Pith. sign in

REVIEW 3 major objections 5 minor 20 references

Nonperturbative refined topological string

T0 review · 3 major / 5 minor · reviewed 2026-08-08 · deepseek-v4-flash

Pith's one-line read A single contour integral is proposed for the full nonperturbative refined topological string free energy, with BPS counts as the only input.

desk verdict A new integral formula for the refined nonperturbative free energy that passes the perturbative check but misses the sign of the Stokes jump, so the central claim needs revision. read the letter →

arxiv 2502.05518 v2 pith:2EZJM67V submitted 2025-02-08 hep-th math-phmath.AGmath.MP

classification hep-thmath-phmath.AGmath.MP MSC 14N3581T30
keywords topologicalstringsrefinedstringnonperturbativefreeenergytrans-seriesStokesautomorphismsresurgenceBPSinvariantsDonaldson-Thomas
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 perturbative refined topological string free energy is an asymptotic series, so the exact free energy must contain exponentially small nonperturbative corrections. This paper proposes that the full nonperturbative free energy of the refined topological string is a single contour integral, Eq. (60), built from refined BPS degeneracies (counts of supersymmetric states) and SU(2) characters, with the parameter $b$ controlling the refinement. Evaluating the residues at different families of poles reproduces both the usual perturbative expansion and the nonperturbative trans-series terms. Rotating the coupling into the complex plane makes the contour pick up exactly the Stokes jumps that resurgence theory predicts. If this is right, the resurgent structure of the refined string is not extra input: it is a residue computation from the same BPS data that gives the perturbative series.

What carries the argument

The load-bearing object is the meromorphic integrand $$\frac{1}{u}\frac{1}{1-$e^{{-2\pi iu}}$}$e^{{-(d\cdot t-2\pi i n_{d\cdot B}}$)u}\frac{\chi_{j_L}($e^{{iu\lambda(b+b^{-1}}$)/2})\chi_{j_R}($e^{{iu\lambda(b-b^{-1}}$)/2})}{4\sin(ub\$\lambda$/2)\sin(u\$\lambda$/2b)}.$$ The factor $1/(1-e^{-2\pi iu})$ encodes an infinite geometric sum over the integer D0-brane charge; the sine denominators encode the two $\Omega$-background graviphoton couplings $b\lambda$ and $-\lambda/b$; the characters $\chi_j$ encode the spin of the BPS multiplet. The mechanism is to evaluate the contour integral by residues and to group the poles by family: the integer-$u$ family gives the perturbative refined expansion, while the families at $u=2\pi b\ell/\lambda$ and $u=2\pi\ell/(\lambda b)$ give the trans-series. When $\lambda$ is moved into the complex plane, the contour picks up extra residue contributions that are the Stokes automorphisms.

What would settle it

Compute the trace in Eq. (47) directly from a microscopic $\Omega$-background or M2-brane index calculation and compare the character combination with the paper's expression; a mismatch would change the pole structure and the trans-series. Alternatively, for a local toric Calabi-Yau with $b\neq 1$, numerically Borel-resum the perturbative refined free energy to high genus and check that the discontinuities across the two Stokes rays agree term by term with the residues of (60).

Watch

Extended reading notes

Core claim

The paper argues that the full nonperturbative refined topological string free energy on a Calabi-Yau threefold is $$F_{\mathrm{ref,full}}=\sum_{d,j_L,j_R} N^d_{j_L,j_R}\oint_C \frac{du}{u}\frac{1}{1-$e^{{-2\pi iu}}$}$e^{{-(d\cdot t-2\pi i n_{d\cdot B}}$)u}\frac{\chi_{j_L}($e^{{iu\lambda(b+b^{-1}}$)/2})\chi_{j_R}($e^{{iu\lambda(b-b^{-1}}$)/2})}{4\sin(ub\$\lambda$/2)\sin(u\$\lambda$/2b)},$$ where $N^d_{j_L,j_R}$ counts BPS states with class $d$ and spin quantum numbers $(j_L,j_R)$, $\chi_j(y)=(y^{2j+1}-y^{-2j-1})/(y-y^{-1})$ is the spin-$j$ SU(2) character, and $b$ is the refinement parameter with $b=1$ recovering the unrefined string. The factor $1/(1-e^{-2\pi iu})$ sums over the D0-brane charge $n$, and the contour $C$ is chosen to circle the poles on the positive real axis. Residues at integer $u$ give the perturbative refined expansion (62); residues at $u=(2\pi b/\lambda)\ell$ and $u=2\pi\ell/(\lambda b)$ give the nonperturbative trans-series; complexifying $\lambda$ and letting the contour cross Stokes rays gives the Stokes automorphisms of [AMP24].

Load-bearing premise

The whole formula rests on one assumed trace: how a BPS multiplet with two spin labels responds to the background field, stated in Eq. (47) without derivation; if that response differed, the pole locations and the nonperturbative corrections would change.

Editorial extensions

If this is right

  • The full refined free energy is determined by the refined BPS degeneracies $N^d_{j_L,j_R}$ alone; no separate nonperturbative constants are introduced.
  • Setting $b=1$ reduces the contour formula to the unrefined full free energy, recovering ordinary topological string theory as a special case.
  • The Borel singularities at $\ell b^{\pm1}A_{d,n}$ and their Stokes constants are identified with pole residues, so resurgence data become geometric data of the integrand.
  • The perturbative part matches the refined topological vertex / refined BPS expansion, providing a consistency check with existing A-model computations.
  • The trans-series sum (69) equals the logarithm of the quantum dilogarithm appearing in the Stokes automorphism, confirming the wall-crossing interpretation of the jumps.

Reading between the lines

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

  • A direct reading of (60) is that the perturbative and exponential terms are residues of one integrand, so any construction that adds independent nonperturbative data to the refined string would be redundant.
  • The pole structure suggests that crossing a Stokes ray in the coupling is equivalent to crossing a wall in the BPS state space, tying refined resurgence to Donaldson-Thomas wall-crossing.
  • Inserting a mass parameter or Wilson line into the integrand would predict deformed trans-series from the new pole locations; comparing those predictions with the dual partition function or tau-function constructions would be a test of the formula beyond the cases treated here.
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.

Referee Report

3 major / 5 minor

Summary. The paper proposes a contour-integral formula, Eq. (60), for the full nonperturbative refined topological string free energy, extending the Hattab-Palti formula to the refined setting. The derivation starts from an integrating-out computation of M2-branes in the Omega background, with a trace over the (jL, jR) BPS multiplet stated in Eq. (47). The paper then shows that the perturbative residues of the integral reproduce the refined Gopakumar-Vafa expansion, Eq. (62), and that the additional poles are located at the Borel singularities identified in [AMP24]. The main claim is that the formula reproduces the trans-series structure and Stokes automorphisms of the refined topological string. The comparison is carried out in §4.2, where the residue sum is matched to the AMP24 trans-series up to a sign.

Significance. If correct, Eq. (60) would provide a compact, all-orders nonperturbative definition of the refined topological string free energy in terms of refined BPS invariants, and would connect the BPS/DT data to the resurgence structure in a direct way. The perturbative check in Eq. (62) is a genuine and useful consistency test, and the identification of the nonperturbative pole locations with the Borel singularities of [AMP24] is a nontrivial structural match. However, the central claim that the formula captures the Stokes automorphisms is not established as written because of the sign discrepancy recorded in §4.2, and the physical input in Eq. (47) is assumed without derivation. These issues are local and potentially fixable, so they warrant a major revision rather than rejection.

major comments (3)
  1. [§4.2, Eq. (69)] The central comparison with [AMP24] has a sign mismatch. The residue computation in Eq. (69) gives +i sum_{d,j} Omega[j](d) log Phi[j]_b(-A_{d,n}/(2 pi lambda)), whereas the trans-series in Eq. (33) is -i times the same expression, and the Stokes automorphism in Eq. (35) acts by multiplication by Phi^{-Omega}. Because the Stokes factor is exponentiated, replacing -i log Phi by +i log Phi gives the inverse of the claimed discontinuity. The manuscript itself states at the end of §4.2 that the result coincides with [AMP24] 'up to a minus sign,' but the abstract and §5 assert that the Stokes automorphisms are reproduced. Please correct the sign, either in the residue evaluation, in the contour orientation in Eq. (59), or in the normalization of F_ref,full, and then re-derive the jump; alternatively, state explicitly a convention under which Eq. (35) has the opposite sign.
  2. [§3, Eq. (47)] Equation (47), the trace over the (jL, jR) multiplet in the Omega background, is stated without derivation and is the physical input on which the whole integrand (48) rests. The formula fixes the coupling of the graviphoton to the two SU(2) factors, namely e tilde H ~ i z (b + b^{-1}) lambda J_L + (b - b^{-1}) lambda J_R, as well as the overall sign (-1)^{jL+jR}. A different coupling or fermion-number assignment would change the poles and hence the trans-series. Please provide either a derivation of this trace from the Omega-background/M2-brane computation or a precise reference where this refined trace is computed, and explain why the fermion number is 2(J_L + J_R).
  3. [§4.1, after Eq. (60)] The assertion that formula (60) is valid 'even if the coupling constant lambda is not real' is not justified in the text. The contour manipulations leading to Eq. (59) use lambda in R_+ so that the poles lie on the positive real axis. The later discussion in §4.2 rotates the nonperturbative poles by taking lambda complex, which requires an analytic-continuation argument, such as a precise definition of the integration contour for complex lambda and a demonstration that no other contributions appear. Please supply this argument or restrict the claim to the value obtained by analytic continuation of the real-lambda contour integral.
minor comments (5)
  1. [Abstract and §5] The abstract and the conclusion state that Eq. (60) 'captures the Stokes automorphisms' and 'reproduces the trans-series structure,' but §4.2 explicitly records a minus-sign discrepancy. Please qualify these statements until the sign issue is resolved.
  2. [§3, before Eq. (46)] There is a typo: 'grviphoton' should be 'graviphoton'.
  3. [§1, Eq. (1)] 'zero convergence radius' should be 'zero radius of convergence'.
  4. [§4.2, Eq. (64)] The angles theta_{n_d dot B + k} are used before they are defined. Please state how these angles are determined from the arguments of e^{-(d dot t - 2 pi i n_d dot B) u} for the relevant values of n.
  5. [§2.3, Eq. (29)] The equality between the logarithmic expansion of Phi[j]_b(z) and the integral representation in Eq. (29) is stated without derivation; a brief indication of how the character enters the integral would improve readability.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the proposed contour integral is independently motivated and the comparison with AMP24 is a genuine residue computation; the sign discrepancy noted near Eq. (69) is a correctness issue, not circularity.

full rationale

The paper's central object, formula (60), is not obtained by fitting the perturbative refined Gopakumar-Vafa expansion or the AMP24 trans-series. The integrand structure, including the sine denominators and SU(2) characters, is motivated by the Omega-background integrating-out calculation, with the trace over (jL,jR) multiplets stated as an assumption in Eq. (47). The perturbative reproduction in Eq. (62) is a genuine residue computation from the proposed contour integral rather than an insertion of the known expansion, and the nonperturbative poles at u = 2πb/λ l and u = 2π/(λb) l follow from the same integrand, so their agreement with the Borel singularities of AMP24 is a substantive check. The self-citations in the paper, such as [CDP14], are used for background or consistency and are not load-bearing for the proposed formula. The trace formula (47) is an unproven assumption, and the sign mismatch between Eq. (69) and Eq. (33) is an explicit correctness defect, but neither constitutes a reduction of the claimed result to its inputs by definition or by fitted parameters. Accordingly, the paper is not circular.

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

The paper introduces no free parameters and no new entities. The central claim rests on the physical assumption that the M2-brane integrating-out result can be continued to a full nonperturbative formula, on the specific Omega-background trace formula (47), and on the AMP24 characterization of Borel singularities and Stokes automorphisms. The quantum dilogarithm machinery is standard.

assumptions (5)
  • domain assumption The refined topological string free energy admits the Gopakumar-Vafa integrality expansion (11) with integers N^d_{jL,jR}.
    Used throughout to justify the integrand of the proposed formula; cited to IKV09 and CDP14.
  • domain assumption The M2-brane integrating-out computation yields the full nonperturbative free energy, not just the perturbative expansion, as claimed in HP24.
    This is the physical basis for proposing the contour integral formula (60).
  • ad hoc to paper The trace over the (jL,jR) multiplet in the Omega background gives formula (47).
    This is a new technical step in Section 3, stated without derivation; the final formula depends on it.
  • domain assumption The Borel singularities are at l/b A_d,n and l b A_d,n, and the Stokes automorphism is (35), as found in AMP24.
    Used in Section 4 to identify the nonperturbative poles and to compare with the known trans-series.
  • standard math The quantum dilogarithm identities (24)-(29) from Faddeev, Kashaev and Volkov.
    Used to sum the exponential series into the log Phi form in Eqs. (32) and (69).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Nonperturbative refined topological string." pith.science (2026). https://pith.science/paper/2EZJM67V

@misc{pith2026250205518,
  author       = {Pith},
  title        = {Pith review of: Nonperturbative refined topological string},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2EZJM67V}},
  note         = {Machine review of arXiv:2502.05518}
}
read the original abstract

A formula for the full nonperturbative topological string free energy was recently proposed by Hattab and Palti \cite{HP24a}. In this work, we extend their result to the refined topological string theory. We demonstrate that the proposed formula for the full nonperturbative refined topological string free energy correctly reproduces the trans-series structure of the refined topological string and captures the Stokes automorphisms associated with its resurgent properties.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

20 extracted references · 18 canonical work pages

  1. [1]

    Alexandrov, M

    S. Alexandrov, M. Mari\ no and B. Pioline, Resurgence of Refined Topological Strings and Dual Partition Functions , SIGMA vol.20, 073 (2024)

  2. [2]

    Bridgeland, Riemann-Hilbert problems from Donaldson-Thomas theory , Invent

    T. Bridgeland, Riemann-Hilbert problems from Donaldson-Thomas theory , Invent. Math. 216 , no. 1, 69-124 (2019)

  3. [3]

    Bridgeland, Riemann-Hilbert problems for the resolved conifold , J

    T. Bridgeland, Riemann-Hilbert problems for the resolved conifold , J. Differential Geom. 115, no. 3, 395-435 (2020)

  4. [4]

    Chuang, D

    W.-Y. Chuang, D. E. Diaconescu and G. Pan, BPS states and the P=W conjecture , Lond. Math. Soc. Lect. Note Ser. 411, 132-150 (2014)

  5. [5]

    Coman, E

    I. Coman, E. Pomoni and J. Teschner, From Quantum Curves to Topological String Partition Functions , Commun. Math. Phys. 399, no.3, 1501-1548 (2023)

  6. [6]

    Couso-Santamaría, J

    R. Couso-Santamaría, J. Edelstein, R. Schiappa and M. Vonk, Resurgent Transseries and the Holomorphic Anomaly , Ann. Henri Poinca\'e, Vol 17, 331–399, (2016)

  7. [7]

    Dedushenko and E

    M. Dedushenko and E. Witten, Some Details On The Gopakumar-Vafa and Ooguri-Vafa Formulas , Adv. Theor. Math. Phys. 20, 1-133 (2016)

  8. [8]

    \'Ecalle, Les fonctions r\'esurgentes , Publ

    J. \'Ecalle, Les fonctions r\'esurgentes , Publ. Math. d'Orsay/Univ. de Paris, Dept de Math. (1981)

Show all 20 references
  1. [9]

    L. D. Faddeev, Discrete Heisenberg-Weyl group and modular group , Lett. Math. Phys. 34, 249-254 (1995)

  2. [10]

    L. D. Faddeev, R. M. Kashaev and A. Y. Volkov, Strongly coupled quantum discrete Liouville theory. 1. Algebraic approach and duality , Commun. Math. Phys. 219, 199-219 (2001)

  3. [11]

    Gopakumar, and C

    R. Gopakumar, and C. Vafa, On the Gauge Theory/Geometry Correspondence , Adv. Theo. Math. Phys., 3(5), 1415-1445,1999

  4. [12]

    J. Gu, A. K. Kashani-Poor, A. Klemm and M. Marino, Non-perturbative topological string theory on compact Calabi-Yau 3-folds , SciPost Phys. 16, no.3, 079 (2024)

  5. [13]

    Hattab and E

    J. Hattab and E. Palti, Non-perturbative topological string theory on compact Calabi-Yau manifolds from M-theory , arXiv:2408.09255 [hep-th]

  6. [14]

    Hattab and E

    J. Hattab and E. Palti, Notes on integrating out M2 branes , arXiv:2410.15809 [hep-th]

  7. [15]

    M. X. Huang, S. Katz, and A. Klemm, Towards refining the topological strings on compact Calabi-Yau 3-folds , JHEP 03 (2021) 266

  8. [16]

    M. X. Huang, S. Katz, A. Klemm and X. Wang, Refined BPS numbers on compact Calabi-Yau 3-folds from Wilson loops , arXiv:2503.16270[hep-th]

  9. [17]

    M. X. Huang and A. Klemm, Direct integration for general backgrounds , Adv. Theor. Math. Phys. 16, no.3, 805-849 (2012)

  10. [18]

    Iqbal, C

    A. Iqbal, C. Kozcaz and C. Vafa, The Refined topological vertex , JHEP 10, 069 (2009)

  11. [19]

    Maulik and Y

    D. Maulik and Y. Toda, Gopakumar-Vafa invariants via vanishing cycles , Invent. Math. 213 no. 3, (2018) 1017–1097

  12. [20]

    Nekrasov and A

    N. Nekrasov and A. Okounkov, Membranes and sheaves , Algebr. Geom. 3 no. 3, (2016) 320–369

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.