AtlasPublic Web 
Internal TWiki
|
Observation of quantum entanglement in top-quark pair production using $pp$ collisions of √ s = 13 TeV with the ATLAS detector ATLAS-CONF-2023-069 28 September 2023
|
||
| Content | Preview | |
|---|---|---|
| Main document (CDS record), Physics Briefing - internal | pdf from CDS | |
| Figures Tables | - | |
| Abstract | ||
| We report the highest-energy observation of entanglement so far in top--antitop quark events produced at the Large Hadron Collider, using a proton$-$proton collision data set with a centre-of-mass energy of $\sqrt{s}=13$ TeV and an integrated luminosity of 140 fb$^{-1}$. Spin entanglement is detected from the measurement of a single observable $D$, inferred by the angle between the charged leptons in their parent top- and antitop-quark rest frames. The observable is measured on a narrow interval around the top-quark--antitop-quark production threshold, where the entanglement detection is expected to be significant. The entanglement observable is measured in a fiducial phase-space with stable particles. The entanglement witness is measured to be $D = -0.547 \pm 0.002 \text{(stat.)} \pm 0.021 \text{(syst.)}$ for $340 < m_{t\bar{t}} < 380$ GeV. The large spread in predictions from several mainstream event generators indicates that modelling this property is challenging. The predictions depend in particular on the parton-shower algorithm used. The observed result is more than five standard deviations from a scenario without entanglement and hence constitutes the first observation of entanglement in a pair of quarks, and the observation of entanglement at the highest energy to date. | ||
| Figures | ||
|
Figure 01: In the left panel, the cos(φ) observable in the signal region at reconstructed level and in the right, the entanglement witness D at the reconstructed detector-level distribution from different three different MC generators; the POWHEGBOX + PYTHIA and and POWHEGBOX + HERWIG heavy quark models, labelled "Pow+Py (hvq)" and "Pow + H7 (hvq)", respectively, and the POWHEGBOX + PYTHIA bb4ℓ models, labelled "Pow+Py (bb4ℓ)", after backgrounds are subtracted. The uncertainty band includes all sources of uncertainty added in quadrature. The ratio of the predictions with respect to data is shown at the bottom of the figure. The quoted value for D for the bb4ℓ model also includes a subtraction of the single-top-quark background. png (123kB) pdf (28kB) |
 |
|
|
Figure 02a: (a): Calibration curve for the dependence between the particle-level value of D to the reconstructed value of D, in the signal region. The yellow band represents the statistical uncertainty, while the gray band represents the total uncertainty due to statistical and systematic uncertainties. The total uncertainty is obtained by adding in quadrature the statistical and all the systematic uncertainties. The measured and the expected values are marked with black and red circles, respectively, and the entanglement limit is shown as a dashed line. (b): The particle-level D results for the signal and validation regions compared to various MC models. The entanglement limit shown is a conversion from its parton-level value of D = -1/3 to the corresponding value at particle level, and the uncertainties which are considered for the band are described in the text. png (56kB) pdf (25kB) |
 |
|
|
Figure 02b: (a): Calibration curve for the dependence between the particle-level value of D to the reconstructed value of D, in the signal region. The yellow band represents the statistical uncertainty, while the gray band represents the total uncertainty due to statistical and systematic uncertainties. The total uncertainty is obtained by adding in quadrature the statistical and all the systematic uncertainties. The measured and the expected values are marked with black and red circles, respectively, and the entanglement limit is shown as a dashed line. (b): The particle-level D results for the signal and validation regions compared to various MC models. The entanglement limit shown is a conversion from its parton-level value of D = -1/3 to the corresponding value at particle level, and the uncertainties which are considered for the band are described in the text. png (62kB) pdf (51kB) |
 |
|
|
Figure 03: Example of the results of the nominal distribution and the reweighting technique with scaling hypotheses X = 0.4, 0.6, 0.8, 1.2 in the m(tt̄) < 380 GeV region at parton level. The ratio shows the various reweighting prediction point "Pred." over the nominal "Nom." png (59kB) pdf (35kB) |
 |
|
|
Figure 04a: Comparison of cos(φ) (inclusive) for (a) POWHEGBOX and (b) HERWIG setups at particle level after an inclusive tt̄ selection but without any requirements on the invariant mass of the top-quark pair. png (61kB) pdf (34kB) |
 |
|
|
Figure 04b: Comparison of cos(φ) (inclusive) for (a) POWHEGBOX and (b) HERWIG setups at particle level after an inclusive tt̄ selection but without any requirements on the invariant mass of the top-quark pair. png (61kB) pdf (35kB) |
 |
|
|
Figure 05a: Comparison of cos(φ) in the entanglement region for (a) POWHEGBOX and (b) HERWIG setups at particle level in the signal region. png (66kB) pdf (25kB) |
 |
|
|
Figure 05b: Comparison of cos(φ) in the entanglement region for (a) POWHEGBOX and (b) HERWIG setups at particle level in the signal region. png (69kB) pdf (36kB) |
 |
|
|
Figure 06: The cos(φ) observable in validation region 1 at reconstructed level (left) and the D at the reconstructed detector level distribution from different three different MC generators; the POWHEGBOX + PYTHIA and and POWHEGBOX + HERWIG heavy quark models, labelled "Pow+Py (hvq)" and "Pow + H7 (hvq)", respectively, and the POWHEGBOX + PYTHIA bb4ℓ models, labelled "Pow+Py (bb4ℓ)", after backgrounds are subtracted (right). The uncertainty band includes all sources of uncertainty added in quadrature. The ratio of the predictions with respect to data is shown at the bottom of the figure. The D quoted for the bb4ℓ model also includes a subtraction of the single-top-quark background. png (124kB) pdf (28kB) |
 |
|
|
Figure 07: The cos(φ) observable in the validation region 2 at reconstructed level (left) and the D at the reconstructed detector level distribution from different three different MC generators; the POWHEGBOX + PYTHIA and and POWHEGBOX + HERWIG heavy quark models, labelled "Pow+Py (hvq)" and "Pow + H7 (hvq)", respectively, and the POWHEGBOX + PYTHIA bb4ℓ models, labelled "Pow+Py (bb4ℓ)", after backgrounds are subtracted (right). The uncertainty band includes all sources of uncertainty added in quadrature. The ratio of the predictions with respect to data is shown at the bottom of the figure. The D quoted for the bb4ℓ model also includes a subtraction of the single-top-quark background. png (122kB) pdf (28kB) |
 |
|
| Tables | ||
|
Table 01: A summary of the effect of the groups of uncertainty at the expected SM value for Dexpected=-0.470, corresponding to the Pow+Py modelling, and the observed value Dobserved=-0.547, both in the signal region. The total systematic uncertainty is calculated as the sum in quadrature of the individual grouped sources of systematic uncertainty. png (34kB) pdf (58kB) |
 |
|
|
Table 02: A comparison of the relative size of the uncertainties related to signal modelling at the SM expectation point with respect to Dparticle. The PDF uncertainties are the quadrature sum of all variations. png (23kB) pdf (48kB) |
 |
|
|
Table 03: Summary of the top 10 most important systematic uncertainties for the signal region, in the measurement at the SM expectation. png (25kB) pdf (48kB) |
 |
|
|
Table 04: Summary of the top 10 most important systematic uncertainties for the signal region, in the measurement at the -20 of the SM expectation. png (23kB) pdf (47kB) |
 |
|
