the leading EFT spectrum and reinterpret it as a QCD dilaton (Crewther et al., 2014).

A distinct but related motivation emerged in near-conformal confining theories, especially large-$N_f$ systems studied on the lattice. There the low-energy spectrum can contain a light $0^{++}$ scalar nearly degenerate with the pions, together with approximate hyperscaling over a broad fermion-mass range. Ordinary ChPT is not designed for such spectra. Dilaton ChPT was developed precisely to describe confining theories just below the conformal window, where explicit scale breaking is parametrically small and a pseudo-Nambu–Goldstone boson of approximate scale symmetry can remain light (Matsuzaki et al., 2013, Golterman et al., 2021).

The resulting subject therefore has two complementary lineages. In one, scale invariance is tied to a conjectured infrared fixed point in three-flavor QCD and to the identification $\sigma\sim f_0(500)$. In the other, it is tied to proximity to the conformal sill in the Veneziano limit, with explicit scale breaking tracked by $n_f-n_f^*$ or equivalent parameters. A plausible implication is that “Scale-Invariant Chiral Perturbation Theory” functions less as a single model than as a family of EFT constructions sharing the same structural move: pions plus a dilaton-like scalar at leading order (Crewther et al., 2015, Li et al., 2018).

2. Symmetry realization and low-energy degrees of freedom

The three-flavor QCD formulation starts from the trace anomaly

$$\theta^\mu_{\ \mu}=\frac{\beta(\alpha_s)}{4\alpha_s}G^a_{\mu\nu}G^{a\mu\nu}+\bigl(1+\gamma_m(\alpha_s)\bigr)\sum_{q=u,d,s}m_q\bar qq.$$

If $f_0(500)$0 flows to an infrared fixed point $f_0(500)$1 with $f_0(500)$2, then at that point the gluonic contribution vanishes and, in the chiral limit, $f_0(500)$3. The central claim of chiral-scale perturbation theory is that the vacuum nevertheless remains non-scale-invariant because $f_0(500)$4, so scale symmetry, like chiral symmetry, is realized in Nambu–Goldstone mode rather than Wigner mode. The NG content is then enlarged from $f_0(500)$5 to $f_0(500)$6, with $f_0(500)$7 identified with $f_0(500)$8 (Crewther et al., 2015).

Near-conformal dilaton ChPT adopts an analogous logic but with different microscopic control parameters. The microscopic Ward identities are written as

$f_0(500)$9

The EFT fields are

$$m_{f_0}=441-i\,272\ \text{MeV},\qquad |g_{f_0\pi\pi}|=3.31\ \text{GeV}.$$0

with $$m_{f_0}=441-i\,272\ \text{MeV},\qquad |g_{f_0\pi\pi}|=3.31\ \text{GeV}.$$1 the dilaton and $$m_{f_0}=441-i\,272\ \text{MeV},\qquad |g_{f_0\pi\pi}|=3.31\ \text{GeV}.$$2 defined through

$$m_{f_0}=441-i\,272\ \text{MeV},\qquad |g_{f_0\pi\pi}|=3.31\ \text{GeV}.$$3

Scale symmetry is nonlinear: $$m_{f_0}=441-i\,272\ \text{MeV},\qquad |g_{f_0\pi\pi}|=3.31\ \text{GeV}.$$4, $$m_{f_0}=441-i\,272\ \text{MeV},\qquad |g_{f_0\pi\pi}|=3.31\ \text{GeV}.$$5 (Matsuzaki et al., 2013).

In the modern dChPT formulation for theories near the conformal window, the low-energy fields are

$$m_{f_0}=441-i\,272\ \text{MeV},\qquad |g_{f_0\pi\pi}|=3.31\ \text{GeV}.$$6

with $$m_{f_0}=441-i\,272\ \text{MeV},\qquad |g_{f_0\pi\pi}|=3.31\ \text{GeV}.$$7 the conformal compensator. Explicit scale breaking is encoded in a low-energy parameter $$m_{f_0}=441-i\,272\ \text{MeV},\qquad |g_{f_0\pi\pi}|=3.31\ \text{GeV}.$$8, proportional to the distance from the conformal sill,

$$m_{f_0}=441-i\,272\ \text{MeV},\qquad |g_{f_0\pi\pi}|=3.31\ \text{GeV}.$$9

The fermion-mass term carries the exponent

$f_0(500)$0

so the scaling dimension of the explicit chiral-breaking operator is built directly into the EFT (Golterman et al., 2021).

3. Effective Lagrangians and expansion schemes

In three-flavor chiral-scale perturbation theory, the effective fields are the usual $f_0(500)$1 and a scalar $f_0(500)$2, with decay constants $f_0(500)$3 and $f_0(500)$4. The EFT is classified simultaneously by chiral order and scaling dimension: $f_0(500)$5 At leading order,

$f_0(500)$6

and the terms are dressed by exponentials in $f_0(500)$7, so that an operator of scaling dimension $f_0(500)$8 is multiplied by $f_0(500)$9. The anomaly sector carries $\pi\pi$0, while the mass sector carries $\pi\pi$1. Ward-identity analysis of the improved energy-momentum tensor gives the leading constraints

$\pi\pi$2

which are among the defining structural relations of the framework (Crewther et al., 2015).

Near-conformal dChPT has a parallel but not identical LO structure: $\pi\pi$3 Its power counting is

$\pi\pi$4

so the expansion is simultaneously in momenta, fermion mass, distance to the conformal sill, and $\pi\pi$5. This is a defining difference from ordinary ChPT, which expands only in $\pi\pi$6 and mass (Golterman et al., 2021).

A separate large-$\pi\pi$7 scale-invariant ChPT writes the compensator as $\pi\pi$8 and the leading Lagrangian as

$\pi\pi$9

Here the hard anomaly keeps the dilaton massive even in the chiral limit (Matsuzaki et al., 2013).

At one loop, the most general chiral $N_f$0 EFT with a light scalar singlet can be renormalized with the background-field method and heat-kernel expansion. The generic LO form is

$N_f$1

with chiral-dimension assignments

$N_f$2

The divergent one-loop effective Lagrangian is

$N_f$3

Specializing to chiral-scale perturbation theory gives

$N_f$4

and shows explicitly that light-scalar loops modify the RG evolution of the NLO couplings relative to standard $N_f$5 ChPT (Cata et al., 2019). A useful methodological parallel is the electroweak chiral EFT review that emphasizes chiral dimensions rather than canonical dimensions for nonlinear EFTs with a singlet scalar; it explicitly states that it is not a paper on scale-invariant ChPT, but its counting logic transfers directly (Cata, 2015).

4. Three-flavor QCD phenomenology

A central phenomenological result of $N_f$6 is the dilaton–pion coupling. After imposing the current-conservation constraint $N_f$7, the on-shell relation becomes

$N_f$8

With the narrow-width approximation,

$N_f$9

Using Roy-equation pole determinations

$0^{++}$0

the inferred dilaton decay constant is

$0^{++}$1

The same paper then estimates

$0^{++}$2

via the dilaton Goldberger–Treiman relation, and from the electromagnetic trace anomaly obtains

$0^{++}$3

using $0^{++}$4 (Crewther et al., 2015).

An earlier conference review had quoted $0^{++}$5 MeV from a Goldberger–Treiman-like analysis of $0^{++}$6 scattering and an estimate $0^{++}$7 from $0^{++}$8 phenomenology. The later bounds therefore represent a substantial phenomenological revision within the same general framework (Crewther et al., 2014).

The other major application is the $0^{++}$9 rule in nonleptonic kaon decay. In $\sigma\sim f_0(500)$0, a $\sigma\sim f_0(500)$1-pole mechanism appears already at lowest order: $\sigma\sim f_0(500)$2 The weak chiral-scale Lagrangian contains octet, 27-plet, and weak-mass operators dressed by $\sigma\sim f_0(500)$3-dependent exponents. If the weak-mass anomalous dimension differs from the strong one, $\sigma\sim f_0(500)$4, the weak mass term cannot be rotated away and a residual mixing

$\sigma\sim f_0(500)$5

survives, with

$\sigma\sim f_0(500)$6

In the authors’ picture, this pole graph dominates the isospin-$\sigma\sim f_0(500)$7 amplitude, so $\sigma\sim f_0(500)$8 need not be unnaturally large (Crewther et al., 2015).

To test that mechanism, the framework revives Crewther’s proposal to compute on-shell $\sigma\sim f_0(500)$9 on the lattice with nonzero momentum transfer $n_f-n_f^*$0, keeping both external mesons on shell. The no-tadpoles theorem remains

$n_f-n_f^*$1

and the allowed kinematic range is

$n_f-n_f^*$2

This setup is intended to extract $n_f-n_f^*$3 and $n_f-n_f^*$4 without contamination from the $n_f-n_f^*$5-pole that contributes to $n_f-n_f^*$6 (Crewther et al., 2015).

5. Near-conformal theories, lattice extrapolations, and hyperscaling

For near-conformal confining theories, the principal lattice-oriented formula is the dilaton mass relation

$n_f-n_f^*$7

At tree level this is often written as

$n_f-n_f^*$8

Its practical use is direct: a plot of $n_f-n_f^*$9 versus $$\theta^\mu_{\ \mu}=\frac{\beta(\alpha_s)}{4\alpha_s}G^a_{\mu\nu}G^{a\mu\nu}+\bigl(1+\gamma_m(\alpha_s)\bigr)\sum_{q=u,d,s}m_q\bar qq.$$0 determines the intercept $$\theta^\mu_{\ \mu}=\frac{\beta(\alpha_s)}{4\alpha_s}G^a_{\mu\nu}G^{a\mu\nu}+\bigl(1+\gamma_m(\alpha_s)\bigr)\sum_{q=u,d,s}m_q\bar qq.$$1 and, once $$\theta^\mu_{\ \mu}=\frac{\beta(\alpha_s)}{4\alpha_s}G^a_{\mu\nu}G^{a\mu\nu}+\bigl(1+\gamma_m(\alpha_s)\bigr)\sum_{q=u,d,s}m_q\bar qq.$$2, $$\theta^\mu_{\ \mu}=\frac{\beta(\alpha_s)}{4\alpha_s}G^a_{\mu\nu}G^{a\mu\nu}+\bigl(1+\gamma_m(\alpha_s)\bigr)\sum_{q=u,d,s}m_q\bar qq.$$3, and $$\theta^\mu_{\ \mu}=\frac{\beta(\alpha_s)}{4\alpha_s}G^a_{\mu\nu}G^{a\mu\nu}+\bigl(1+\gamma_m(\alpha_s)\bigr)\sum_{q=u,d,s}m_q\bar qq.$$4 are known, the slope determines $$\theta^\mu_{\ \mu}=\frac{\beta(\alpha_s)}{4\alpha_s}G^a_{\mu\nu}G^{a\mu\nu}+\bigl(1+\gamma_m(\alpha_s)\bigr)\sum_{q=u,d,s}m_q\bar qq.$$5 (Matsuzaki et al., 2013).

The same framework derives explicit one-loop pion-log corrections. Defining

$$\theta^\mu_{\ \mu}=\frac{\beta(\alpha_s)}{4\alpha_s}G^a_{\mu\nu}G^{a\mu\nu}+\bigl(1+\gamma_m(\alpha_s)\bigr)\sum_{q=u,d,s}m_q\bar qq.$$6

the corrected dilaton mass is written as

$$\theta^\mu_{\ \mu}=\frac{\beta(\alpha_s)}{4\alpha_s}G^a_{\mu\nu}G^{a\mu\nu}+\bigl(1+\gamma_m(\alpha_s)\bigr)\sum_{q=u,d,s}m_q\bar qq.$$7

The authors state that for then-current $$\theta^\mu_{\ \mu}=\frac{\beta(\alpha_s)}{4\alpha_s}G^a_{\mu\nu}G^{a\mu\nu}+\bigl(1+\gamma_m(\alpha_s)\bigr)\sum_{q=u,d,s}m_q\bar qq.$$8 data these chiral logs were numerically negligible because the simulations had not reached the truly soft regime (Matsuzaki et al., 2013).

The 2021 review emphasized a further distinctive feature: a controlled large-mass regime. With

$$\theta^\mu_{\ \mu}=\frac{\beta(\alpha_s)}{4\alpha_s}G^a_{\mu\nu}G^{a\mu\nu}+\bigl(1+\gamma_m(\alpha_s)\bigr)\sum_{q=u,d,s}m_q\bar qq.$$9

the large-mass solution is

$f_0(500)$00

Consequently,

$f_0(500)$01

This approximate hyperscaling occurs even though the theory remains confining and chirally broken. The EFT remains systematic provided

$f_0(500)$02

so the large-mass regime is parametrically broad: $f_0(500)$03 This regime has no analogue in ordinary QCD ChPT (Golterman et al., 2021).

Applications to lattice data reflect that structure. Tree-level dChPT fits to lower-mass LSD $f_0(500)$04 data reportedly described

$f_0(500)$05

well, with $f_0(500)$06-values as high as $f_0(500)$07 for a four-point fit and fitted parameters

$f_0(500)$08

By contrast, LatKMI data at larger masses could not be fit over the full range at tree level, and the review interpreted this as the onset of sizable NLO effects (Golterman et al., 2021).

6. Related formulations, extensions, and unresolved issues

Not every EFT with pions plus a light scalar is a scale-invariant chiral EFT in the strict sense. A generic scalar-extended ChPT writes

$f_0(500)$09

with a separate scalar potential. In its pseudo-dilaton specialization,

$f_0(500)$10

The framework therefore contains a pseudo-dilaton as a constrained subspace of parameter space, but the generic theory is deliberately agnostic about whether the scalar is truly a dilaton (Hansen et al., 2016, Hansen et al., 2018).

Scale-chiral EFT has also been extended to baryons and hidden local symmetry vector mesons in the Veneziano limit. There the compensator is

$f_0(500)$11

and the counting is

$f_0(500)$12

At leading order, the resulting “LOSS” formulation gives Brown–Rho-type scaling relations such as

$f_0(500)$13

with $f_0(500)$14 (Li et al., 2017, Li et al., 2018).

A recent Skyrme-model application based on scale-invariant chiral perturbation theory used

$f_0(500)$15

to separate quark-mass and gluonic contributions to the effective trace anomaly and to study nucleon gravitational form factors. In that construction the integrated anomalous part gives

$f_0(500)$16

and the paper reported that the gluonic part dominates that anomalous quarter, about $f_0(500)$17 of the total mass. It also found that the gluonic scale anomaly is crucial for the negative outer pressure, negative tangential pressure, negative $f_0(500)$18-term, and inward confining force, with representative values such as $f_0(500)$19 for $f_0(500)$20 MeV, $f_0(500)$21 MeV, and $f_0(500)$22 (Tanaka et al., 28 Jul 2025).

The subject remains controversial in several respects. The existence of the three-flavor infrared fixed point required by $f_0(500)$23 is not established, and the older proceedings explicitly state that there was “currently no consensus” on whether such an $f_0(500)$24 exists in real $f_0(500)$25 QCD (Crewther et al., 2014). The broad width of $f_0(500)$26 is a persistent concern, although the later status review argues that this is no worse than perturbative treatments of unstable particles because the EFT coefficients are defined in the exact chiral-scale limit where $f_0(500)$27 is massless and has zero width (Crewther et al., 2015). In near-conformal dChPT, the review criticizes generalized dilaton potentials of the form

$f_0(500)$28

when $f_0(500)$29 is not close to $f_0(500)$30, on the grounds that they do not admit a consistent EFT power counting (Golterman et al., 2021).

Taken together, these developments define Scale-Invariant Chiral Perturbation Theory as a technically rich but non-uniform field. Its common core is clear: approximate scale symmetry is implemented nonlinearly by a compensator or dilaton field, explicit breaking is inserted in symmetry-controlled form, and the light scalar is retained as a genuine low-energy mode. Its open question is equally clear: whether the underlying infrared dynamics of the relevant gauge theory truly supports that symmetry interpretation in each case.