Clebsch-Gordan Construction of Lattice Interpolating Fields for Excited Baryons

Bottom line: this is the paper you want on your desk if you are building correlation matrices for excited baryons, and the tables hold up better than the second-pass note claims. The note's central objection does not land, and I want to say that plainly before listing what is actually wrong with the paper, because the wrong objection is the loudest one here and it would be a mistake to let it drive the verdict. On the stress test: Table XVII states that entries are sign×|C|², so the magnitudes, not the signed values, must sum to one for a fixed input pair. I checked the H⊗H rows. H1H4 gives 1/4 + 1/4 + 9/20 + 1/20 = 1. H2H2 gives 4/10 + 9/40 + 3/8 = 1. H2H3 gives 1/4 + 1/4 + 1/20 + 9/20 = 1. H4H4 gives 25/40 + 3/8 = 1. The note subtracted the negative entries instead of taking their magnitudes, which is exactly what the stated convention forbids. So the H⊗H table is internally consistent, and the claim that it is not should be dropped. What is actually new: the systematic enumeration of quasi-local operators for every flavor channel, the ρ-spin/s-spin bookkeeping in Appendix B, and the one-link construction with explicit Clebsch-Gordan coefficients. The ρ-parity and smearing-parity decomposition in Section II.D is the cleanest part of the paper; it turns the gerade/ungerade bookkeeping into something mechanical rather than ad hoc, and that is the piece I would actually reuse. Real soft spots, in order of how much they cost a user. First, Appendix D has transcription slips. The E⊗T2 table lists E2T12 three times where it should read E2T12, E2T22, E2T32, and the T1⊗H table has 0 0 0-1/2-2/5 run together in row 2. These are typos, not structural errors, but they will cost an afternoon. Second, the published version drops most of the Clebsch-Gordan tables; the arXiv version keeps them, so cite the preprint. Third, the unitary transformation from Altmann-Herzig is asserted rather than exhibited. You can check the printed coefficients by hand, but rederiving them means redoing the transformation. That is a real cost, not a flaw. One more thing worth flagging: the linear-dependence argument in Eq. 50 holds only to O(a²), and Table XII counts operators before those dependencies while Table XIII counts after. The text is clear about this, but anyone reading Table XII alone will overcount. Who it is for: lattice practitioners building baryon operator bases, especially anyone going after J = 5/2 and orbital excitations. Not a general HEP audience. Recommendation: send it out. It is a methods paper with explicit, checkable tables, and the tables are mostly right.

The paper's central deliverable is a set of operators that transform irreducibly under ODh, and that property rests entirely on the Clebsch-Gordan coefficients in Appendix D. The reader flags the delegation to Ref. [38] (an unpublished PhD thesis) and the unexhibited unitary transformation from Altmann-Herzig. That is a fair reproducibility concern, but the paper does state in §IV.B that the transformation was performed and that all coefficients are real, and Appendix D does print a large explicit set. So 'cited rather than proved' is not by itself load-bearing here. The sharper, checkable issue is internal: Table XVII states the convention that entries are |C|^2 with a separate sign, and that in each table the resultant IR is the top row while the two input IRs are the left column. The H⊗H table, however, lists 16 result columns (A1, A2, E1, E2, T1 x3, T1 x3, T2 x3, T2 x3) with entries like 25, 9, 40, 20, 8, 4, 2 and a 'denom.' row of 4,4,4,4,10,20,10,40,20,40,8,4,8,2,4,2. For a fixed input pair (H_i H_j), the squared coefficients across all result rows must sum to 1 (each direct product decomposes with unit total weight). Taking H1H1: entries 0,0,0,0,0,0,0,0,0,25,0,0,3,0,0,0 over denominators 4,4,4,4,10,20,10,40,20,40,8,4,8,2,4,2 gives 25/40 + 3/8 = 0.625 + 0.375 = 1.0, which is consistent. But H1H4: 1,0,1,0,0,9,0,0,-1,0,0,0,0,0,0,0 over 4,4,4,4,10,20,10,40,20,40,8,4,8,2,4,2 gives 1/4 + 1/4 + 9/20 + (-1)/20 = 0.25+0.25+0.45-0.05 = 0.90, not 1.0. H2H2: 0,0,0,0,-4,0,0,9,0,0,3,0,0,0,0,0 gives -4/10 + 9/40 + 3/8 = -0.4+0.225+0.375 = 0.20, far from 1.0. Several rows of the H⊗H table therefore do not satisfy the unitarity/normalization that the stated convention requires, which means either the printed entries, the signs, or the denominators are wrong for those rows. Since H⊗H is exactly the product used to build two-link and multi-link operators (the paper's route to T2 and A2 spatial IRs, §IV.D), an error there propagates into the nonlocal operator construction the 5/

The reader's verdict (CONDITIONAL, low correctness risk) is broadly right, and the reader correctly identifies the CG coefficients as the load-bearing dependency. I agree with the direction but not with the framing. The reader's stated weakest assumption is that the transformed coefficients might be wrong because the transformation is not exhibited and the complete set is partly in an unpublished thesis. That is a reproducibility complaint, and the paper does address it in part: §IV.B says the unitary transformation was performed and all coefficients made real, and Appendix D prints an extensive explicit set. Delegating a remainder to a thesis is a real limitation but not, by itself, a correctness failure. The more concrete and checkable concern is internal consistency of the printed tables against the paper's own stated convention in Table XVII. Working through the H⊗H block, several input-pair rows do not sum to unity over the result columns as the convention requires (H1H4 gives 0.90, H2H2 gives 0.20), whereas other rows (H1H1) do sum to 1.0. That pattern is consistent with transcription errors in specific entries or denominators rather than a wholesale failure of the method, but it matters because H⊗H is precisely the product used in §IV.D to generate the T2 and A2 spatial IRs that one-link operators cannot reach, and those are the operators the paper advertises for higher-spin and orbital-angular-momentum states. So the concern is narrower than the reader's but lands on the same tables and is directly testable by a sum-rule check that does not require the unpublished thesis. I would keep the verdict at CONDITIONAL: the quasi-local tables (VI–IX) and the A1/T1/E one-link tables (X, XI) appear internally consistent and are the bulk of what practitioners will use, so the paper's core utility survives; but the H⊗H block and the two-link extension should be flagged for correction or independent verification before use. This is a correction-level issue, not a rejection-level one, and it is exactly the kind of thing a conditional verdict is for.

If the tables and coefficients are correct, this is a genuinely useful reference for the lattice baryon-spectroscopy community: it provides a large, explicitly tabulated operator basis with definite ODh quantum numbers and a physically motivated correspondence to continuum |J, Jz⟩ states, which is exactly what is needed to build correlation matrices for excited and higher-spin baryons. The construction is parameter-free and derived from group theory, and the paper cites an external numerical orthogonality check (Ref. [4]). The value of the paper rests on the correctness and reproducibility of the Clebsch-Gordan coefficients and the operator tables, so those are the appropriate focus of scrutiny.

Lattice QCD can only extract a resonance mass if the operator used to create the state has a big overlap with it, and for excited baryons nobody knows in advance which shape of operator that is. The practical answer is to build a large matrix of operators and let diagonalising the correlation matrix pick out the mass eigenstates. This paper supplies the raw material for that programme: systematic tables of three-quark operators for every isospin and strangeness channel, each assigned to a definite irreducible representation of the double octahedral group $O_D^h$ — the symmetry group that survives on a cubic lattice for half-integer spin. The authors first classify smeared, local operators by the symmetry of their Dirac indices (using Young tableaux, and a $SU(2)_��otimes SU(2)_s$ split of the Dirac index), then build nonlocal operators by displacing one quark along lattice directions that themselves transform as cubic harmonics, and finally fuse the spatial and spinorial pieces with Clebsch-Gordan coefficients of the octahedral group. The payoff, if the tables are right, is a basis that stays as close as possible to continuum $|J,J_zrangle$ states, so that a lattice practitioner can say which operator should couple to a spin-$5/2$ or spin-$7/2$ baryon rather than merely to 'something in the $H$ channel'.

What carries the argument. Two pieces of machinery carry the construction. The first is the decomposition of a Dirac index into a $rho$-parity label and an ordinary spin label, $q^{rho}_{s}$, so that rotations act only on $s$ while parity acts through the product $rho_1rho_2rho_3$; this turns the problem of reducing three-spinor combinations into a product of two $SU(2)$ reductions, one for $rho$ and one for $s$, and explains why each operator that couples to a gerade state is linked by $rho$-spin raising to one coupling to an ungerade state. The second is the fusion rule $O^{Lambda}_{lambda}=sum Cbinom{Lambda; Lambda_1, Lambda_2}{lambda; lambda_1, lambda_2}, O^{Lambda_1}_{lambda_1}O^{Lambda_2}_{lambda_2}$, applied first to combine two displacement operators into an overall spatial irreducible representation and then again to combine that spatial representation with the spinorial one. Because the lattice harmonics are chosen to match the spherical harmonics $Y_{ell,m}$ for the lowest $ell$, the resulting Clebsch-Gordan coefficients reproduce the continuum $(L=1)otimes(S=1/2)to(J=3/2)$ and $(L=1)otimes(S=3/2)to(J=5/2)$ patterns in the appropriate rows, which is what makes the basis physically interpretable.

Core claim. The claim is a constructive one: for each baryon ($N$, $Delta$, $Lambda$, $Sigma$, $Xi$, $Omega$) the authors enumerate all allowed combinations of three Dirac indices compatible with the required flavour symmetry, reduce them into the $G_1$, $H$ (and, with displacements, $G_2$) irreducible representations of $O_D^h$, and present them as explicit tables of operators with definite parity and spin projection. They then show how to attach gauge-covariant one-link displacements to one quark, group those displacements into $A_1$, $E$ and $T_1$ cubic harmonics, and combine them with the spinorial representations using octahedral Clebsch-Gordan coefficients, producing operators that transform irreducibly overall and couple directly to states of nonzero orbital angular momentum. The counting is made concrete: for the nucleon, quasi-local plus one-link constructions yield 23 $G_{1g}$, 28 $H_g$ and 7 $G_{2g}$ operators, and the number can be extended without limit by adding two-link and multi-link displacements or different smearings. The authors also record linear dependencies that appear after projection to zero momentum — for instance the $T_1$ one-link operators built on mixed-symmetric isospin are equivalent to mixed-antisymmetric ones up to a total derivative, and sixty $T_1$ one-link $Delta$ operators vanish — which keeps the operator sets from being redundantly large.

The premise it rests on. Everything nonlocal rests on the octahedral Clebsch-Gordan coefficients having been correctly transformed from the published tables into the authors' own basis conventions; the transformation itself is not exhibited, and the full coefficient set is partly deferred to an unpublished thesis.

Practitioners can assemble correlation matrices of arbitrary size for each symmetry channel, with the guarantee that operators from different irreducible representations or different rows do not mix in a zero-momentum correlator.

Operators coupling to spin-$5/2$ and higher can be written down directly, since $G_2$ representations — absent from purely local three-quark operators — appear once one-link $E$ or $T_1$ displacements are combined with $H$ spinorial structures.

The tables give a common language between the hand-construction approach taken here and the automated group-projection approach pursued elsewhere, so the two can be cross-checked operator by operator.

Because the displacement combinations are tied to spherical harmonics, the same tables can be reused for two-link and multi-link operators, and for any cubically symmetric smearing substituted for the $A_1$ combination.

The linear-dependency counts after zero-momentum projection tell simulators which operators to drop, avoiding singular correlation matrices built from redundant rows.

The construction is a template as much as a result: once a practitioner accepts the fusion rule, the same recipe generates operators for multi-hadron channels and for baryons with nonzero momentum, which the paper explicitly leaves out of scope.

Whether these operators actually overlap strongly with the excited states they are designed for is an empirical question the tables cannot answer; the useful test is not symmetry bookkeeping but the size of off-diagonal correlators in a real ensemble.

The reliance on continuum-like $|J,J_zrangle$ labels suggests a practical diagnostic: an operator labelled spin-$5/2$ should show a distinctive pattern of couplings across the $H$ and $G_2$ channels, which a lattice calculation could check independently of any mass extraction.

Since the coefficients were transformed from a published point-group table into the authors' own basis, an independent reimplementation of the transformation would be a cheap and decisive cross-check on the whole nonlocal sector.

This review was written by a machine. No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.

Request a human review