Renormalon singularities beyond the leading ultraviolet pole: the higher-dimension operator hierarchy, infrared renormalons and nonlocal operator content in massless asymptotically free theories, and the fate of renormalons in the $4-\epsilon$ expansion

In renormalizable quantum field theories the perturbation series in the coupling diverges factorially, and its Borel transform (which divides the $k$-th coefficient by $k!$ to restore convergence) has branch-point singularities generated by the running of the coupling, called renormalons. Parisi's 1978 analysis of these singularities established that in a massive renormalizable theory the leading ultraviolet renormalon ambiguity of a Green function factorizes over matrix elements of dimension-6 local operators, with coefficients fixed by a renormalization-group equation, and showed how the singularity maps onto the $4-\epsilon$ expansion at the Wilson-Fisher fixed point. Within the renormalon program so initiated - extended to massless asymptotically free theories by Parisi's 1979 companion paper 'On Infrared Divergences' and systematized in later reviews - three issues remain unresolved as of 2026: (1) the extension of the factorization and operator-compensation mechanism to dimension-8 and higher operators, i.e. the full ultraviolet renormalon hierarchy; the compensation idea (ultraviolet renormalons can be compensated by adding higher-dimension operators to the Lagrangian) and the renormalization-group treatment originate with Parisi 1978, while the explicit acknowledgment that the extension to the next singularity is possible in principle, requiring single insertions of dimension-8 operators and double insertions of dimension-6 operators, is Beneke's 1999 review, which judged it 'probably already too complicated to be useful'; the hierarchy has never been worked out; (2) the classification and construction, in massless asymptotically free theories, of the nonlocal operator insertions analogous to local dimension-6 operators, needed to specify the infrared renormalon singularities - the infrared side belongs to Parisi's 1979 companion paper rather than to the 1978 paper itself, nonlocal Wilson-line operator insertions were first analyzed for jet observables by Korchemsky-Sterman (1995), and a general classification explicitly stated not to exist in the standard review is still missing, with 2026 work proposing only case-by-case workarounds; (3) the quantitative fate of renormalon non-analyticities in the $4-\epsilon$ expansion - Parisi's analysis supplies the mapping only to leading orders, and whether the non-analyticity survives in $\epsilon$-expanded quantities is contested between analyses that find them suppressed or absent in MS-scheme (minimal subtraction) renormalization-group functions and critical exponents, and the framework expectation that they are present.

Background

Perturbation expansions in quantum field theory (QFT) are divergent asymptotic series: the coefficient of $g^K$ typically grows like $K!$, where $g$ is the coupling. The standard diagnostic device is the Borel transform $B(t)=\sum_K c_K t^K/K!$, which divides the $K$-th coefficient by $K!$ to restore convergence in the Borel variable $t$; if $B(t)$ is analytic along the path of integration and the Borel integral converges, the series is Borel summable and defines the function unambiguously. 't Hooft (1977 Erice lectures) and Parisi (1978) identified a new source of Borel-plane singularities, distinct from instantons (classical finite-action solutions of the field equations, whose neighborhoods dominate the path integral at large order and fix the leading factorial growth of the series in a number of scalar theories): in a renormalizable theory, chains of bubble subdiagrams (the simplest chain-like classes of loop insertions) feel the running coupling $\beta(g)=\mu\,dg/d\mu=\beta_2 g^2-\beta_3 g^3-\cdots$ ($\mu$ is the renormalization scale), and the integral over loop momentum produces branch points at $t=2n/\beta_2$ (with $\beta_2,\beta_3$ the first two beta-function coefficients and $n=1,2,\ldots$). Singularities from the large-momentum region are called ultraviolet (UV) renormalons; those from the small-momentum region in massless asymptotically free theories - theories whose coupling vanishes at short distances, of which quantum chromodynamics (QCD), the theory of the strong interaction, is the prototype - are located at negative $t$ (equivalently at positive $u$ in the standard QCD convention $u=-t$) and are called infrared (IR) renormalons. They are called renormalons because they are tied to renormalization: UV renormalons reflect the ambiguity of subtracting power divergences, and IR renormalons signal power corrections of order $(\Lambda/Q)^p$, where $\Lambda$ is the QCD scale and $Q$ a hard (large-momentum) scale, which are parametrized by the operator product expansion (OPE, the expansion of a short-distance amplitude in matrix elements of local operators of increasing dimension).

Parisi's 1978 paper analyzed the leading UV renormalon of a massive renormalizable theory. Writing $G(g)$ for a Green function (a correlation function of the fields) and $G_i$ for matrix elements of dimension-6 local operators, the leading renormalon ambiguity (the imaginary part of the Borel integral around the singularity) factorizes as $\operatorname{Im} G(g)=\sum_i f_i(g)\,G_i(g)+O\left(e^{-4/(\beta_2 g)}\right)$: the ambiguity is absorbed by adding dimension-6 operators to the Lagrangian, just as logarithmic UV divergences are compensated by dimension-4 counterterms. The coefficient functions obey the renormalization-group (RG) equation $\left[-\beta(g)\frac{d}{dg}+2+\gamma_6(g)\right]f(g)=0$, where $\gamma_6$ is the anomalous dimension of the dimension-6 operator (the quantum correction to its classical scaling dimension), with solution $f(g)=g^{\alpha_6}e^{-2/(\beta_2 g)}$ and $\alpha_6=\gamma_6/\beta_2+2\beta_3/\beta_2^2$. The corresponding Borel-plane branch point sits at $t=2/\beta_2$ and the perturbative coefficients grow like $(\beta_2/2)^K\Gamma(K-\alpha_6)$. Parisi verified the pattern against the exactly solvable large-$N$ limit (the limit of infinitely many field components, where $\alpha_6\to 3$), and mapped the singularity onto the $4-\epsilon$ expansion (expanding the space-time dimension as $d=4-\epsilon$, in which the theory has the Wilson-Fisher fixed point $g_c$, the nontrivial fixed point of the renormalization-group flow at which the epsilon expansion is evaluated): the four-dimensional non-analyticity appears as $(g_c-g)^{2/\epsilon-1+\alpha_6+O(\epsilon)}$ with $g_c=\epsilon/\beta_2+\epsilon^2\beta_3/\beta_2^3+O(\epsilon^3)$.

Later literature systematized but did not complete this program. Beneke-Braun-Kivel (1997) worked out the RG treatment for QCD; Beneke's 1999 review fixed the standard framework, noting that the extension to the next UV singularity is in principle possible but 'probably already too complicated to be useful', and that a general classification of IR renormalon singularities comparable to the UV classification 'is not known at this time'. The IR side requires nonlocal operators - gauge-invariant generalizations of local operators containing Wilson lines (path-ordered exponentials of the gauge field along a contour) - first analyzed for jet observables (event-shape measures of the collimated sprays of particles produced in collider processes) by Korchemsky-Sterman (1995). The $\epsilon$-expansion question has become more, not less, contested: large-$N$ solutions (Moshe-Zinn-Justin 2003; Brezin 2023) find no renormalon non-analyticity in critical exponents (the power-law exponents governing correlation functions at a critical point), while the framework predicts one, and the exact six- and seven-loop MS-scheme (minimal subtraction) renormalization-group functions of $O(N)$ $\phi^4$ theory favor instanton-dominated large-order behavior (Dunne-Meynig 2022), leaving the suppression of renormalons unexplained. Recent work on removing IR renormalons by gradient flow (a smoothing evolution of the fields in an auxiliary flow time) (Beneke-Takaura 2026) confirms the issue is still open.

Problem Statement

Within the renormalon framework initiated by Parisi's 1978 analysis of Borel-transform singularities in renormalizable theories (massive theory, leading ultraviolet renormalon at Borel variable $t=2/\beta_2$, factorization over dimension-6 operators), extended to massless asymptotically free theories by his 1979 companion paper and systematized in the cited reviews, three problems remain open. They are independent.

(a) The higher-dimension ultraviolet renormalon hierarchy. Extend the factorization and operator-compensation mechanism - originating with Parisi and acknowledged in the standard review as extendable in principle to the next singularity but 'probably already too complicated to be useful' - beyond the leading pole: derive the analogue of $\operatorname{Im} G(g)=\sum_i f_i(g)G_i(g)+O\left(e^{-4/(\beta_2 g)}\right)$ once dimension-8 operators are included, which requires treating single insertions of dimension-8 operators together with double insertions of dimension-6 operators and their operator mixing; obtain the renormalization-group equations and anomalous dimensions governing the new coefficient functions, the positions and exponents of the subsequent Borel-plane singularities ($t=4/\beta_2, 6/\beta_2, \ldots$) and the corresponding large-order behavior; and determine how far the hierarchy can be carried and in what sense it is universal (scheme-independent, as the leading pole is).

(b) Infrared renormalons and nonlocal operators in massless asymptotically free theories. In a massless asymptotically free theory the infrared renormalon singularities (at negative $t$, or positive $u$ in the standard QCD convention $u=-t$) cannot be parametrized by local operators, since no general structural theorem for infrared divergences exists. Construct and classify the nonlocal operator insertions - gauge-invariant operators involving Wilson lines - that play the role analogous to local dimension-6 operators in the massive case; specify the detailed form, location and exponent of the infrared renormalon singularities for a general Green function, a classification comparable to the ultraviolet one, which the standard review explicitly states 'is not known at this time' and which still does not exist (partial results cover only specific observables: twist-4 nonlocal operators (operators of dimension minus spin equal to 4) for jet observables, the pole mass (the quark mass defined as the pole of the propagator, whose perturbative value is ambiguous by a renormalon), and, most recently, a gradient-flowed OPE for the Adler function (the QCD vacuum-polarization function underlying tau-lepton decay rates) that bypasses the classification rather than supplying it).

(c) The fate of renormalons in the $4-\epsilon$ expansion. Parisi's analysis maps the four-dimensional renormalon singularity to the non-analyticity $(g_c-g)^{2/\epsilon-1+\alpha_6+O(\epsilon)}$ at the Wilson-Fisher fixed point $g_c=\epsilon/\beta_2+\epsilon^2\beta_3/\beta_2^3+O(\epsilon^3)$, to leading orders only. Determine quantitatively whether this non-analyticity survives in $\epsilon$-expanded quantities (for example critical exponents) or is suppressed - for instance through the higher-order $O(\epsilon)$ corrections to the exponent and the subleading terms in $g_c$ - and reconcile the prediction with the modern high-loop MS data for $O(N)$ $\phi^4$ theory (which favor instanton-dominated large-order behavior), with large-$N$ exact solutions that find no renormalon non-analyticity, and with proposed renormalon-freeness of MS-scheme renormalization-group functions.

References

1. DOI:10.1016/0370-2693(78)90101-6 - G. Parisi, Singularities of the Borel transform in renormalizable theories, Phys. Lett. B 76 (1978) 65-66 - https://doi.org/10.1016/0370-2693(78)90101-6
2. DOI:10.1007/978-1-4684-0991-8_17 - G. 't Hooft, Can We Make Sense Out of 'Quantum Chromodynamics'? (1977 Erice lectures), in A. Zichichi (ed.), The Whys of Subnuclear Physics, Subnuclear Series vol. 15, Plenum Press, New York, 1979, pp. 943-982 - https://doi.org/10.1007/978-1-4684-0991-8_17
3. DOI:10.1016/0550-3213(79)90298-0 - G. Parisi, On Infrared Divergences, Nucl. Phys. B 150 (1979) 163-172 - https://doi.org/10.1016/0550-3213(79)90298-0
4. DOI:10.1016/0370-2693(82)90691-8 - G. 't Hooft, Borel Summability of a Four-dimensional Field Theory, Phys. Lett. B 119 (1982) 369-371 - https://doi.org/10.1016/0370-2693(82)90691-8
5. DOI:10.1016/0550-3213(85)90485-7 - A. H. Mueller, On the Structure of Infrared Renormalons in Physical Processes at High Energies, Nucl. Phys. B 250 (1985) 327-350 - https://doi.org/10.1016/0550-3213(85)90485-7
6. DOI:10.1016/0550-3213(89)90168-5 - I. I. Balitsky, V. M. Braun, Evolution Equations for QCD String Operators, Nucl. Phys. B 311 (1989) 541-584 - https://doi.org/10.1016/0550-3213(89)90168-5
7. DOI:10.1016/0550-3213(92)90054-F - V. I. Zakharov, QCD perturbative expansions in large orders, Nucl. Phys. B 385 (1992) 452-480 - https://doi.org/10.1016/0550-3213(92)90054-F
8. DOI:10.1103/PhysRevLett.73.1207 - A. I. Vainshtein, V. I. Zakharov, Ultraviolet renormalon calculus, Phys. Rev. Lett. 73 (1994) 1207 [Erratum: Phys. Rev. Lett. 75 (1995) 3588] - https://doi.org/10.1103/PhysRevLett.73.1207 (arXiv:hep-ph/9404248)
9. DOI:10.1016/0550-3213(94)00006-Z - G. P. Korchemsky, G. Sterman, Nonperturbative corrections in resummed cross-sections, Nucl. Phys. B 437 (1995) 415-432 - https://doi.org/10.1016/0550-3213(94)00006-Z (arXiv:hep-ph/9411211)
10. DOI:10.1016/S0370-2693(97)00562-5 - M. Beneke, V. M. Braun, N. Kivel, Large-order behaviour due to ultraviolet renormalons in QCD, Phys. Lett. B 404 (1997) 315-320 - https://doi.org/10.1016/S0370-2693(97)00562-5 (arXiv:hep-ph/9703389)
11. DOI:10.1016/S0370-1573(98)00130-6 - M. Beneke, Renormalons, Phys. Rept. 317 (1999) 1-142 - https://doi.org/10.1016/S0370-1573(98)00130-6 (arXiv:hep-ph/9807443)
12. DOI:10.1016/S0370-1573(03)00263-1 - M. Moshe, J. Zinn-Justin, Quantum field theory in the large N limit: a review, Phys. Rept. 385 (2003) 69-228 - https://doi.org/10.1016/S0370-1573(03)00263-1 (arXiv:hep-th/0306133)
13. DOI:10.1142/S0217751X15430010 - M. Shifman, New and old about renormalons: In memoriam Kolya Uraltsev, Int. J. Mod. Phys. A 30 (2015) 1543001 - https://doi.org/10.1142/S0217751X15430010 (arXiv:1310.1966)
14. DOI:10.1016/j.aop.2018.04.027 - A. Maiezza, J. C. Vasquez, Renormalons in a general Quantum Field Theory, Annals Phys. 394 (2018) 84-99 - https://doi.org/10.1016/j.aop.2018.04.027 (arXiv:1802.06022)
15. DOI:10.1088/1751-8121/aaf768 - A. J. McKane, Perturbation expansions at large order: results for scalar field theories revisited, J. Phys. A 52 (2019) 055401 - https://doi.org/10.1088/1751-8121/aaf768 (arXiv:1807.00656)
16. DOI:10.1103/PhysRevD.105.025019 - G. V. Dunne, M. Meynig, Instantons or renormalons? Remarks on $\phi^4_{d=4}$ theory in the MS scheme, Phys. Rev. D 105 (2022) 025019 - https://doi.org/10.1103/PhysRevD.105.025019 (arXiv:2111.15554)
17. arXiv:2301.01174 - E. Brezin, Should we worry about renormalons in the epsilon-expansion? (2023) - https://arxiv.org/abs/2301.01174
18. DOI:10.1007/jhep03(2026)033 - M. Beneke, H. Takaura, Gradient-flowed operator product expansion without IR renormalons, J. High Energy Phys. 2026 (2026) 033 - https://doi.org/10.1007/jhep03(2026)033 (arXiv:2510.12193)