Positivity at 1-Loop: Bounds on Photon and Gluon EFTs
Jay Desai, Diptimoy Ghosh · 2025
Read Positivity at 1-Loop: Bounds on Photon and Gluon EFTs by Jay Desai, Diptimoy Ghosh (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-21864.txt.
Introduction
Jay Desai and Diptimoy Ghosh pushed positivity bounds for photon and gluon theories to one loop. Their paper, from twenty twenty-five, deals with bounds that tie a low energy theory to any sane high energy whole. At tree level such bounds are stock. At one loop the photon case stays finite and clean in the forward limit, but old tries hit a snag: the arc term hung on a fake low scale.
A close read of the high side of the spread link kills that hang. The bound comes out safe and flow fixed. The gluon case, with worse flaws, still yields a sound bound. Readers in field theory get loop grade fences for light and strong force toys.
Catalog record
| CREATOR | Jay Desai, Diptimoy Ghosh |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 61 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-21864
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Positivity at 1-Loop: Bounds on Photon and Gluon EFTs Jay Desai, Diptimoy Ghosh Abstract In this paper, we attempt to derive “positivity” bounds on Photon and Gluon Effective Field Theories (EFTs) at one loop level. While for the Photon case, the one loop amplitude is IR finite and well defined in the forward limit, earlier studies failed to obtain a dispersive bound on dimension-12 operators due to the dependence of the “arc integral” on the artificial low-energy scale. We show that this awkward dependence can be taken care of by analysing the ultra-violet (UV) side of dispersion relation closely. In particular, we derive an IR safe and RG improved bound at 1-loop. Thereafter, we perform a similar analysis on the Gluon EFT, which has additional complications due to ill-defined forward limit and IR divergences at 1-loop. We show that even in this case, one can get a meaningful bound at 1-loop. 1 Introduction Effective Field Theories (EFTs) (see [ 1 , 2 ] for a review) play a crucial role in understanding low energy physics. From a reductionist point of view, they can be obtained by integrating out heavy degree(s) of freedom from a full “UV complete” theory. However, an effective field theorist can never have access to physics at the shortest scales. So, one writes down an EFT based on symmetries such as Lorentz Invariance, Gauge Invariance, etc. with arbitrary coefficients. The EFT is understood to be a low energy description of physics below a scale \Lambda . It is an expansion in E/\Lambda (where E is the relevant low energy scale of the process of interest), with higher dimensional operators suppressed by higher powers of E/\Lambda . So one can work at a given order by including only a finite number of operators in computations, and hence, only a finite number of coefficients, which can be measured by low energy experiments. Without the knowledge of the corresponding UV complete theory, these coefficients are a-priori arbitrary. One can ask if they can be constrained using general principles, without assuming any particular UV completion. In fact, in [ 3 ] , the authors showed that not all low-energy effective field theories described by a local, Lorentz invariant Lagrangian are compatible with a microscopic S-matrix satisfying the usual analyticity conditions and Unitarity. These constraints are known as positivity constraints in the literature. We first review the analysis of [ 3 ] , as relevant to this paper. Lorentz invariance, micro-causality and Unitarity of the S -matrix in the full UV theory and some (well-motivated) fundamental assumptions on the scattering amplitudes and their analytic structure, can be exploited to derive a relation between an IR contribution of an integral involving scattering amplitude (i.e. at low energies, where the EFT is valid) to a UV integral of its imaginary part, up to arbitrarily high energies. This relation between IR and UV parts of the amplitude is called a dispersion relation . The UV side cannot be fully computed within the EFT but its positivity properties (which crucially relies on Unitarity of the full UV theory) can be exploited, which puts non-trivial constraints on the EFT coefficients of operators involved in the process, which a priori could have taken any values. One makes the following assumptions: (i) The forward amplitude as a function of s ( \mathcal{A}(s) ), in the complex s plane, is analytic in the upper half plane, which follows from causality [ 4 , 5 , 6 ] (see [ 7 , 8 ] for a review and Sec 1.2 of [ 9 ] for a pedagogical explanation). The singularities are all on the Real- s axis such as simple poles (due to tree-level exchanges) and branch cuts (due to logs at loop-level). That the amplitude is well defined at arbitrarily high energies is a consequence of Boost invariance. See [ 10 , 11 , 12 , 13 ] for discussions on the fate of such bounds in theories without Lorentz Invariance. (ii) The Froissart-Martin Bound ( \mathcal{A}(s)|_{s\rightarrow\infty}<\text{ const. }s\log^{2}(s) ) [ 14 , 15 , 16 ] , which was proved rigorously for only gapped theories, also holds for mass-less exchanges. To make this more believable, we can use a mass regulator by assigning a mass M to mass-less particles in the theory and at the end, take the limit M\rightarrow 0 . We can make a slightly weaker assumption: \lim_{s\rightarrow\infty}\mathcal{A}(s)/s^{2}=0 , known as “Regge Bound” in the literature (see [ 17 ] for a discussion). Consider a 2\rightarrow 2 scattering process AB\rightarrow CD with the Mandelstam variables defined as usual: s=(p_{A}+p_{B})^{2},~t=(p_{C}-p_{A})^{2},~u=(p_{D}-p_{A})^{2} with s+t+u=m_{A}^{2}+m_{B}^{2}+m_{C}^{2}+m_{D}^{2} where A,B,C and D are one-particle states with 4-momenta p_{A},p_{B},p_{C} and p_{D} respectively. Since s,t and u satisfy this constraint, a 2\rightarrow 2 scattering amplitude \mathcal{M} is a function of only two Mandelstam variables, usually chosen as s and t . The scattering amplitude in the forward limit is denoted as \lim_{t\rightarrow 0}\mathcal{M}(s,t)\equiv\mathcal{A}(s) In this paper, since we are working with massless particles, the poles introduced at tree level (if any) will be all at origin. In case of massless exchanges, it is important to choose an appropriate 2\rightarrow 2 process such that the t\rightarrow 0 limit is well defined. s_{0} Re(s) Im(s) \cap_{s_{0}} \cap_{\infty} Consider the contour as shown above, as was first considered in [ 18 ] . Since this does not enclose any residues, we have \oint ds\frac{\mathcal{A}(s)}{s^{2n+1}}=0 for some positive integer n . We can decompose this as follows: \oint=\int_{\infty}^{s_{0}}+\int_{\cap_{s_{0}}}^{~}+\int_{-s_{0}}^{-\infty}+\int_{\cap_{\infty}} where second and fourth integrals are over semi-circular arcs of radius s_{0} and \infty respectively. After a substitution s\rightarrow-s in the third integral, it can be clubbed with the first. 0=\int_{s_{0}}^{\infty}ds~\frac{\mathcal{A}(-s+i\epsilon)-\mathcal{A}(s+i\epsilon)}{s^{2n+1}}+\int_{\cap_{s_{0}}}ds~\frac{\mathcal{A}(s)}{s^{2n+1}}+\int_{\cap_{\infty}}ds~\frac{\mathcal{A}(s)}{s^{2n+1}} The last term on the RHS - the integral over the arc at infinity - gives zero (for n>0 ) due to Regge boundedness. Then, we have \int_{\cap_{s_{0}}}ds~\frac{\mathcal{A}(s)}{s^{2n+1}}=\int_{s_{0}}^{\infty}ds~\frac{\mathcal{A}(s+i\epsilon)-\mathcal{A}(-s+i\epsilon)}{s^{2n+1}} With analytic continuation (or real analyticity: \mathcal{A}(s)=\mathcal{A}^{*}(s^{*}) ) and choosing an s - u crossing-symmetric amplitude [ 19 , 20 , 21 ] , ( \mathcal{A}(s)=\mathcal{A}(-s) ), we have for s on real axis: \mathcal{A}(s+i\epsilon)=\mathcal{A}^{*}(s-i\epsilon)=\mathcal{A}^{*}(-s+i\epsilon) , (which we refer to as “crossing symmetry”) so that we can relate the negative s branch to the (physical) positive one. Under this assumption, we get \int_{\cap_{s_{0}}}\frac{ds}{i\pi}~\frac{\mathcal{A}(s)}{s^{2n+1}}=\frac{2}{\pi}\int_{s_{0}}^{\infty}ds~\frac{Im\mathcal{A}(s)}{s^{2n+1}} (1) We make use of the Optical Theorem, which for massless particles in the initial state reads: Im\mathcal{A}(s)=s\sigma(s) . So, we finally get the dispersion relation \int_{\cap_{s_{0}}}\frac{ds}{i\pi}~\frac{\mathcal{A}(s)}{s^{2n+1}}=\frac{2}{\pi}\int_{s_{0}}^{\infty}ds~\frac{\sigma(s)}{s^{2n}} (2) One can take s_{0}<\Lambda^{2} , so that the LHS can be computed within the EFT, in terms of Wilson coefficients of EFT operators ( \mathcal{A}=\mathcal{A}_{\text{EFT}} for s<\Lambda^{2} ). The integer n has to be chosen to get the desired contribution of Wilson coefficients from the LHS, which we can then constrain to be positive due to the RHS being positive. \textit{Amplitude's Positivity:}~~~\int_{\cap_{s_{0}}}\frac{ds}{i\pi}~\frac{\mathcal{A}_{\text{EFT}}(s)}{s^{2n+1}}>0 (3) For instance, consider a tree level contribution \mathcal{A}\sim c_{{}_{2}}\frac{s^{2}}{\Lambda^{4}}+c_{{}_{6}}\frac{s^{6}}{\Lambda^{12}}+... Then, n=1 in ( 3 ) gives c_{{}_{2}}>0 and n=3 gives c_{{}_{6}}>0 . This analysis has been explicitly done at tree level in various contexts (e.g. [ 3 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 ] ) in the forward limit and [ 33 , 34 , 35 , 36 , 37 , 38 ,