Mar 17, 2025
First "polarized lenses" for the scattered "light" of the weak interaction

Max Stange (left), Mareen Hoppe (center) and Erik Bachmann (right) symbolize transverse T-, longitudinal L- and once again transverse T-polarization with their arms.
Whether fishing, sailing, canoeing, kitesurfing, windsurfing or SUP: with polarized goggles, light reflections from the water surface can be eliminated, eye strain is reduced, you keep a clear view and don't miss any big fish under the water surface. The reason for this is that the electromagnetic fields of reflected light waves are aligned ("polarized") in a very specific direction, which can be filtered out with a suitable goggle structure. You get less scattered light in your eyes and can see below the surface of the water.
In proton-proton collisions at the Large Hadron Collider (LHC) at CERN, the analogon for the photons of the electromagnetic interaction are the W particles emitted during the sub-nuclear weak interaction. In contrast to the light of the electromagnetic interaction that we are all familiar with, these heavy W particles not only have two transverse (T) polarizations with respect to the direction of flight, but also an additional longitudinal (L) polarization in the direction of flight. This L-polarization would increase in the W±W± scattering with higher energy without any limitation - and thus lead to a violation of the basic rules of physics - if the contribution of the Higgs boson (Fig. 1) did not compensate exactly for the increase in L-polarization.

Abb. 1: The contribution of the Higgs boson (H) in the scattering of two W± particles, each emitted by a quark (q) in protons colliding in the LHC.
If the Higgs had not already been found, measuring the L-polarization in the W±W± scattering would have proved or disproved its existence in any case. This "no-lose theorem of W±W± scattering" was scientifically decisive in the decision to build the 27-kilometre LHC. However, even after the Higgs discovery, the measurement of the fully L-polarized W±LW±L (LL) component in the W±W± scattering remains the "big fish" among the LHC's goals, as it allows further conclusions to be drawn about the properties of the novel Higgs boson and tests the consistency and completeness of the underlying Brout-Englert-Higgs mechanism.
In 2014, the Dresden particle physics group, together with other researchers from the ATLAS experiment, was able to prove for the first time that W± particles scatter from one another, whereas light beams are known to pass through each other unhindered. The Dresden researchers have now succeeded in observing these scattered W particles with "polarized lenses" separating different polarization states from one another.
Using the data recorded between 2015 and 2018 from the ATLAS experiment in the so-called "Run 2" of the LHC, they have developed a method to separately measure the respective proportions of their polarizations (TT, TL/LT and LL) from the energy and angular distribution of the two W particles produced in the W±W± -> W±W± scattering. Highly optimized artificial intelligence in the form of a combination of two deep neural networks (DNN) was used. Fig. 2 shows that the "Inclusive DNN" enriches the W±W± -> W±W± scattering (blue) especially in bin 2 and within this bin the "Signal DNN" then accumulates the TL/LT components (light blue) and the LL component (medium blue) on the right. This results for the first time in a significantly non-zero sum of the TL/LT+LL components (26% of the observed total scattering cross-section of 3.4 fb, corresponding to (0.9±0.3) fb) and thus confirms the hypothesis that L-polarized W-particles are produced in this scattering.

Fig. 2: Distribution of observed events in bins of two deep neural networks (DNN). The colored areas represent the best fit of the contributions of WW-scattering (blue) and background (colored) to the data (black dots with error bars).
With these "polarized lenses", it was thus possible to filter out the distracting TT scattered "light" (dark blue) and take a first look at the L-polarization. The pure LL scattering (medium blue, the "fat fish" under the water surface) has not yet been observed significantly, but an upper limit of 0.45 fb can be derived for it, not much higher than the (0.29±0.07) fb (10% of the predicted total scattering cross-section) expected with the Higgs properties of the standard model. This already places strong constraints on theories beyond the Standard Model. A significant establishment and measurement of the LL fraction will allow a very precise verification of the Higgs properties, but will require significantly more data and/or more advanced "polarized lenses", which will be collected and developed at the LHC in the coming years.
In the development of this new method, three doctoral students in the Dresden group led by Frank Siegert played a leading role in various aspects of data analysis, without whose contributions the development would not have been possible:
Max Stange, as the main person responsible, worked in an international research team with colleagues from Germany, China, the UK and the USA and teased out the highest possible significance from the available data. At the same time, he paid meticulous attention to ensuring that the systematic uncertainties were reliably estimated. He was supported by Erik Bachmann, whose expertise in machine learning and DNN parameter optimization helped to separate the tiny measurement signal from the background. Mareen Hoppe provided the theoretical input required for this method. Her novel development of a simulation of polarized W-production in the SHERPA program package, which was created in Dresden and is now used worldwide, enabled a robust theoretical interpretation of the measurement for the first time.
The relevance of these contributions is also reflected in the fact that all three Dresden researchers have been invited to give presentations at highly-renowned conferences in the coming weeks. Max Stange will present this result as an ATLAS highlight on March 29 at one of the most important conferences in particle physics in Moriond. Erik Bachmann has been invited to give a lecture at the annual meeting of the largest European research action in this field, COMETA, in Krakow in April. As both experiment and theory are involved in this network, Mareen Hoppe will also be attending and has been invited to present the latest theoretical developments in the entire field.
This successful interplay between experiment and theory is a special feature of the Dresden Chair of Particle Physics. For further measurements at the LHC, it will be increasingly important to tease out greater significance from existing and future data through methodological and theoretical improvements.
Authors: Michael Kobel and Frank Siegert
Further information:
Original publication:
https://cds.cern.ch/record/2926878
https://arxiv.org/abs/2503.11317
Conference presentations:
Max Stange, 3/29/2025 https://moriond.in2p3.fr/2025/EW/Program.html
Erik Bachmann and Mareen Hoppe, 28-30.4.2025: https://indico.cern.ch/e/cometa-gm2
Research on WW scattering at the TU Dresden:
https://tu-dresden.de/mn/physik/iktp/arbeitsgruppen/teilchenphysik/forschung
LHC and ATLAS-Germany:
https://lhc-deutschland.de/aktuelles
https://lhc-deutschland.de/lhc_deutschland/fsp_atlas
WW scattering at ATLAS at CERN:
https://atlas.cern/updates/feature/vector-boson-scattering
Contact persons:
Erik Bachmann
Mareen Hoppe
Max Stange