Three-Body Renormalization and Universality

In a resonant three-body channel, pairwise scattering lengths do not determine the leading amplitude. Iterating the zero-range two-body interaction exposes an ultraviolet log-periodic mode, so a three-body counterterm must run on a limit cycle and one three-body datum must fix its phase. The resulting discrete scale invariance produces Efimov spectra and universal correlations, but only within the window between the infrared scales and the inverse range.

Required background. Pionless EFT and Shallow Two-Body Systems supplies the promoted two-body amplitude, pole conventions, and regulator checks used here.

Helpful background. Multiple Couplings and Coupled RG Flows supplies the distinction between fixed points, trajectories, and cycles in coupling space.

The zero-range three-body equation

Consider three identical nonrelativistic bosons of mass $M$, or a spin–isospin channel with the same attractive S-wave ultraviolet kernel, at momenta

$$1/|a|,\sqrt{M|E|},k\sim Q\ll\Lambda_b.$$

The degrees of freedom are the particles themselves (or an auxiliary dimer plus a spectator); pions and finite-range structure are integrated out. The two-body scattering length is resummed at leading order. With a sharp spectator cutoff $\Lambda$, define

$$\tau(q;E)= \frac{1}{-1/a+\sqrt{3q^2/4-ME-i0}}$$

and the exchange kernel

$$\mathcal K(p,q;E)=\frac{1}{pq} \ln\!\left( \frac{p^2+pq+q^2-ME}{p^2-pq+q^2-ME} \right).$$

Up to channel-dependent overall normalizations, the S-wave Skorniakov–Ter-Martirosian equation has the structure

$$\begin{aligned} t(p,k;E)={}&b(p,k;E) +\frac{2}{\pi}\int_0^\Lambda dq\,q^2 \left[\mathcal K(p,q;E)+\frac{H(\Lambda)}{\Lambda^2}\right]\\ &\times\tau(q;E)t(q,k;E). \end{aligned}$$

The inhomogeneous term $b$ contains the same exchange and three-body-contact structures for scattering; for a bound state it is absent and the equation is homogeneous. Precise factors can be redistributed by rescaling the dimer field, but the on-shell amplitude and the running described below cannot. The derivation and channel projections are given in Hammer, König, and van Kolck 2020, § II.C, pp. 15–22.

Setting $H=0$ does not define a regulator-independent leading theory. As $q\gg1/|a|,\sqrt{M|E|}$, the equation becomes scale free. A homogeneous ansatz $t(q)\propto q^{s-1}$ gives

$$1=\frac{8}{\sqrt3\,s} \frac{\sin(\pi s/6)}{\cos(\pi s/2)}.$$

Its leading roots are $s=\pm i s_0$, with $s_0\simeq1.00624$, so the ultraviolet solution is

$$t(q)\propto\frac{1}{q} \sin\!\left[s_0\ln(q/\Lambda_*)+\phi\right].$$

Changing $\Lambda$ changes the phase sampled at the boundary. Two-body renormalization fixes $a$ but supplies no condition on this phase: observables therefore oscillate with $\ln\Lambda$. This is the momentum-space form of the Thomas collapse and is the concrete reason a new leading operator is required—not merely an empirical preference for a three-body force Bedaque, Hammer, and van Kolck 1999, pp. 463–467.

Limit-cycle renormalization

For a conventional sharp-cutoff normalization, cutoff independence is restored by

$$H(\Lambda)= \frac{\cos\!\left[s_0\ln(\Lambda/\Lambda_*)+\arctan s_0\right]} {\cos\!\left[s_0\ln(\Lambda/\Lambda_*)-\arctan s_0\right]}.$$

The dimensionless function $H$ multiplies the leading three-body contact; its precise phase convention changes with regulator and operator normalization. The dimensional transmutation scale $\Lambda_*$ is not predicted from $a$. One measured three-body binding energy, atom–dimer scattering observable, or equivalent datum fixes it, after which other leading-order observables in that channel follow.

Three exact checks expose the cycle:

$$\begin{aligned} H(\Lambda_*)&=1,\\ H\!\left(\Lambda_*e^{\pi/(2s_0)}\right)&=-1,\\ H\!\left(\Lambda e^{\pi/s_0}\right)&=H(\Lambda). \end{aligned}$$

Thus the preferred momentum or cutoff scaling factor is

$$\lambda_0=e^{\pi/s_0}\simeq22.694,$$

while binding-energy magnitudes of adjacent states at unitarity differ by $\lambda_0^2$. A reproducible calculation checks these values against the supplied convention. A different regulator can distort the plotted trajectory of its bare coupling, but it must reproduce the same renormalized spectrum and scattering observables after the same three-body datum is matched.

Universal predictions and what fixes them

At $|a|\Lambda_b\gg1$, an observable with dimensions $[Q]^d$ has the leading form

$$\mathcal O_3=\Lambda_*^d F_{\mathcal O}\!\left(a\Lambda_*,\operatorname{sgn}a\right),$$

where the dimensionless function is log periodic at unitarity. Eliminating $\Lambda_*$ between two three-body observables produces a universal correlation. Such a curve is a prediction only after the following are kept distinct:

Input or assumption	Role
Two-body scattering lengths (and, at higher order, ranges)	Fix the pair amplitudes entering $\tau$
One three-body datum at leading order	Fixes $\Lambda_*$, the ultraviolet phase
Regulator and renormalization prescription	Define bare $H(\Lambda)$; must disappear from observables to the claimed order
Statistics, spin, isospin, Coulomb, and open channels	Select the kernel and whether the attractive log-periodic mode exists
Higher-order range and three-body operators	Control departures of relative size set by $Q/\Lambda_b$ and by the chosen counting

Two-body data are therefore necessary but insufficient in the resonant attractive channel. Conversely, a leading three-body counterterm is not universal to every three-particle system. Pauli repulsion or nonzero angular momentum can make the ultraviolet kernel less attractive—for example, the neutron–deuteron quartet channel does not require the same leading three-body datum. The eigenvalue analysis, not particle count alone, decides promotion.

Checks and limits

A controlled calculation should pass independent tests:

1. Cutoff test. Refit the one three-body datum as $\Lambda$ varies through a window $Q\ll\Lambda\lesssim\Lambda_b$; other observables should approach a plateau up to the predicted order. Holding the bare coupling fixed is not this test.
2. Dimensional test. The kernel and $H/\Lambda^2$ have the same momentum dimension, and any binding energy scales as momentum squared divided by $M$.
3. Unitarity test. Above breakup, use the correct $i0$ prescription and verify coupled-channel flux conservation; a real bound-state equation is insufficient there.
4. Order test. Insert natural effective ranges and higher operators according to the declared perturbative or iterative scheme, then check that their effects scale with $Q/\Lambda_b$.

Zero-range universality does not by itself control states with binding momenta near $\Lambda_b$, deep molecular or nuclear states, strong inelastic loss, long-range Coulomb forces, or channels whose statistics change the kernel. With deep open channels, a complex short-distance parameter can encode loss and the discrete scaling relations concern pole patterns rather than stable bound levels. Finite-range corrections break exact discrete scale invariance. These limitations and the scope of universal correlations are reviewed in Braaten and Hammer 2006, §§ 3–5, pp. 269–313.

The three-body force renormalizes strong amplitudes; it does not determine independent short-distance couplings to external probes. Those are organized in Electroweak Currents in Few-Body Systems, and correlated truncation and calibration effects belong in Nuclear Predictions, Uncertainties, and Evidence Across Methods.

Common pitfalls

Removing the cutoff before adding the counterterm. The oscillation is not numerical nonconvergence that a larger grid cures. Add the allowed three-body operator, fit one datum at each cutoff, and only then study residual dependence.

Counting one input as one prediction. A fitted trimer energy fixes $\Lambda_*$; it cannot also validate the theory. Test another binding energy, scattering observable, or held-out correlation.

Applying the Efimov factor to every level. The factor $\lambda_0^2$ is exact only in the zero-range unitary limit for the relevant attractive channel. Range effects, finite $a$, thresholds, and deep-state loss shift finite levels.

References

• Bedaque, Paulo F., Hans-Werner Hammer, and U. van Kolck. “Renormalization of the Three-Body System with Short-Range Interactions.” Physical Review Letters 82 (1999): 463–467. DOI.
• Braaten, Eric, and Hans-Werner Hammer. “Universality in Few-Body Systems with Large Scattering Length.” Physics Reports 428 (2006): 259–390. DOI.
• Hammer, Hans-Werner, Sebastian König, and U. van Kolck. “Nuclear Effective Field Theory: Status and Perspectives.” Reviews of Modern Physics 92 (2020): 025004. DOI.