Pith. sign in

REVIEW 1 major objections 5 minor 52 references

Three-loop form factors for Higgs boson pair production in the large top mass limit

T0 review · 1 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read This paper computes five expansion terms for the three-loop box-type form factors and eight for the triangle form factor in large-top-mass Higgs boson pair production, giving analytic results in the soft-virtual approximation.

desk verdict Genuinely new three-loop large-top-mass form factor coefficients for gg→HH, solidly presented with the usual caveat that the new box-type terms lack independent confirmation. read the letter →

arxiv 1909.01361 v1 pith:VVUMI6JN submitted 2019-09-03 hep-ph

Joshua Davies , Matthias Steinhauser This is my paper · ORCID
classification hep-ph
keywords Higgsbosonpairproductionthree-loopformfactorslargetopmasslimitasymptoticexpansionNNLOcorrectionsgluonfusionsoft-virtualapproximationanalyticresults
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

Higgs boson pair production via gluon fusion is the main way to probe the Higgs self-coupling, but at the three-loop level the full amplitude is too hard to compute exactly. This paper extends the large-top-mass expansion of the three-loop virtual amplitude, computing five expansion terms (through $1/m_t^{8}$) for the two box-type form factors and eight terms (through $1/m_t^{14}$) for the triangle form factor. The triangle form factor is the part tied to the triple-Higgs coupling, which makes the deeper triangle expansion directly relevant to the physics goal. The paper presents finite, analytic expressions after ultraviolet renormalization and infrared subtraction, and explains the computational restructuring that made the calculation feasible.

What carries the argument

The argument rests on two techniques. The first is the asymptotic expansion in $1/m_t^2$ defined by taking $m_t$ much larger than every external momentum: it splits each three-loop diagram into hard subgraphs, which are expanded in the external momenta and the co-subgraph loop momenta, and co-subgraphs, reducing the integrals to products of vacuum integrals and massless one- or two-loop integrals with at most one additional scale. The second is a projection method: each diagram is written as a polynomial in the external scalar products, and derivative operators extract the coefficients one at a time, so that three-loop tensor vacuum integrals never have to be computed directly. Applying the derivatives after the $1/m_t^2$ expansion, and storing the expanded super-diagrams in compressed form, keeps the huge intermediate expressions manageable. This is what allows the fifth-order box expansion and the eighth-order triangle expansion to be obtained.

What would settle it

One concrete test is to compute the next order in the expansion, the $1/m_t^{10}$ box term (or the $1/m_t^{16}$ triangle term), and check that it is suppressed relative to the deepest computed term by roughly the external squared momentum divided by $m_t^2$; failure of that suppression would show the series is not converging where it is used. A future exact numerical evaluation of the three-loop virtual amplitude with finite top mass at center-of-mass energies near 300 to 400 GeV would settle the same question directly.

Watch

Extended reading notes

Core claim

The central claim is that the three-loop amplitude for Higgs boson pair production via gluon fusion, in the limit where the top quark mass is much larger than all external momenta, can now be given analytically to order $1/m_t^{8}$ for the two box-type form factors $F_{box1}$ and $F_{box2}$, and to order $1/m_t^{14}$ for the triangle form factor $F_{tri}$. The box-type results are new; the triangle result extends the previously known expansion to three more orders. The finite forms are obtained by ultraviolet renormalization and by subtracting infrared poles with a standard prescription, leaving finite form factors expressed in terms of the colour factors, the number of light and heavy quark flavours, and the kinematical variables. The analytic results are presented for the soft-virtual approximation and supplied in computer-readable form for direct use in approximation procedures.

Load-bearing premise

The load-bearing premise is that the top quark is heavy enough that a power-series expansion in external momenta divided by the top mass is an accurate representation of the amplitude in the kinematic region where the form factors are applied, which includes energies above the two-top-quark threshold.

Editorial extensions

If this is right

  • The new box-type form-factor terms can be inserted into existing approximation schemes that combine exact next-to-leading-order results with effective-theory NNLO building blocks, improving their kinematic coverage.
  • The deeper triangle expansion provides additional input for rational-approximant constructions of the full top-mass dependence, a route the paper notes has already been used for the related Higgs-gluon form factor.
  • Because the curves for the deepest expansions lie close together, rescaling the expanded higher-order corrections by the exact leading-order ratio should give numerically stable approximations across the plotted phase space.
  • The projection technique for bypassing three-loop tensor vacuum integrals can be reused in other multiloop calculations with a large internal mass, reducing both memory and CPU requirements.

Reading between the lines

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

  • If the three-loop series converges like the one- and two-loop series shown below threshold, the new terms could make rational-approximant predictions trustworthy well above the two-top-quark threshold, where no exact three-loop result exists; this goes beyond what the paper explicitly demonstrates.
  • The observation that one box-type form factor starts only at order $1/m_t^2$ for most colour structures suggests the box corrections are more suppressed than the triangle corrections, so the effective-theory uncertainty in NNLO approximations may come mainly from the triangle sector.
  • A useful stress test would be comparing the deepest box expansion against the approximate NNLO cross-section predictions after the new terms are included, to see whether the size of the shift decreases with each added order.
  • If the projection method's efficiency holds, the same approach could target the next $1/m_t^{10}$ box terms rather than waiting for exact finite-top-mass integrals.
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

1 major / 5 minor

Summary. The manuscript computes the three-loop (NNLO) virtual form factors for gg -> HH in the large top quark mass limit. Using qgraf, q2e/exp, FORM-based expansions, and FIRE reduction, the authors obtain five expansion terms (up to 1/m_t^8) for the two box-type form factors F_box1 and F_box2 and eight expansion terms (up to 1/m_t^14) for the triangle form factor F_tri. After UV renormalization and Catani IR subtraction, the finite expressions are provided in a Mathematica ancillary file (resFF.m). The paper also describes substantial computational optimizations, including projection of three-loop vacuum integrals onto an ansatz, graph symmetries, ArgToExtraSymbol, and gzip compression of intermediate results. Numerical plots show the convergence of the 1/m_t expansion and comparisons with exact results where available.

Significance. If correct, these are the first three-loop results beyond the leading power for the box-type gg -> HH form factors, and they extend the triangle form factor by several powers. The results are useful input for NNLO approximations of Higgs-pair production, including Padé-based constructions and soft-virtual approximations. Strengths of the paper include: no parameters are fitted to the target result; the computational pipeline is described in unusual detail; machine-readable analytic results are supplied; and the method is validated at one loop (all form factors) and at two loops (triangle form factor) against exact results. The technical optimizations described are likely to benefit future multiloop calculations in the same framework.

major comments (1)
  1. [Section 5, Figs. 3-4 and Section 4.2] The three-loop box-type form factors are entirely new, and the only reported checks for them are cancellation of IR poles after the Catani subtraction and the apparent convergence of the same 1/m_t expansion. These checks do not constrain the finite transcendental parts of the new coefficients. Moreover, while the two-loop triangle form factor is compared with the exact expression (text below Fig. 3), no comparison with exact results is shown for the two-loop box form factors, so the validation chain skips the complexity level closest to the new three-loop box calculation. Please add an independent check, for example a comparison of the two-loop box form factors with the exact NLO results from Refs. [4-6] at representative kinematic points, or a numerical evaluation of a subset of the new three-loop master integrals. If no independent check is feasible, the conclusions should explicitly state that the three-loop box coefficients currently rest on internal-consistency checks alone.
minor comments (5)
  1. [Abstract] The phrase 'We present analytic results for the form factors in the soft-virtual approximation' is misleading: the results contain full dependence on s, t and u through the logarithms and dilogarithms defined in Section 5, not only the soft-virtual limit. Please reword to say 'in the large top mass limit' or 'with full kinematic dependence within the large-mt expansion'.
  2. [Section 5, first paragraph] There is a typo: 'three-loop for factors' should read 'three-loop form factors'.
  3. [Introduction, paragraph 1] The expression '1/m12 t ' appears to be a typesetting issue; it should be 1/m_t^12.
  4. [Figure 3 caption] Since the exact two-loop comparison is made only for the triangle form factor, the caption should state this explicitly so that readers do not infer that all two-loop curves are checked against exact results.
  5. [Appendix A and Ref. [49]] The full results are not printed in the manuscript but reside in an external ancillary file. Please ensure that the file is included as a permanent ancillary file with the published version (e.g., arXiv ancillary file or journal supplementary material) and that Ref. [49] remains accessible; also state the file format and any restrictions on use.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the three-loop 1/m_t expansion is derived from Feynman-diagram integrals with no fitted or pre-supplied target coefficients.

full rationale

The paper computes new 1/m_t expansion coefficients for the three-loop gg->HH form factors by generating Feynman diagrams with qgraf, applying the hard-mass asymptotic expansion with q2e/exp, projecting onto Lorentz structures, reducing vacuum integrals with MATAD and massless integral families with FIRE, and expressing the master integrals through known analytic results. None of the target quantities—the coefficients of Ftri^(2), Fbox1^(2), and Fbox2^(2)—is used as an input: no parameter is fitted to the final expansion, and the Catani IR subtraction is a fixed scheme with known Ig^(1) and Ig^(2) that removes poles but does not determine the finite constants. The checks against exact one-loop results and the exact two-loop triangle form factor are external validations rather than inputs. Self-citations are to computational infrastructure and to previously computed master integrals or lower-order results; they are not invoked as uniqueness theorems and do not assume the new coefficients. The projection ansatz of Eq. (13) is a complete Lorentz/tensor decomposition, not an assumption of the answer. The absence of an exact three-loop comparison is a verification or correctness risk, not a circularity, because it does not make the derivation depend on its own conclusion.

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

The computation introduces no fitted parameters and no new physical entities. It relies on standard QCD inputs, the asymptotic expansion limit, known master integrals, IBP and vacuum-integral tools, and the Catani IR subtraction scheme. The numerical values used in the plots, mt = 173 GeV, mH = 125 GeV, nl = 5, nh = 1, are external inputs rather than fitted parameters.

assumptions (5)
  • domain assumption The large-top-mass asymptotic expansion is valid: m_t is large compared to all external momenta, as stated in Eq. (11).
    Section 3.1 defines the expansion limit and every result is an asymptotic series in 1/m_t^2. The later evaluation at sqrt(s) up to 400 GeV reaches the edge of the assumed hierarchy.
  • domain assumption The top quark is treated in the decoupling scheme, with alpha_s renormalized in five-flavour QCD and the top mass converted to the on-shell scheme.
    Section 4.1 describes the decoupling and renormalization; the final expressions use alpha_s^(5) and the pole mass. This is a standard scheme choice, not a derived result of the paper.
  • standard math The analytic master integrals for one- and two-loop massless integral families from Refs. [21,39-41] are correct.
    Section 3.1 states that analytic expressions for the master integrals are taken from the literature. The paper does not re-derive them.
  • standard math The integral reduction and vacuum integration performed with FIRE and MATAD are correct.
    Section 3.1 relies on FIRE for integral tables and MATAD for vacuum integrals. These are established tools but are imported without formal verification in this paper.
  • standard math The Catani IR subtraction procedure of Refs. [47,48] removes all infrared poles at the amplitude level.
    Section 4.2 applies Eq. (21) and verifies that the poles cancel, but the subtraction scheme itself is taken from the literature.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Three-loop form factors for Higgs boson pair production in the large top mass limit." pith.science (2026). https://pith.science/paper/VVUMI6JN

@misc{pith2026190901361,
  author       = {Pith},
  title        = {Pith review of: Three-loop form factors for Higgs boson pair production in the large top mass limit},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VVUMI6JN}},
  note         = {Machine review of arXiv:1909.01361}
}
read the original abstract

We consider the virtual corrections to Higgs boson pair production at next-to-next-to-leading order, in the large top quark mass limit. We compute five expansion terms for the box-type form factors and eight expansion terms for the triangle form factor, which serve as useful input for the construction of approximations. We present analytic results for the form factors in the soft-virtual approximation. From a technical point of view the calculation is quite challenging since huge intermediate expressions are produced. We describe our methods and optimizations to overcome these difficulties, which might be useful for other calculations.

Figures

Figures reproduced from arXiv: 1909.01361 by the authors.

Figure 1
Figure 1. Sample Feynman diagrams contributing to gg → HH. For simplicity we show diagrams with a triple-Higgs boson coupling only at one-loop order. A sample colour factor is shown below each diagram. However, note that in general a diagram contributes to more than one colour structure. Solid, dashed and curly lines denote quarks, Higgs bosons and gluons respectively. It is furthermore convenient to express the final result … view at source ↗
Figure 2
Figure 2. Sample three-loop diagrams (left) and the corresponding [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Real parts of one- and two-loop form factors as a funct [PITH_FULL_IMAGE:figures/full_fig_p018_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Real parts of three-loop form factors as a function of [PITH_FULL_IMAGE:figures/full_fig_p019_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 11 canonical work pages

  1. [17]

    Davies, R

    J. Davies, R. Gr¨ ober, A. Maier, T. Rauh and M. Steinhauser, P hys. Rev. D 100 (2019) 034017 [arXiv:1906.00982 [hep-ph]]

    arXiv 2019

  2. [1]

    Baglio, A

    J. Baglio, A. Djouadi, R. Gr¨ ober, M. M. M¨ uhlleitner, J. Quevillon a nd M. Spira, JHEP 1304 (2013) 151 [arXiv:1212.5581 [hep-ph]]

    arXiv 2013

  3. [2]

    E. W. N. Glover and J. J. van der Bij, Nucl. Phys. B 309 (1988) 282

    work page 1988

  4. [3]

    Plehn, M

    T. Plehn, M. Spira and P. M. Zerwas, Nucl. Phys. B 479 (1996) 46 Erratum: [Nucl. Phys. B 531 (1998) 655] [hep-ph/9603205]

    arXiv 1996

  5. [4]

    Borowka, N

    S. Borowka, N. Greiner, G. Heinrich, S. P. Jones, M. Kerner, J. Schlenk, U. Schubert and T. Zirke, Phys. Rev. Lett. 117 (2016) no.1, 012001 Erratum: [Phys. Rev. Lett. 117 (2016) no.7, 079901] [arXiv:1604.06447 [hep-ph]]

    arXiv 2016

  6. [5]

    Borowka, N

    S. Borowka, N. Greiner, G. Heinrich, S. P. Jones, M. Kerner, J. Schlenk and T. Zirke, JHEP 1610 (2016) 107 [arXiv:1608.04798 [hep-ph]]

    arXiv 2016

  7. [6]

    Baglio, F

    J. Baglio, F. Campanario, S. Glaus, M. M¨ uhlleitner, M. Spira and J. Streicher, Eur. Phys. J. C 79 (2019) no.6, 459 [arXiv:1811.05692 [hep-ph]]

    arXiv 2019

  8. [7]

    Dawson, S

    S. Dawson, S. Dittmaier and M. Spira, Phys. Rev. D 58 (1998) 115012 [hep-ph/9805244]

    arXiv 1998

Show all 52 references
  1. [8]

    Grigo, K

    J. Grigo, K. Melnikov and M. Steinhauser, Nucl. Phys. B 888 (2014) 17 [arXiv:1408.2422 [hep-ph]]

    2014 arXiv

  2. [9]

    Degrassi, P

    G. Degrassi, P. P. Giardino and R. Gr¨ ober, Eur. Phys. J. C 76 (2016) no.7, 411 [arXiv:1603.00385 [hep-ph]]

    2016 arXiv

  3. [10]

    Bonciani, G

    R. Bonciani, G. Degrassi, P. P. Giardino and R. Gr¨ ober, Phys. R ev. Lett. 121 (2018) no.16, 162003 [arXiv:1806.11564 [hep-ph]]

    2018 arXiv

  4. [11]

    Davies, G

    J. Davies, G. Mishima, M. Steinhauser and D. Wellmann, JHEP 1803 (2018) 048 [arXiv:1801.09696 [hep-ph]]

    2018 arXiv

  5. [12]

    Davies, G

    J. Davies, G. Mishima, M. Steinhauser and D. Wellmann, JHEP 1901 (2019) 176 [arXiv:1811.05489 [hep-ph]]

    2019 arXiv

  6. [13]

    Davies, G

    J. Davies, G. Heinrich, S. P. Jones, M. Kerner, G. Mishima, M. St einhauser and D. Wellmann, arXiv:1907.06408 [hep-ph]

    1907 arXiv

  7. [14]

    https://github.com/mppmu/hhgrid

  8. [15]

    Maltoni, E

    F. Maltoni, E. Vryonidou and M. Zaro, JHEP 1411 (2014) 079 [arXiv:1408.6542 [hep-ph]]

    2014 arXiv

  9. [16]

    Gr¨ ober, A

    R. Gr¨ ober, A. Maier and T. Rauh, JHEP 1803 (2018) 020 [arXiv:1709.07799 [hep- ph]]. 21

    2018 arXiv

  10. [18]

    R. V. Harlander, M. Prausa and J. Usovitsch, arXiv:1907.06957 [hep-ph]

    1907 arXiv

  11. [19]

    de Florian and J

    D. de Florian and J. Mazzitelli, Phys. Lett. B 724 (2013) 306 [arXiv:1305.5206 [hep- ph]]

    2013 arXiv

  12. [20]

    de Florian and J

    D. de Florian and J. Mazzitelli, Phys. Rev. Lett. 111 (2013) 201801 [arXiv:1309.6594 [hep-ph]]

    2013 arXiv

  13. [21]

    Grigo, J

    J. Grigo, J. Hoff and M. Steinhauser, Nucl. Phys. B 900 (2015) 412 [arXiv:1508.00909 [hep-ph]]

    2015 arXiv

  14. [22]

    Davies, F

    J. Davies, F. Herren, G. Mishima and M. Steinhauser, JHEP 1905 (2019) 157 [arXiv:1904.11998 [hep-ph]]

    2019 arXiv

  15. [23]

    Grazzini, G

    M. Grazzini, G. Heinrich, S. Jones, S. Kallweit, M. Kerner, J. M. L indert and J. Mazz- itelli, JHEP 1805 (2018) 059 [arXiv:1803.02463 [hep-ph]]

    2018 arXiv

  16. [24]

    D. Y. Shao, C. S. Li, H. T. Li and J. Wang, JHEP 1307 (2013) 169 [arXiv:1301.1245 [hep-ph]]

    2013 arXiv

  17. [25]

    de Florian and J

    D. de Florian and J. Mazzitelli, JHEP 1509 (2015) 053 [arXiv:1505.07122 [hep-ph]]

    2015 arXiv

  18. [26]

    de Florian, M

    D. de Florian, M. Grazzini, C. Hanga, S. Kallweit, J. M. Lindert, P. Maierh¨ ofer, J. Mazzitelli and D. Rathlev, JHEP 1609 (2016) 151 [arXiv:1606.09519 [hep-ph]]

    2016 arXiv

  19. [27]

    De Florian and J

    D. De Florian and J. Mazzitelli, JHEP 1808 (2018) 156 [arXiv:1807.03704 [hep-ph]]

    2018 arXiv

  20. [28]

    Banerjee, S

    P. Banerjee, S. Borowka, P. K. Dhani, T. Gehrmann and V. Rav indran, JHEP 1811 (2018) 130 [arXiv:1809.05388 [hep-ph]]

    2018 arXiv

  21. [29]

    Spira, JHEP 1610 (2016) 026 [arXiv:1607.05548 [hep-ph]]

    M. Spira, JHEP 1610 (2016) 026 [arXiv:1607.05548 [hep-ph]]

    2016 arXiv

  22. [30]

    Gerlach, F

    M. Gerlach, F. Herren and M. Steinhauser, JHEP 1811 (2018) 141 [arXiv:1809.06787 [hep-ph]]

    2018 arXiv

  23. [31]

    R. V. Harlander and K. J. Ozeren, Phys. Lett. B 679 (2009) 467 [arXiv:0907.2997 [hep-ph]]

    2009 arXiv

  24. [32]

    A. Pak, M. Rogal and M. Steinhauser, Phys. Lett. B 679 (2009) 473 [arXiv:0907.2998 [hep-ph]]

    2009 arXiv

  25. [33]

    Nogueira, J

    P. Nogueira, J. Comput. Phys. 105 (1993) 279

    1993

  26. [34]

    Harlander, T

    R. Harlander, T. Seidensticker and M. Steinhauser, Phys. Let t. B 426 (1998) 125 [hep-ph/9712228]. 22

    1998 arXiv

  27. [35]

    Seidensticker, hep-ph/9905298

    T. Seidensticker, hep-ph/9905298

    arXiv

  28. [36]

    http://sfb-tr9.ttp.kit.edu/software/html/q2eexp.html

  29. [37]

    Ruijl, T

    B. Ruijl, T. Ueda and J. Vermaseren, arXiv:1707.06453 [hep-ph]

    arXiv

  30. [38]

    V. A. Smirnov, Springer Tracts Mod. Phys. 250 (2012) 1

    2012

  31. [39]

    T. G. Birthwright, E. W. N. Glover and P. Marquard, JHEP 0409 (2004) 042 [hep-ph/0407343]

    2004 arXiv

  32. [40]

    R. K. Ellis and G. Zanderighi, JHEP 0802 (2008) 002 [arXiv:0712.1851 [hep-ph]]

    2008 arXiv

  33. [41]

    Chavez and C

    F. Chavez and C. Duhr, JHEP 1211 (2012) 114 [arXiv:1209.2722 [hep-ph]]

    2012 arXiv

  34. [42]

    K. G. Chetyrkin, In *Oberammergau 1993, New computing tech niques in physics research III* 559-563 [hep-ph/0212040]

    1993 arXiv

  35. [43]

    Steinhauser, Comput

    M. Steinhauser, Comput. Phys. Commun. 134 (2001) 335 [hep-ph/0009029]

    2001 arXiv

  36. [44]

    A. V. Smirnov, Comput. Phys. Commun. 189 (2015) 182 [arXiv:1408.2372 [hep-ph]]

    2015 arXiv

  37. [45]

    Fleischer and O

    J. Fleischer and O. V. Tarasov, Z. Phys. C 64 (1994) 413 [hep-ph/9403230]

    1994 arXiv

  38. [46]

    K. G. Chetyrkin, B. A. Kniehl and M. Steinhauser, Nucl. Phys. B 510 (1998) 61 [hep-ph/9708255]

    1998 arXiv

  39. [47]

    Catani, Phys

    S. Catani, Phys. Lett. B 427 (1998) 161 [hep-ph/9802439]

    1998 arXiv

  40. [48]

    de Florian and J

    D. de Florian and J. Mazzitelli, JHEP 1212 (2012) 088 [arXiv:1209.0673 [hep-ph]]

    2012 arXiv

  41. [49]

    https://www.ttp.kit.edu/preprints/2019/ttp19-028/

    2019

  42. [50]

    Harlander and P

    R. Harlander and P. Kant, JHEP 0512 (2005) 015 [hep-ph/0509189]

    2005 arXiv

  43. [51]

    Anastasiou, S

    C. Anastasiou, S. Beerli, S. Bucherer, A. Daleo and Z. Kunszt, JHEP 0701 (2007) 082 [hep-ph/0611236]

    2007 arXiv

  44. [52]

    Aglietti, R

    U. Aglietti, R. Bonciani, G. Degrassi and A. Vicini, JHEP 0701 (2007) 021 [hep-ph/0611266]. 23

    2007 arXiv

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.