The infrared problem in QED: a lesson from a model with Coulomb interaction and realistic photon emission (Q331469)

To study the infrared problem in QED the authors introduce a simplified model of a fixed number \(N\) of massive charged classical particles interacting with transverse photons of helicity \(\pm 1\). The particles are described with the Koopman formalism and the photons with the formalism of Fock space, \(\mathcal{H}_F\) based on \(L^2(\mathbb{R}^3, dk)\). The Hilbert space of the system is therefore \(\mathcal{H}^N=L^2(\Gamma^N;\mathcal{H}_F)\) where \(\Gamma^N\) is phase space for \(N\) particles. The particles interact with each other through the Coulomb potential and with the photons via a Hamiltonian of the form \(a(f(q,p)) + \mathrm{h.c.}\). In this, \(a\), \(a^*\) are photon annihilation and creation operators and \(f(q,p)\) is a test function of a very particular form. It depends on the particle velocity function, the polarization vectors and a rotationally invariant function normalized to \(1\) at the origin and acting as a uv cutoff. Variants of it appear throughout the article. There is also a subtraction of a \(\mathcal{C}^1\) function of \(p\) which is a mass counter term. The choice of this function is made further on so as to cancel logarithmic infrared divergences. The Hamiltonian is invariant under the space translations from which it is deduced that the particles suffer no recoil as a result of emitting photons. In view of the non-triviality of the final \(S\)-matrix and many other results, one has to agree with the authors that this is a realistic interaction. Scattering theory is based on comparing the asymptotic behavior of quantities evolving under interaction with those same quantities evolving under a reference dynamics. For short range interactions the free evolution is used but that is unsustainable for the long range Coulomb interaction. A certain type of replacement of the free evolution in potential scattering was the suggestion of \textit{J. D. Dollard} [``Asymptotic convergence and the Coulomb interaction'', J. Math. Phys. 5, 729--738 (1964); Rocky Mt. J. Math. 1, 5--88 (1971; Zbl 0226.35074)]. This has subsequently been extended to QED by various authors with varying degrees of success -- though these efforts have been of value. A crucial part of the authors' modelling is the choice of reference dynamics. Judging by their results, the authors have made an effective choice. They stress the need to accompany this with both regularizing counter terms and an adiabatic switch. Starting with the particles on their own, then including more of the model in stages, they end up with a proof of the existence of Möller operators and a non-trivial \(S\)-matrix as strong limits \(t\to \pm \infty\). In the process, regularizing terms and a switch are added and then removed in a sequence of strong limits: no divergences remain at the end. The article continues with an application to LSZ asymptotics ``in the presence of Coulomb interactions and infinite photon emissions''. To this end the model is extended by replacing \(\mathcal{H}^N\) with its antisymmetric subspace and then taking the direct sum of copies for all positive integers \(N\); the photon component remains the same. One is now considering a field theoretic model over a compound Fock space where for every \(n\in \mathbb{N}\), the \(n\)-particle component is a copy of the original model with \(N=n\). The Hamiltonians and time translations which are the analogues of the \(N\)-particle model are determined with the help of the second quantization calculus and good judgement. The particle motions are relativistic and the annihilation and creation operators \(\Psi^\sharp\) for the particles satisfy the CAR. The determination of the time translations, including the Dollard reference dynamics, and applying and removing an adiabatic switch are obtained us methods similar to the \(N\)-particle case. There is much in this second part. On consideration, the reviewer feels that the best way to summarize the results for this part is quoting the abstract of the article in part. Then ``i) asymptotic limits of the charged fields. \(\Psi_{out/in}(x)\), are obtained as strong limits of modified LSZ formulas, with corrections given by a Coulomb phase operator and an exponential of the photon field; ii) free asymptotic electromagnetic fields, \(B_{out/in}(x)\), are given by the massless LSZ formula, as in the Buchholz approach; iii) the asymptotic field algebras are semidirect products of the canonical field algebras generated by \(B_{out/in}(x)\) and \(\Psi_{out/in}(x)\); iv) on the asymptotic spaces the Hamiltonian is sum of free (commuting) Hamiltonians of \(B_{out/in}(x)\) and \(\Psi_{out/in}(x)\) and the same holds for the generators of space translations.'' In summary, this is a densely written forty page paper containing a carefully rendered treatment of the infrared problem in a model which closely mimics the corresponding portion of QED. The authors justifiably refer to the critical part of their model as a ``realistic interaction''. They have moved forward from the work of the authors they cite, adapting earlier work to the requirements of the model when appropriate and correcting it when necessary. This article has impressed the reviewer. Together with previous work of theirs, especially their reference [the authors, Rev. Math. Phys. 28, No. 1, Article ID 1650001, 26 p. (2016; Zbl 1335.81128)], and in anticipation of the paper they have promised at the end of section 6, Morchio and Strocchi have advanced the treatment of the infrared problem. Their work should be considered a programme. The reviewer recommends this article to everyone interested in the infrared problem and rigorous results in QED.