Observation of renormalization group invariance in symmetry-restored nuclear lattice effective field theory

Bing-Nan Lu; Chen-Can Wang; Jia-ai Shi

Observation of renormalization group invariance in symmetry-restored nuclear lattice effective field theory

T0 review · 3 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Adding Galilean-invariance-restoring counterterms makes the NLEFT prediction for the 4He binding energy independent of the contact cutoff from 250 to 400 MeV, matching experiment at the 100 keV level.

desk verdict GIR counterterms flatten the 4He binding energy across the soft cutoff in NLEFT, a solid and novel result, though the RG-invariance claim is limited by the fixed lattice spacing and untested GIR sufficiency in A=4. read the letter →

arxiv 2509.02953 v2 pith:MEPOGA6P submitted 2025-09-03 nucl-th hep-lat

Jia-Ai Shi , Chen-Can Wang , Bing-Nan Lu This is my paper

classification nucl-thhep-lat

keywords nuclearlatticeeffectivefieldtheoryrenormalizationgroupinvarianceGalileanrestorationhelium-4bindingenergyperturbativequantumMonteCarlochiralcutoffindependencelightnuclei

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper reports the first systematic verification of renormalization group invariance for a realistic nuclear few-body system in lattice effective field theory. The strategy is to repair the Galilean invariance broken by the lattice regulator, refit all low-energy constants to A<=3 observables at each cutoff, and then use the 4He binding energy as a pure prediction. Across contact cutoffs from 250 to 400 MeV, the predicted 4He energy stays essentially flat, between -28.52(11) and -28.22(15) MeV, within about 0.1 MeV of the experimental value. Extrapolating the cutoff to infinity gives -28.33(6) MeV versus -28.30 MeV experimentally. If correct, this establishes that a lattice-regulated ab initio nuclear calculation can produce cutoff-independent, accurate predictions once lattice symmetry breaking is handled.

What carries the argument

The central object is the Galilean-invariance-restoration counterterm, V_GIR = [g1 Q^2 + g2 Q^2 (sigma1·sigma2)] f2N, a momentum-dependent contact interaction proportional to the pair's total momentum squared, Q^2, with coefficients g1 and g2 fixed by demanding S-wave scattering lengths be independent of Q. It is regulated by the same soft single-particle cutoff used for the other contact terms. The GIR counterterm cancels the c2/Lambda^2 term in the expression E(Lambda) = E(infinity) + c2/Lambda^2 + c4/Lambda^4, which is the visible signature of renormalization group invariance in the 4He binding energy.

What would settle it

Compute the 4He binding energy with the same Hamiltonian plus a three-nucleon Galilean-restoration counterterm analogous to Eq. (9), varying its coefficient over the range implied by the two-body GIR fit; if E(Lambda) develops a nonzero 1/Lambda^2 slope, the observed flatness was not complete. Alternatively, compute 4He in a moving frame and check whether the energy is independent of total momentum Q at the 100 keV level.

Watch

Extended reading notes

Core claim

The central claim is that Galilean-invariance-restoration (GIR) counterterms, with coefficients fixed by requiring neutron-proton S-wave scattering lengths to be independent of the pair's total momentum, remove the dominant cutoff dependence from the four-nucleon system. With the three-body force refit to the triton at each cutoff, the 4He binding energy follows E(Lambda) = E(infinity) + c2/Lambda^2 + c4/Lambda^4, and the GIR terms drive c2 to zero: the corrected energies are flat in Lambda, while the uncorrected energies fall linearly in 1/Lambda^2. The continuum extrapolation gives E(infinity) = -28.33(6) MeV, agreeing with the experimental -28.30 MeV. The same pattern persists when the tw

Load-bearing premise

The assumption that the two-body Galilean-invariance-restoration counterterms, with coefficients set from two-body scattering data, fully remove Galilean-breaking effects in the four-body system; without a three-nucleon GIR term or a direct many-body consistency check, the observed flatness could be partly accidental.

Editorial extensions

If this is right

Across cutoffs 250 to 400 MeV, the GIR-corrected 4He binding energy stays within -28.52 to -28.22 MeV, matching experiment to about 0.1 MeV.
Extrapolating Eq. (11) to infinite cutoff gives E(infinity) = -28.33(6) MeV, statistically identical to the experimental -28.30 MeV, so the continuum limit is reached without a residual offset.
The GIR correction scales approximately as Lambda^-2: about 1 MeV for 4He at Lambda = 400 MeV and about 5 MeV at Lambda = 250 MeV, consistent with an NLO lattice artifact.
Omitting the GIR terms leaves a pronounced 1/Lambda^2 slope in the 4He energy; including them kills that slope, showing that symmetry restoration is what makes the prediction RG-invariant.
Varying the two N2LO three-body parameters cE and cD produces a Tjon band passing through the experimental 4He point, indicating the cutoff-independence is robust across different three-body force shapes.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If the flatness also survives the inclusion of the one-pion-exchange three-nucleon force at every cutoff—the paper only varies cD at a single cutoff—then RG invariance would extend to the full N2LO three-body sector, not just the contact term.
The same symmetry-restoration logic could be applied to ordinary lattice regulators, where rotational symmetry breaking is also present; demonstrating c2 = 0 there would show the mechanism is general, not specific to the soft regulator used here.
A sharper validation would be to compute 4He in a moving frame and check directly that the energy is independent of total momentum; the paper fixes g1 and g2 from two-body scattering but does not test the four-body Galilean Ward identity itself.
If the flatness holds, few-body correlations such as the Tjon line can be compared across different lattice regularizations without extrapolation, placing three-nucleon-force fits on firmer footing.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The paper is worth a serious look. The main new result is that adding Galilean-invariance-restoration counterterms to an N2LO lattice chiral Hamiltonian makes the 4He binding energy essentially flat across soft contact cutoffs from 250 to 400 MeV, with values within ~0.3 MeV of experiment. Without the GIR terms, the same quantity varies by ~1.7 MeV and converges slowly. The infinite-cutoff extrapolation from Eq. (11) gives -28.33(6) MeV with GIR versus -28.30 experimental. Notably, the no-GIR extrapolation gives the same value within larger errors, so the physical limit is robust; the GIR terms mainly accelerate convergence and reduce systematic uncertainty.

What's genuinely new is the combination of ptQMC (second-order perturbation) with GIR counterterms in a sliding-cutoff study of A<=4. The prior Tjon-line work used first-order perturbation and no GIR. The paper is transparent: Table 1 gives all fitted LECs and observables, and the 4He energy is a genuine prediction, not used in the fit. The Tjon-band plot with cD/cE variations supports that the pattern persists with more general 3NFs, at least at fixed cutoff.

The soft spots are real but not fatal. First, the advertised "RG invariance" is tested only by varying the soft contact cutoff at a fixed lattice spacing (a=0.987 fm, Lambda_a~628 MeV). The claim that this is sufficient rests on a separation of scales that is plausible but not directly demonstrated. A single calculation at a smaller lattice spacing would have strengthened the claim. Second, the GIR terms are fitted only to S-wave np scattering in moving frames; the regulator breaks Galilean invariance for all operator structures, and no many-body Galilean boost check is provided. The flatness of 4He could in principle be partly accidental, though the observed Lambda^-2 scaling of g1 and g2 and the magnitude of corrections are consistent with the leading artifact being captured. These are addressable limitations, not demonstrated errors.

Who is this for? Anyone using lattice EFT or chiral EFT with regulators, and anyone concerned with cutoff dependence of ab initio nuclear calculations. It deserves a serious referee. I'd send it to review with a request for a discussion of the lattice-spacing caveat and possibly a moving-frame consistency check, but the core result looks solid.

Recommendation: engage with it; accept for peer review, probably with minor revisions.

Referee Report

3 major / 4 minor

Summary. The paper presents NLEFT calculations of 3H, 3He, and 4He with N2LO chiral interactions regulated by a soft single-particle momentum cutoff Lambda. To compensate the Galilean invariance breaking induced by this regulator, two Q^2 counterterms g1,g2 (Eq. 9) are introduced and fixed to S-wave np scattering lengths as functions of the total pair momentum Q. Two-body LECs are fit to Nijmegen phase shifts in the center-of-mass frame, and cE is fit to the triton binding energy at each Lambda. The central numerical result is that with the GIR terms, the predicted 4He binding energy in Table 1 and Fig. 3 is flat across Lambda = 250-400 MeV and consistent with experiment, whereas without GIR it is not; extrapolation through Eq. (11) gives E(infinity) = -28.33(6) MeV versus -28.30 MeV experimentally. This is presented as the first systematic verification of RG invariance in NLEFT for realistic few-body systems.

Significance. If the result holds, it is a valuable demonstration that Symanzik-style symmetry-restoration counterterms can restore cutoff independence in a realistic NLEFT calculation at the 0.1 MeV level. The 4He energy is a genuine prediction: calibration uses only A<=3 observables, and the paired comparison with and without GIR at identical two-body LECs isolates the effect. The ptQMC second-order treatment gives statistical uncertainties of order 0.1 MeV, and the infinite-cutoff extrapolation provides a direct experimental benchmark. The paper also honestly identifies deferred issues (rotational restoration, lattice-spacing variation, cutoff-dependent radii), which helps delimit the scope of the claim.

major comments (3)

[Sec. 3.1, Eqs. (6), (9), (11)] The GIR coefficients are fixed using only S-wave np scattering lengths in moving frames, but the single-particle regulator Eq. (2) multiplies every term in Eqs. (1) and (4), including spin-orbit, tensor, and P-wave structures, and the three-body regulator Eq. (6) also breaks Galilean invariance. No GIR counterterm is introduced for the three-body force. The flatness of E(4He) in Fig. 3 could therefore be a partial artifact of an incomplete restoration. I ask for a concrete many-body Galilean check: compute E(3H) and E(4He) at nonzero total momentum P and verify that E(P)-E(0)-P^2/(2M) is below the ~0.1 MeV scale, or repeat the sliding-cutoff study after adding Q^4 and channel-dependent GIR operators and show that the cancellation of the c2 term in Eq. (11) persists.
[Sec. 3.1 (fixed lattice spacing)] Only the soft cutoff Lambda is varied while the lattice spacing is fixed at a = 0.987 fm. The assertion that Lambda_a ~ 628 MeV makes the results independent of a is an assumption, not a numerical result. In NLEFT the lattice spacing is itself the regulator, so the advertised 'RG invariance' is not yet lattice-spacing independence; it is soft-regulator independence of a fixed-lattice theory. The title and abstract should be qualified, or at least one additional lattice spacing should be computed to support the claim.
[Sec. 3.2, Fig. 4] The statement that the RG-invariance improvement 'robustly persists for more general 3NFs' is not supported by the data. Fig. 4 shows the Tjon correlation only at a single cutoff Lambda = 350, with cE and cD varied while all other LECs are held fixed; it does not demonstrate that a two-parameter 3NF refit at each Lambda yields a flat 4He energy. Since cE is refit at every Lambda and absorbs part of the cutoff dependence, this is a load-bearing gap. A concrete test is to repeat the sliding-cutoff analysis with both cE and cD refit (e.g., to 3H and 3He) and check E(4He) versus Lambda.

minor comments (4)

[Table 1 and Eq. (11)] The fitted values of c2 and c4, and their uncertainties, are not quoted. Please provide them so the statement 'consistent with c2 = c4 = 0' is numerically verifiable.
[References and text] Ref. [27] (Scarduelli et al.) appears unrelated to hypernuclei; please verify the citation. Also fix typos: 'ommission' -> 'omission', 'persued' -> 'pursued'.
[Fig. 3 caption/text] The text in Sec. 3.2 refers to 'dots' while the figure caption/legend says 'circles'; make the terminology consistent.
[Abstract and Sec. 4] The phrase 'parameter-free predictions' for 4He is loose: the LECs are fit to A<=3 observables. Consider 'prediction after A<=3 calibration' to avoid ambiguity.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the 4He binding energy is a genuine prediction from LECs calibrated to A≤3 observables.

full rationale

The paper's central claim is cutoff independence of the 4He binding energy. This quantity is not used in any LEC fit: two-body LECs (B, C) are fit to neutron-proton phase shifts in the center-of-mass frame, CSB coefficients to nn/pp scattering lengths, GIR coefficients (g1,g2) to the Q-independence of S-wave np scattering lengths, and the three-body LEC cE is calibrated to reproduce the triton binding energy at each Λ. The 4He energy is then computed with the calibrated Hamiltonian, so its near-constancy across Λ is a nontrivial prediction. The only by-construction result is the triton binding energy, which the text explicitly identifies as calibration ('calibrate cE to reproduce the experimental triton binding energy') and later as 'trivially reproduces the experimental value by construction'; it is not presented as evidence for RG invariance. The GIR counterterm form is imported from prior work by the same group (Refs [12,68]), but this is a method citation; the present paper independently re-fits g1,g2 at each Λ and tests the consequences in the many-body system, so the citation is not the load-bearing proof of the central claim. Remaining concerns—GIR sufficiency for A=4, fixed lattice spacing a=0.987 fm, and fixed pion/Coulomb cutoffs—are correctness/scope limitations rather than circular reductions. Hence the derivation chain is self-contained for the central 4He prediction.

Assumptions & free parameters 7 free parameters · 5 assumptions · 1 invented entities

The central 4He prediction rests on standard chiral EFT power counting, on the assumption that a fixed fine lattice spacing does not affect the Lambda-scan, on the sufficiency of two-body GIR counterterms in A=4, on a 3NF truncated to the cE contact term, and on the convergence of the ptQMC expansion. The calibration parameters are listed in Table 1; additional fit parameters c2 and c4 appear in the Eq. (11) extrapolation.

free parameters (7)

B1, B2 = Table 1, e.g., B1=-4.931 to -3.997, B2=-0.369 to -0.119
Leading-order contact LECs fitted to Nijmegen np phase shifts for each Lambda.
C1..C7 = Table 1
NLO contact LECs fitted to np phase shifts at each Lambda.
cnn, cpp = Table 1
Charge-symmetry-breaking contact LECs fitted to nn and pp scattering lengths.
g1, g2 = Table 1
GIR counterterm LECs fitted by requiring S-wave np scattering lengths to be independent of total momentum Q.
cE = Table 1 (1.863 to 0.289 as Lambda goes 250 to 400 MeV)
Three-body contact LEC fitted to reproduce the triton binding energy -8.482 MeV at each Lambda.
c'E = Table 1 (5.389 to 0.459)
Same as cE but for the no-GIR reference Hamiltonian.
c2, c4 in Eq. (11) = not tabulated; fit to Monte Carlo results
Extrapolation coefficients fitted to the computed 4He energies as functions of Lambda^-2; with GIR they are consistent with zero.

assumptions (5)

domain assumption Low-energy observables for A<=4 are governed by the smooth contact cutoff Lambda, and the lattice cutoff Lambda_a near 628 MeV is large enough that lattice discretization artifacts are negligible.
Invoked in Sec. 3.1: 'Given that Lambda_a substantially exceeds the other cutoffs, the low-energy physics is independent of the lattice spacing.' Central to interpreting the Lambda scan as an RG-invariance test.
domain assumption Two-pion-exchange and higher-order chiral terms are negligible below Prel <= 200 MeV and can be absorbed into contact terms.
Stated in Sec. 3.1 when omitting TPEP from the fit; supports the N2LO truncation.
ad hoc to paper The two-body GIR counterterms of Eq. (9), calibrated in the NN sector, restore Galilean invariance in A=3,4 nuclei.
Used in Sec. 3.2 to compute GIR-corrected 3H and 4He; no three-nucleon GIR term or many-body validation is presented.
domain assumption The N2LO three-body force is dominated by the cE contact term; omitting the cD one-pion-exchange term does not change the conclusion.
Sec. 3.2 supports this via a fixed-cutoff Tjon-band correlation at Lambda=350 MeV, not via a cutoff scan.
domain assumption ptQMC second-order perturbation around the Wigner-SU(4) action accurately converges for 4He.
Relied on from Refs. [10,71] for the central 4He numbers; not independently demonstrated here.

invented entities (1)

Galilean-invariance-restoration (GIR) counterterms V_GIR independent evidence
purpose: Compensate the total-momentum dependence of the single-particle regulator in moving frames and many-body systems.
The coefficients are constrained by requiring S-wave np scattering lengths independent of total momentum Q (Table 1), so the term has an external falsifiable handle; its extension to A=4 is assumed rather than independently tested.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Observation of renormalization group invariance in symmetry-restored nuclear lattice effective field theory." pith.science (2026). https://pith.science/paper/MEPOGA6P

@misc{pith2026250902953,
  author       = {Pith},
  title        = {Pith review of: Observation of renormalization group invariance in symmetry-restored nuclear lattice effective field theory},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MEPOGA6P}},
  note         = {Machine review of arXiv:2509.02953}
}

abstract

Renormalization group (RG) invariance implies that the predictions of effective field theory are independent of the momentum cutoffs introduced during regularization. Here we report the first systematic verification of RG invariance for realistic nuclear few-body systems within nuclear lattice effective field theory. To restore broken continuum rotational and Galilean symmetries, we employ Galilean-invariance-restoration counterterms and use a soft momentum regulator. We calibrate the two- and three-body next-to-next-to leading order (N$^2$LO) chiral forces using $A\leq 3$ observables and perform precision quantum Monte Carlo calculations to compute the $^4$He binding energy. The predicted energy remains constant across cutoffs from $250$~MeV to $400$~MeV and agrees well with the experimental value, with discrepancies of order 100 keV. Our results demonstrate the capability of extracting accurate, cutoff-independent predictions within lattice-regulated \textit{ab initio} nuclear theory.

Figures

Figures reproduced from arXiv: 2509.02953 by the authors.

**Figure 3.** Calculated ground-state energies of 4He as functions of Λ −2 . Circles (diamonds) denote results calculated without (with) the GIR terms. Lines represent the extrapolations according to Eq. (11). Red star marks the experimental value. We further examine the asymptotic behavior at large Λ [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

**Figure 2.** Binding energies of 3H (Upper panel) and 4He (Lower panel) calculated with two-body force only (triangles) and two- plus three-body forces (circles/diamonds) as functions of Λ. Diamonds represents the results including the GIR terms. Dotted lines denote the experimental values. The trends observed in the numerical results are more clearly visualized in [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Binding energies of 3H and 4He calculated with varying three-body force LECs cD and cE. The calculations cover the ranges △cE ∈ [−0.3,0.3] and △cD ∈ [−3,3]. Red star marks the experimental values. Empirical Tjon band is depicted to guide the eyes. data alone is difficu…

Discussion (0). Sign in to comment.

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.