Gradient properties of $\varphi^3$ in $d=6-\varepsilon$

Lorenzo Benfatto; Omar Zanusso

REVIEW 4 major objections 4 minor 1 cited by

Gradient properties of $\varphi^3$ in $d=6-\varepsilon$

T0 review · 4 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The renormalization group flow of a general multiscalar φ³ theory remains a gradient flow at d = 6 − ε, but only if the three-loop beta-function coefficients satisfy a new linear constraint that has no analogue at d = 6.

desk verdict The new d=6−ε constraint is numerically plausible and probably right, but the printed derivation is internally inconsistent, so the paper needs major revision before the result can be trusted. read the letter →

arxiv 2507.20761 v3 pith:6MTT4KKM submitted 2025-07-28 hep-th

Lorenzo Benfatto , Omar Zanusso This is my paper

classification hep-th

keywords renormalizationgroupflowgradientstructureφ³theoryεexpansionMSschemeintegrabilityconstraintsA-functionF-function

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

The paper claims that the renormalization-group flow of a general multiscalar φ³ theory, which is known to have a gradient structure exactly in d = 6, continues to have that gradient structure slightly below the upper critical dimension, at d = 6 − ε. The proof works by expanding the gradient condition ∂_I A = T_{IJ} B^J order by order in the ε expansion and eliminating the unknown coefficients of A and the tensor T. This produces one new linear constraint among the three-loop $\beta$-function coefficients, Eq. (22), that has no counterpart in d = 6; the published MS values satisfy it exactly. The result matters because it extends the 'A-function' picture to non-integer dimensions and strengthens the case that the F-function, rather than the topological anomaly alone, governs the irreversibility of the flow. The authors note that the analysis is carried out in the MS scheme and that a general scheme-dependence study is left to future work.

What carries the argument

The central object is the gradient (integrability) equation ∂_I A = T_{IJ} B^J, where A is a scalar function of the tensor couplings g_{ijk}, B^J are 'B-functions' related to the $\beta$ functions by a rotation induced by the antisymmetric part of the anomalous dimension, and T_{IJ} is a tensor in the space of couplings whose symmetric part can be read as a metric. The paper expands this equation simultaneously in the loop order and in ε, writing the coefficients of A and T as regular power series in ε whose ε = 0 boundary values are the known d = 6 solutions. The new constraint (22) emerges when the unknown ε-corrections of the next-to-leading-order coefficients are eliminated; demanding that the T-coefficients have no 1/ε poles gives an equivalent formulation. The specific new identity is a linear combination of three-loop coefficients of the gamma function, the antisymmetric part s_{3,a}, and the $\beta$ function that must vanish for the flow to be a gradient flow at order $g^{9}$ in d = 6 − ε.

What would settle it

Take any multiscalar φ³ theory with N ≥ 3 and compute its three-loop beta function in a different renormalization scheme, for example a momentum-subtraction scheme, at d = 6 − ε; then check whether the transformed coefficients satisfy Eq. (22) (or its general form, Eq. (A12)). A scheme in which the d = 6 integrability constraints still hold but the new relation fails would show that the d = 6 − ε gradient structure is scheme-dependent. Alternatively, a direct diagrammatic calculation of the coefficient s_{3,a} that does not borrow from the d = 6 integrability equations would independently test the numerical validity of Eq. (22).

Watch

Extended reading notes

Core claim

At d = 6 − ε the leading term of the B-functions is $B^{{(0)}}$ = −(1/2)ε g, so the expansion of the gradient equation acquires a new term $T^{{(n)}}$ * $B^{{(0)}}$ that was absent at d = 6. Demanding that the coefficient functions of A and T be regular series in ε turns the gradient condition into linear constraints on the $\beta$-function coefficients. At next-to-leading order the authors find exactly one new independent constraint, Eq. (22): $$36 c_{3,a} - 1728 s_{3,a} - 144 c_{3,b} + 36 b_{3,f} - 144 b_{3,g} + 72 b_{3,h} - 432 c_{3,d} + 216 c_{3,e} + 5184 c_{3,g} + 432 c_{3,c} - 7 = 0.$$ The MS coefficients from Eq. (A8) satisfy it, and the constraint is not a linear combination of the ten constraints that already follow from integrability at d = 6. Hence, in the MS scheme, the flow of B is integrable at d = 6 − ε, and the existence of an A-function below the critical dimension is not an automatic consequence of the d = 6 result—it requires an extra relation specific to the ε extension.

Load-bearing premise

The derivation is done entirely in the MS renormalization scheme, and the paper explicitly postpones studying whether the gradient structure survives in other schemes, so the central claim could collapse if the new constraint turned out to be an artifact of MS subtraction rather than a property of the flow.

Editorial extensions

If this is right

The MS flow of B in the multiscalar φ³ model admits an A-function and a metric G built order by order in 6 − ε, so the gradient-structure description extends below the upper critical dimension.
The three-loop coefficients of the MS B-function must satisfy 1 + 10 + 1 = 12 independent linear constraints, one of which is new and specific to the ε extension.
If the identification with the sphere free energy is right, this opens a perturbative route to proving the F-theorem for the cubic model in the ε expansion, analogous to what has been done for φ⁴ in 4 − ε.
The scalar curvature of the coupling-space metric computed from the gradient structure grows as R ∝ N^5 in the large-N limit, with a leading coefficient that is independent of the scheme ambiguities α_1 and α̃_1.

Reading between the lines

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

An immediate extension is to ask whether at four loops a second ε-dependent constraint appears or whether the pattern stops; a single new constraint per loop order would suggest a systematic form of the A-function in the ε expansion.
Because s_{3,a} enters Eq. (22) with the largest coefficient but was fixed in the literature using the d = 6 integrability equations themselves, an independent diagrammatic computation of s_{3,a} would test both the beta-function values and the claim that the new relation is genuinely independent.
The regularity requirement (absence of 1/ε poles in the metric coefficients) is a general criterion that could be applied to other models, such as O(N) cubic models or other couplings, to locate which of them admit a gradient structure away from criticality.
If scheme dependence turns out to be significant, the physically meaningful statement would have to be formulated scheme-independently, for instance by fixing α_1 and α̃_1 through a sphere computation that matches A with F̃ and G with Zamolodchikov's metric, as the paper suggests.

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

One genuinely new thing here: Eq. (22), the extra three-loop constraint needed for integrability of the phi^3 RG flow in d=6−ε. I checked it against the MS values in (A8); it vanishes exactly. That is a real check, and the extension of the Pannell–Stergiou construction from phi^4 to phi^3 is new, as far as I know. The framing is sensible: integrability at d=6 does not automatically descend to the epsilon expansion, and if the A-function is to interpolate to F, the scheme-dependent story has to be addressed. The authors are upfront about deferring scheme dependence, and the curvature discussion is honest about the free parameters. The citation pattern looks right—[3] and [19] are the obvious sources, and the reliance on an unpublished draft is flagged.

The problem is that the derivation as printed is not reproducible. Eq. (14) is not what follows from (A6). Plugging (A1) and (A3) into (14) gives 175/432, not zero; the general formula gives b2,b + c2,a + 6c2,b = 0, which is a different equation. The ten constraints in (A9), said to be satisfied by (A8), are not: the first evaluates to 41/8, the third to 15/4 − 14ζ3. Eq. (A11) uses b2,f, which is never defined. These are not peripheral typos; they are the intermediate equations that are supposed to justify the elimination leading to (22). On top of that, the independence of (22) from the ten d=6 constraints is asserted without showing the rank computation, and s3,a, the coefficient with the largest weight in (22), was fixed in Ref. [3] using one of the d=6 integrability equations, so the new constraint is not a fully independent test of that coefficient.

None of this shows (22) is wrong. The direct substitution is compatible. But compatibility is not derivation, and as it stands the only route the paper offers to its main result is an elimination whose printed steps contradict each other. In a subject where these constraints are cheap to verify algebraically, that is fixable, but it has to be fixed before the claim "we have proven" is justified.

This paper is for people working on the A-function/F-function program and perturbative integrability of multiscalar models. I would send it to a serious referee and ask for major revision: fix the appendix, show the elimination and the rank computation, and clarify the status of s3,a. If those come back clean, the result is worth having. I would not cite it in its current form.

Referee Report

4 major / 4 minor

Summary. The paper claims that the multiscalar cubic scalar field theory in d = 6 - epsilon dimensions admits a gradient (integrable) renormalization-group flow, in the sense of the existence of a scalar function A and a tensor T such that partial_I A = T_IJ B^J. The main new result is a single linear constraint, Eq. (22), on the three-loop MS-scheme coefficients of the B-functions, in addition to the one- and two-loop constraints and the ten three-loop constraints already known in d = 6. The authors verify numerically that the MS values in Eq. (A8) satisfy this new constraint, and they interpret the result as evidence for an F-function interpolation picture. They also discuss the parametric freedom in A and the metric G, and provide a large-N expression for the curvature scalar.

Significance. If the derivation is made fully reproducible, the result is significant and nontrivial: it extends the known d = 6 integrability of the multiscalar phi^3 renormalization group to the epsilon expansion, gives a concrete consistency condition on three-loop MS coefficients, and reinforces the proposed connection between the A-function and the F-function. The numerical compatibility check in the paper is correct: substituting the printed MS values of Eq. (A8) into Eq. (22) gives exactly zero, and the paper is honest in stating that the scheme-dependence analysis is deferred. However, the printed supporting derivation contains several internally inconsistent equations, so the advertised central claim is not independently established by the text as it stands.

major comments (4)

[§II, Eq. (14) and Appendix A, Eq. (A6)] Eq. (14), b2,b + 6c2,a + c2,b = 0, is not the specialization of the general constraint (A6) to the one-loop values (A1). Substituting b1,a = -1 and c1,a = 1/12 into (A6) yields b2,b + c2,a + 6c2,b = 0, with the coefficients of c2,a and c2,b interchanged relative to (14). With the MS values (A3), Eq. (14) evaluates to 175/432, not zero, while the corrected relation is satisfied. This is a load-bearing inconsistency because the two-loop constraint is used in the elimination that leads to Eq. (22).
[Appendix A, system (A9)] The claim that the ten constraints (A9) are satisfied by the MS values (A8) is false as printed. For example, the first equation evaluates to 41/8 and the third to 15/4 - 14 zeta(3) when the values of (A8) are inserted, neither of which vanishes. Since these constraints are the stated d = 6 consistency conditions used to simplify the epsilon-expansion problem, the text must be corrected, and the corrected system must be verified before the central elimination can be accepted.
[Appendix A, Eq. (A11)] Eq. (A11) contains the coefficient b2,f, which is not defined anywhere in the two-loop data of Eq. (11) or (A3); the two-loop coefficients are b2,a, b2,b, and b2,c. This makes the claimed simplification of the new constraint unusable and further blocks reproduction of the derivation.
[§III and §IV; Eq. (22)] The coefficient s3,a enters Eq. (22) with the largest weight, -1728, but it was not computed directly in this paper. The text states that its MS value was fixed indirectly in Ref. [3] using one of the d = 6 integrability equations, with the direct determination relegated to the unpublished draft Ref. [19]. The numerical check of Eq. (22) therefore depends on an external, partially unpublished input, and an independent check of the new constraint requires that input. The authors should either provide the direct determination of s3,a or clearly separate the verification of Eq. (22) from the derivation that establishes it as an independent constraint.

minor comments (4)

[Introduction, Eq. (2) and surrounding text] There are typographical errors: "tradpoles" should be "tadpoles" and "unintresting" should be "uninteresting".
[§III, Eq. (21)] The notation ~a'_{n,x}(0) is used without an explicit definition; it should state that the prime denotes differentiation with respect to epsilon and the argument zero means evaluation at epsilon = 0.
[§IV, Eqs. (23)-(24)] The footnote thanking a colleague for pointing out mistakes in an earlier version of Eq. (24) is unusual in a formal journal report and should be removed or moved to the acknowledgments; the final published version should present only the corrected result.
[References] Ref. [19] is repeatedly cited as an unpublished draft; if it remains unpublished, the authors should provide a more complete citation or a preprint number, since the direct determination of s3,a is load-bearing for the verification of Eq. (22).

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: Eq. (22) is a derived constraint, not a fitted input; internal algebraic issues are correctness concerns, not circularity.

full rationale

The central result, Eq. (22), is presented as a linear necessary condition obtained by imposing the gradient condition ∂_I A = T_IJ B^J order by order in the ε-expansion and eliminating the A and G coefficients; it is not defined in terms of the MS beta-function coefficients it constrains. The general form (A12) and the one-loop/two-loop inputs (A1), (A3) show that the constraint is not fitted to the three-loop values, and the check against (A8) is an external numerical test, not the derivation. The only provenance concern is s3,a, whose MS value was 'determined indirectly in Ref. [3] (using the last equation of the system (A9))'; since s3,a enters (22) with the largest coefficient, the numerical check is not fully independent for that coefficient, but the constraint itself is not constructed from that value. The unsupported parts of the paper are algebraic/reproducibility defects—Eq. (14) does not follow from (A6), the equations in (A9) are not satisfied by (A8) as printed, and (A11) uses the undefined coefficient b2,f—and these are correctness risks, not circular reductions. No self-definitional step, fitted input called prediction, self-citation chain, or renaming of a known result is exhibited.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The central claim rests on imports: the d = 6 gradient structure from the trace-anomaly/local RG framework, the three-loop MS coefficients of Refs. [3,4] (with s3,a's value fixed in part by integrability itself or by unpublished Ref. [19]), the regularity in epsilon of the series solutions for A(n) and G(n), and the conjectural interpolation of A by the sphere free energy F. Against those inputs the paper's own contribution is a linear-elimination computation producing one new constraint. The free parameters alpha_1 and tilde_alpha_1 are left undetermined, so the A and metric solutions are a family, and the curvature (24) inherits this ambiguity at subleading orders.

free parameters (2)

alpha_1 (a2,c) = not fixed (free)
Parameter of the two-loop A function (Eq. A4) not fixed by integrability; it enters the epsilon corrections of A(1) via Eq. (A7) and the curvature (24). The paper notes A can be redefined by alpha T_IJ B^I B^J, so the physical value awaits a sphere matching.
tilde_alpha_1 (m1,c + a2,c) = not fixed (free)
Parameter of the one-loop metric G(1) (Eq. A5), not fixed by integrability; it affects the curvature scalar (24) at subleading orders in the large-N expansion. Fixing it requires an unmotivated matching of G to Zamolodchikov's metric.

assumptions (5)

domain assumption The local RG / trace-anomaly framework establishes the gradient structure (Eq. 3) in d = 6.
Invoked in the Introduction via Refs. [5,6,7,8,10]; the paper relies on this structure as the boundary condition for the epsilon expansion.
domain assumption A is identified, up to normalization, with the topological trace-anomaly coefficient in d = 6, interpolated by the sphere free energy F at d = 6 − epsilon.
Motivating premise in Section I and the Conclusions; the interpolation away from the critical dimension is a conjecture whose validity is the very property under test.
domain assumption The three-loop MS beta, gamma and upsilon coefficients from Refs. [3,4] are correct.
Input data listed in Eq. (A8). The printed constraint systems in the paper do not all check against these values, so the values themselves rest on the external references.
ad hoc to paper The epsilon-series coefficients of A(n) and G(n) are regular in epsilon, with boundary values at epsilon = 0 equal to the d = 6 solution.
Section III: the new constraint (22) is obtained by demanding regularity, i.e., no 1/epsilon poles, in the solution for the order-epsilon coefficients. This assumption selects the notion of integrability in d = 6 − epsilon; without it the constraint would not be forced.
standard math Perturbative renormalizability and MS-scheme validity in d = 6 − epsilon.
Background for the epsilon expansion, standard in the field and cited to Refs. [1,2].

how reviews work

0 comments

Cite this review

Pith. "Pith review of Gradient properties of $\varphi^3$ in $d=6-\varepsilon$." pith.science (2026). https://pith.science/paper/6MTT4KKM

@misc{pith2026250720761,
  author       = {Pith},
  title        = {Pith review of: Gradient properties of $\varphi^3$ in $d=6-\varepsilon$},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6MTT4KKM}},
  note         = {Machine review of arXiv:2507.20761}
}

abstract

The renormalization group flow of the multiscalar interacting $\varphi^3$ theory in $d=6$ dimensions is known to have a gradient structure, in which suitable generalizations of the beta functions $B^{I}$ emerge as the gradient of a scalar function $A$, $\partial_I A = T_{IJ} B^J $, with a nontrivial tensor $T_{IJ}$ in the space of couplings. This has been shown directly to three loops in schemes such as $\overline{\rm MS}$ and can be argued in general by identifying $A$ with the coefficient of the topological term of the trace-anomaly in $d=6$ up to a normalization. In this paper we show that the same renormalization group has a gradient structure in $d=6-\varepsilon$. The requirement of a gradient structure is translated to linear constraints that the coefficients of the $\overline{\rm MS}$ beta functions must obey, one of which is new and pertinent only to the extension to $d \neq 6$.

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Gradient RG Flow in Scalar-Fermion QFTs
hep-th 2025-11 conditional novelty 7.0 of 10

RG flow in scalar-fermion theories is gradient through four loops only when the 'beta shift' is included, via over a thousand scheme-independent constraints that hold wherever current data allow a check.

Reference graph

Works this paper leans on

24 extracted references · 13 canonical work pages · cited by 1 Pith paper

[3]

In general, the leading term B(0) of the beta func- tion conspires in Eq

The beta func- tion of the dimensionless coupling must be complemented by the leading term β(0) ijk = B(0) ijk =− 1 2εgijk, which, in turn, forces the leading term of A to be A(0) =−ε 4 , (19) which solves the “zeroth” order (zeroth order if compared to the loop-expansion). In general, the leading term B(0) of the beta func- tion conspires in Eq. (3) with...

work page
[19]

A. B. Zamolodchikov, JETP Lett. 43, 730-732 (1986)

work page 1986
[1]

M. E. Fisher, Phys. Rev. Lett. 40, 1610-1613 (1978)

work page 1978
[2]

There are 5 irreducible bubbles that contribute to the A function at this order A(2) =a2,a +a2,b +a2,c +a2,d +a2,e

As in the one loop case υ(2) ij = 0 because it is impossible to form antisymmetric contrac- tions also at order g4, which means that β(2) = B(2). There are 5 irreducible bubbles that contribute to the A function at this order A(2) =a2,a +a2,b +a2,c +a2,d +a2,e . (12) A general parametric form of the tensor T (1) IJ is obtained by labelling two independent...

work page
[4]

(A5) The free parameters ˜α1 andα1, discussed in Sect

(A3) 7 The coefficients of the A-function are solved as a2,a =− 1 12, a 2,b = 1 6− 2α1, a 2,c =α1, a2,d =− 7 576− α1 12 , a 2,e = 1 3456− α1 24 , (A4) Similarly for the coefficients of GIJ are solved as m1,a =−7 4 + 12α1, m 1,b =− 7 48−α1, m1,c = ˜α1−α1, m 1,d = 1 6−α1− ˜α1. (A5) The free parameters ˜α1 andα1, discussed in Sect. IV, are unconstrained by i...

work page
[5]

Grinstein, A

B. Grinstein, A. Stergiou and D. Stone, JHEP 11, 195 (2013) [arXiv:1308.1096 [hep-th]]

arXiv 2013
[6]

O. F. de Alcantara Bonfim, J. E. Kirkham and A. J. McKane, J. Phys. A 13, L247 (1980) [erratum: J. Phys. A 13, 3785 (1980)]

work page 1980
[7]

J. A. Gracey, I. Jack and C. Poole, JHEP 01, 174 (2016) [arXiv:1507.02174 [hep-th]]

arXiv 2016

Show all 24 references

[8]

L. Fei, S. Giombi and I. R. Klebanov, Phys. Rev. D 90, no.2, 025018 (2014) [arXiv:1404.1094 [hep-th]]

2014 arXiv
[9]

Herren and A

F. Herren and A. E. Thomsen, JHEP 06, 116 (2021) [arXiv:2104.07037 [hep-th]]

2021 arXiv
[10]

C. W. Poole, PhD thesis, doi:10.17638/03015643

doi
[11]

Grinstein, A

B. Grinstein, A. Stergiou, D. Stone and M. Zhong, Phys. Rev. D 92, no.4, 045013 (2015) [arXiv:1504.05959 [hep- th]]

2015 arXiv
[12]

Stergiou, D

A. Stergiou, D. Stone and L. G. Vitale, JHEP 08, 010 (2016) [arXiv:1604.01782 [hep-th]]

2016 arXiv
[13]

Codello, M

A. Codello, M. Safari, G. P. Vacca and O. Zanusso, Phys. Rev. D 101, no.6, 065002 (2020) [arXiv:1910.10009 [hep- th]]

2020 arXiv
[14]

Jack and H

I. Jack and H. Osborn, Nucl. Phys. B883, 425-500 (2014) [arXiv:1312.0428 [hep-th]]

2014 arXiv
[15]

Osborn and A

H. Osborn and A. Stergiou, JHEP 05, 051 (2018) [arXiv:1707.06165 [hep-th]]

2018 arXiv
[16]

(A2) At two loops the MS values of B(2) in (11) are c2,a = 1 18, c 2,b =− 11 432, b2,a =−1 4, b 2,b = 7 72, b 2,c =−1
[17]

Osborn and A

H. Osborn and A. Stergiou, JHEP 04, 128 (2021) [arXiv:2010.15915 [hep-th]]

2021 arXiv
[18]

Safari, G

M. Safari, G. P. Vacca and O. Zanusso, Eur. Phys. J. C 80, no.12, 1127 (2020) [arXiv:2009.02589 [cond-mat.stat- mech]]

2020 arXiv
[20]

Osborn, Nucl

H. Osborn, Nucl. Phys. B 363, 486-526 (1991)

1991
[21]

L. Fei, S. Giombi, I. R. Klebanov and G. Tarnopolsky, JHEP 12, 155 (2015) [arXiv:1507.01960 [hep-th]]. 9

2015 arXiv
[22]

W. H. Pannell and A. Stergiou, Phys. Lett. B 853, 138701 (2024) [arXiv:2402.17817 [hep-th]]

2024 arXiv
[23]

Jack and H

I. Jack and H. Osborn, unpublished draft
[24]

W. H. Pannell and A. Stergiou, [arXiv:2502.06940 [hep- th]]

arXiv

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [3]

In general, the leading term B(0) of the beta func- tion conspires in Eq

The beta func- tion of the dimensionless coupling must be complemented by the leading term β(0) ijk = B(0) ijk =− 1 2εgijk, which, in turn, forces the leading term of A to be A(0) =−ε 4 , (19) which solves the “zeroth” order (zeroth order if compared to the loop-expansion). In general, the leading term B(0) of the beta func- tion conspires in Eq. (3) with...

work page

[2] [19]

A. B. Zamolodchikov, JETP Lett. 43, 730-732 (1986)

work page 1986

[3] [1]

M. E. Fisher, Phys. Rev. Lett. 40, 1610-1613 (1978)

work page 1978

[4] [2]

There are 5 irreducible bubbles that contribute to the A function at this order A(2) =a2,a +a2,b +a2,c +a2,d +a2,e

As in the one loop case υ(2) ij = 0 because it is impossible to form antisymmetric contrac- tions also at order g4, which means that β(2) = B(2). There are 5 irreducible bubbles that contribute to the A function at this order A(2) =a2,a +a2,b +a2,c +a2,d +a2,e . (12) A general parametric form of the tensor T (1) IJ is obtained by labelling two independent...

work page

[5] [4]

(A5) The free parameters ˜α1 andα1, discussed in Sect

(A3) 7 The coefficients of the A-function are solved as a2,a =− 1 12, a 2,b = 1 6− 2α1, a 2,c =α1, a2,d =− 7 576− α1 12 , a 2,e = 1 3456− α1 24 , (A4) Similarly for the coefficients of GIJ are solved as m1,a =−7 4 + 12α1, m 1,b =− 7 48−α1, m1,c = ˜α1−α1, m 1,d = 1 6−α1− ˜α1. (A5) The free parameters ˜α1 andα1, discussed in Sect. IV, are unconstrained by i...

work page

[6] [5]

Grinstein, A

B. Grinstein, A. Stergiou and D. Stone, JHEP 11, 195 (2013) [arXiv:1308.1096 [hep-th]]

arXiv 2013

[7] [6]

O. F. de Alcantara Bonfim, J. E. Kirkham and A. J. McKane, J. Phys. A 13, L247 (1980) [erratum: J. Phys. A 13, 3785 (1980)]

work page 1980

[8] [7]

J. A. Gracey, I. Jack and C. Poole, JHEP 01, 174 (2016) [arXiv:1507.02174 [hep-th]]

arXiv 2016