Nonperturbative completion of Parisi's infrared-counterterm program: uniqueness of multilocal infrared counterterms, the nonperturbative remainder, and renormalon-shielded instanton effects

In 'On Infrared Divergences' (Nucl. Phys. B 150 (1979) 163), G. Parisi proposed that the infrared divergences of massless and superrenormalizable quantum field theories be absorbed into multilocal counterterms — products of local operators at separated points, completed by Wilson lines in gauge theories — and conjectured an infrared analogue of the Bogoliubov–Parasiuk–Hepp renormalization theorem. He connected the same infrared structure to the large-order behaviour of the perturbation series: infrared renormalon singularities of the Borel transform, a nonperturbative remainder $A_{\mathrm{NP}}(\alpha)$ tied to power corrections, and instanton singularities shielded by nearer renormalon singularities in asymptotically free gauge theories. Later literature resolved two layers of this program: the R*-operation of Chetyrkin, Tkachov and Smirnov gives an algorithmic, all-orders subtraction of Euclidean infrared divergences in dimensional regularization, and the renormalon–operator-product-expansion correspondence established by David, Mueller and others fixes the operator interpretation of the leading nonperturbative ambiguities. The open core is the nonperturbative layer: whether there exists — and how to characterize or classify — a choice of the infrared counterterms, equivalently a trans-series/nonperturbative completion of the infrared-divergent perturbative expansion, that restores full physical consistency (cluster decomposition, unitarity, locality of the completed Green functions and gauge invariance) beyond perturbation theory, together with the definition or extraction of instanton-induced contributions whose Borel singularities are shielded by nearer infrared renormalons (as in pure SU(N) Yang–Mills in the source's comparison).

Background

Perturbative quantum field theory suffers two complementary kinds of divergence. Ultraviolet (short-distance) divergences are controlled by the Bogoliubov–Parasiuk–Hepp (BPH) renormalization theorem: the divergent part of any Feynman diagram can be absorbed into local counterterms — polynomials in external momenta and masses added to the Lagrangian — and the recursive subtraction of overlapping divergent subgraphs (Bogoliubov's R-operation, solved by Zimmermann's forest formula, the closed-form prescription that enumerates all nestings of divergent subgraphs) terminates for arbitrary diagrams (Bogoliubov and Parasiuk 1957; Hepp 1966). Infrared (long-distance) divergences are not covered by this theorem. In massless and superrenormalizable theories — theories with only finitely many ultraviolet divergences, such as scalar quartic theories below four dimensions — individual Feynman amplitudes become singular from the small-momentum (soft) region of loop integrations: in coordinate space the large-separation behaviour of propagators decays too slowly for the Fourier transform to exist, so that off-shell Euclidean Green functions (correlation functions at distinct points, at generic non-exceptional external momenta — momenta for which no partial sum vanishes) are ill-defined order by order in the coupling. Infrared-safe cross sections evade this through the cancellation between virtual soft radiation and real soft emission, but the Green-function level has no such escape.

In 'On Infrared Divergences' (1979), building on his 1978 analysis of Borel-transform singularities, Parisi proposed to treat infrared divergences the way the BPH theorem treats ultraviolet ones, with two essential modifications. First, the required counterterms cannot be local: they must be multilocal, i.e., products of local operators evaluated at separated points, such as $O(x_1)\,O(x_2)$ integrated with kernels depending on $x_1-x_2$; a typical infrared counterterm is bilocal. Second, in gauge theories a bilocal operator is not gauge invariant by itself; it must be completed with a Wilson line — the path-ordered exponential $P\exp\left(i\int_x^y dz_\mu A_\mu(z)\right)$ of the gauge field along a path connecting the two points, the symbol $P$ ordering the gauge-field factors along the path by increasing path parameter — and Parisi posed the question of which gauge-invariant completion (which path, if any) the dynamics selects. Parisi conjectured the infrared analogue of the BPH theorem: a finite, algorithmic, consistent prescription built from such counterterms that renders the Green functions finite, together with a classification of admissible choices under physical requirements such as cluster decomposition (the property that Euclidean Green functions factorize when their arguments separate into distant groups, which multilocal counterterms naively threaten).

The same paper connected infrared divergences to the large-order behaviour of perturbation theory. The coefficients of the perturbation series of a renormalizable theory grow factorially; the Borel transform — the transform $B(t)=\sum_n a_n t^n/n!$ whose singularities on the $t$-plane encode this growth — contains, in asymptotically free theories (theories whose effective coupling decreases logarithmically at short distances and therefore grows at long distances), both renormalon poles (arising from the running coupling at small loop momenta, with infrared renormalons on the positive Borel axis at positions fixed by the first coefficient of the beta function, the function that governs the scale dependence of the coupling) and instanton–anti-instanton singularities farther out, associated with semiclassical field configurations: an instanton is a finite-action solution of the Euclidean field equations, interpretable as a tunneling event, and instanton–anti-instanton pairs contribute through their interactions. Parisi conjectured that the ambiguities of the infrared renormalons are compensated by the power corrections of the operator product expansion (OPE) — the expansion of a correlation function at short distances in local operators of increasing dimension, with vacuum matrix elements (condensates) — and noted that when the first infrared renormalon lies closer to the origin than the instanton singularity (as for pure SU(N) Yang–Mills), the instanton contribution is shielded and its practical computability is obstructed. Writing the exact answer as a perturbative part plus a nonperturbative remainder $A_{\mathrm{NP}}(\alpha)\sim\exp(-\mathrm{const}/\alpha)$, the paper left the operator content, scale dependence and logarithmic dressing of that remainder, beyond the leading exponential estimate, undetermined.

Two layers of this program were settled later and are treated as background here. The R*-operation of Chetyrkin–Tkachov (1982), Chetyrkin–Smirnov (1984) and Smirnov–Chetyrkin (1985), placed in a general asymptotic-expansion framework by Smirnov (1990), realizes Parisi's conjectured subtraction algorithm within dimensional regularization (continuing the spacetime dimension away from four so that divergences appear as poles in a regulator $\varepsilon$) and the minimal-subtraction scheme (subtracting exactly those pole parts): it subtracts ultraviolet and Euclidean infrared divergences of massless diagrams simultaneously, with a termination and consistency proof, and is now standard multi-loop technology (Herzog 2018). And the renormalon–OPE correspondence was confirmed and sharpened by David (1982, 1984) and Mueller (1985) and systematized in Beneke's 1999 review: the ambiguity of an infrared renormalon equals the ultraviolet renormalization ambiguity of the composite operator that appears in the OPE at the corresponding dimension, so the two cancel in physical quantities; this was verified quantitatively in two-dimensional models and in expansions in powers of $1/N_f$ (the inverse number of quark flavours), and quantified in resurgence analyses of the Adler function (Laenen–Marinissen–Vonk 2023) — resurgence being the framework that reconstructs a function from its divergent power series through the singularities of its Borel transform, and the Adler function being the QCD vacuum-polarization derivative observable that governs the running of the electromagnetic coupling. What no later work settles, for four-dimensional massless and asymptotically free theories themselves, is the nonperturbative layer of Parisi's program, stated below.

Problem Statement

Consider the class of theories addressed by Parisi's 1979 program: massless or superrenormalizable Euclidean field theories whose off-shell Green functions at non-exceptional momenta are order-by-order infrared divergent, and asymptotically free gauge theories whose perturbation series carry infrared renormalon and instanton singularities in the Borel plane. Within this program, the subtraction algorithm (the infrared R*-operation) and the operator interpretation of the leading renormalon ambiguities (via the operator product expansion) have been settled; the open core, common objective of both parts below, is the nonperturbative completion of the infrared-divergent perturbative expansion — by a trans-series where applicable, i.e., a generalized expansion combining the power series in the coupling with exponentially small sectors of order $\exp(-\mathrm{const}/\alpha)$.

(1) Existence, uniqueness and classification of the infrared-counterterm completion. Does there exist a choice — possibly unique — of infrared counterterms (multilocal operators of the type introduced by Parisi, with gauge invariance of bilocal operators implemented through Wilson-line completions), or equivalently a nonperturbative trans-series completion of the perturbative series including the remainder $A_{\mathrm{NP}}(\alpha)$ with its full scale and logarithm dependence, such that the completed theory satisfies beyond any fixed perturbative order the physical requirements of the original theory: cluster decomposition of Euclidean Green functions, locality for the stated theories (understood as the requirement that the multilocal counterterms preserve the local structure of physical quantities), gauge invariance for the gauge-theory completions, and unitarity of the corresponding Minkowskian continuation? Characterize and classify the admissible choices; determine in particular whether the physical requirements select a unique scheme, and establish the multilocality structure of the required counterterms by proof rather than by diagrammatic heuristic, in a framework that does not presuppose it.

(2) Renormalon-shielded instanton contributions. In asymptotically free gauge theories where the first infrared renormalon singularity of the Borel transform lies closer to the origin than the instanton (instanton–anti-instanton) singularities — as in pure SU(N) Yang–Mills in the source's comparison — provide a definition, or an operational extraction procedure, for the instanton-induced nonperturbative contributions when the formal semiclassical expressions are themselves infrared divergent; establish whether these contributions are separately meaningful or only defined jointly with the renormalon sectors of a trans-series.

Part (2) presupposes a resolution of part (1): a construction that realizes the nonperturbative completion of (1) fixes the joint definition of all nonperturbative sectors and thereby determines the sectors that (2) asks to define, whereas a rigorously argued no-go result for (1) rules out any admissible completion and so leaves no nonperturbative completion whose sectors (2) would define.

References

1. G. Parisi, 'On Infrared Divergences', Nuclear Physics B 150 (1979) 163-172, DOI 10.1016/0550-3213(79)90298-0, https://doi.org/10.1016/0550-3213(79)90298-0
2. N. N. Bogoliubow, O. S. Parasiuk, 'Über die Multiplikation der Kausalfunktionen in der Quantentheorie der Felder', Acta Mathematica 97 (1957) 227-266, DOI 10.1007/BF02392399, https://doi.org/10.1007/BF02392399 (the original Bogoliubov–Parasiuk renormalization paper)
3. S. Weinberg, 'High-Energy Behavior in Quantum Field Theory', Physical Review 118 (1960) 838-849, DOI 10.1103/PhysRev.118.838, https://doi.org/10.1103/PhysRev.118.838
4. K. H. Hepp, 'Proof of the Bogoliubov-Parasiuk theorem on renormalization', Communications in Mathematical Physics 2 (1966) 301-326, DOI 10.1007/bf01773358, https://doi.org/10.1007/bf01773358
5. M. C. Bergère, Y.-M. P. Lam, 'Asymptotic expansion of Feynman amplitudes', Communications in Mathematical Physics 39 (1974) 1-32, DOI 10.1007/BF01609168, https://doi.org/10.1007/BF01609168
6. B. Lautrup, 'On high order estimates in QED', Physics Letters B 69 (1977) 109-111, DOI 10.1016/0370-2693(77)90145-9, https://doi.org/10.1016/0370-2693(77)90145-9
7. G. Parisi, 'Singularities of the Borel transform in renormalizable theories', Physics Letters B 76 (1978) 65-66, DOI 10.1016/0370-2693(78)90101-6, https://doi.org/10.1016/0370-2693(78)90101-6
8. P. Olesen, 'On vacuum instability in quantum field theory', Physics Letters B 73 (1978) 327-329, DOI 10.1016/0370-2693(78)90526-9, https://doi.org/10.1016/0370-2693(78)90526-9
9. K. G. Chetyrkin, F. V. Tkachov, 'Infrared R-operation and ultraviolet counterterms in the MS-scheme', Physics Letters B 114 (1982) 340-344, DOI 10.1016/0370-2693(82)90358-6, https://doi.org/10.1016/0370-2693(82)90358-6
10. K. G. Chetyrkin, V. A. Smirnov, 'R* operation corrected', Physics Letters B 144 (1984) 419-424, DOI 10.1016/0370-2693(84)91291-7, https://doi.org/10.1016/0370-2693(84)91291-7
11. V. A. Smirnov, K. G. Chetyrkin, 'R* operation in the minimal subtraction scheme', Theoretical and Mathematical Physics 63 (1985) 462-469, DOI 10.1007/BF01017902, https://doi.org/10.1007/BF01017902
12. V. A. Smirnov, 'Asymptotic expansions in limits of large momenta and masses', Communications in Mathematical Physics 134 (1990) 109-137, DOI 10.1007/BF02102092, https://doi.org/10.1007/BF02102092
13. C. de Calan, A. P. C. Malbouisson, 'Infrared and ultraviolet dimensional meromorphy of Feynman amplitudes', Communications in Mathematical Physics 90 (1983) 413-416, DOI 10.1007/BF01206890, https://doi.org/10.1007/BF01206890
14. Y. Frishman, A. R. White, 'Contributions of renormalon singularities', Nuclear Physics B 158 (1979) 221-233, DOI 10.1016/0550-3213(79)90197-4, https://doi.org/10.1016/0550-3213(79)90197-4
15. G. 't Hooft, 'Borel summability of a four-dimensional field theory', Physics Letters B 119 (1982) 369-371, DOI 10.1016/0370-2693(82)90691-8, https://doi.org/10.1016/0370-2693(82)90691-8
16. F. David, 'Non-perturbative effects and infrared renormalons within the 1/N expansion of the O(N) non-linear sigma model', Nuclear Physics B 209 (1982) 433-460, DOI 10.1016/0550-3213(82)90266-8, https://doi.org/10.1016/0550-3213(82)90266-8
17. F. David, 'On the ambiguity of composite operators, IR renormalons and the status of the operator product expansion', Nuclear Physics B 234 (1984) 237-251, DOI 10.1016/0550-3213(84)90235-9, https://doi.org/10.1016/0550-3213(84)90235-9
18. A. H. Mueller, 'On the structure of infrared renormalons in physical processes at high energies', Nuclear Physics B 250 (1985) 327-350, DOI 10.1016/0550-3213(85)90485-7, https://doi.org/10.1016/0550-3213(85)90485-7
19. M. Beneke, 'Renormalons', Physics Reports 317 (1999) 1-142, DOI 10.1016/S0370-1573(98)00130-6, arXiv:hep-ph/9807443, https://arxiv.org/abs/hep-ph/9807443
20. N. Agarwal, L. Magnea, C. Signorile-Signorile, A. Tripathi, 'The infrared structure of perturbative gauge theories', Physics Reports 994 (2023) 1-120, DOI 10.1016/j.physrep.2022.10.001, arXiv:2112.07099, https://arxiv.org/abs/2112.07099
21. F. Herzog, 'Zimmermann's forest formula, infrared divergences and the QCD beta function', Nuclear Physics B 926 (2018) 370-380, DOI 10.1016/j.nuclphysb.2017.11.011, arXiv:1711.06121, https://arxiv.org/abs/1711.06121
22. F. Brown, 'Feynman Amplitudes and Cosmic Galois group', arXiv:1512.06409, https://arxiv.org/abs/1512.06409
23. M. Mariño, T. Reis, 'Renormalons in integrable field theories', Journal of High Energy Physics 04 (2020) 160, DOI 10.1007/JHEP04(2020)160, arXiv:1909.12134, https://arxiv.org/abs/1909.12134
24. M. Mariño, R. Miravitllas, T. Reis, 'Instantons, renormalons and the theta angle in integrable sigma models', SciPost Physics 15 (2023) 184, DOI 10.21468/scipostphys.15.5.184, arXiv:2205.04495, https://arxiv.org/abs/2205.04495
25. M. Mariño, 'Anatomy of the simplest renormalon', arXiv:2504.12044, https://arxiv.org/abs/2504.12044
26. M. Ünsal, 'TQFT at work for IR-renormalons, resurgence and Lefschetz decomposition', arXiv:2106.14971, https://arxiv.org/abs/2106.14971 (TQFT: topological quantum field theory)
27. M. Beneke, 'Pole mass renormalon and its ramifications', European Physical Journal Special Topics 230 (2021) 2565-2579, DOI 10.1140/epjs/s11734-021-00268-w, arXiv:2108.04861, https://arxiv.org/abs/2108.04861
28. E. Laenen, C. Marinissen, M. Vonk, 'Resurgence analysis of the Adler function at order $1/N_f^2$', Journal of High Energy Physics 09 (2023) 103, DOI 10.1007/JHEP09(2023)103, arXiv:2302.13715, https://arxiv.org/abs/2302.13715