PHOTOS Monte Carlo: a precision tool for QED corrections in Z and W decays

TL;DR

The study assesses the precision of PHOTOS, a universal Monte Carlo for QED radiative corrections in particle decays, by extensive comparisons with established reference generators using MC-TESTER. It demonstrates that PHOTOS, including iterative and exponentiated multi-photon emission and a universal interference weight, reaches ~0.1% accuracy for Z and W leptonic decays across full final states. The results validate PHOTOS as a high-precision tool for radiative corrections in common LHC-era processes, while highlighting the need for process-specific checks for broader applicability. The work also documents algorithmic improvements, stability enhancements, and high-energy feasibility, with implications for precision measurements in Higgs, CKM studies, and beyond.

Abstract

We present a discussion of the precision for the PHOTOS Monte Carlo algorithm, with improved implementation of QED interference and multiple-photon radiation. The main application of PHOTOS is the generation of QED radiative corrections in decays of any resonances, simulated by a "host" Monte Carlo generator. By careful comparisons automated with the help of the MC-TESTER tool specially tailored for that purpose, we found that the precision of the current version of PHOTOS is of 0.1% in the case of Z and W decays. In the general case, the precision of PHOTOS was also improved, but this will not be quantified here.

Figures (8)

• Figure 1: Predictions of KORALZ (with $O(\alpha)$ matrix-element and single-photon emission) are compared with predictions of PHOTOS (running in single-photon option) for the $Z^{0}\rightarrow\mu^{+}\mu^{-}(\gamma)$ channel, $E_\mathrm{test}=1.0$ GeV. The plot presents the distribution of invariant mass of $\mu^- \mu^+$ pair coming from KORALZ (in red, or darker-grey) and PHOTOS (in green, or lighter-grey); the black line is the ratio of the two normalized distributions. The red and green lines are hard to separate - they practically overlap.
• Figure 2: Predictions for exact $O(\alpha)$ in TAUOLA are compared with single-photon emission of PHOTOS in the $\tau^- \rightarrow \mu^- \bar{\nu}_\mu \nu_\tau (\gamma)$ channel, $E_\mathrm{test}=0.05$ GeV. The plot presents the distribution of invariant mass of the $\bar{\nu}_\mu \nu_\tau \gamma$ system coming from TAUOLA (in red, or darker-grey) and PHOTOS (in green, or lighter-grey); the black curve is the ratio of the two normalized distributions.
• Figure 3: Comparison of predictions from KKMC ($O(\alpha^2)$ ME with exponentiation at the level of spin amplitudes) and KORALZ ($O(\alpha^2)$ ME with exponentiation) in the $Z^0 \rightarrow \mu^+ \mu^- (\gamma)$ channel, $E_{test}=1.0$ GeV, test1 used for analysis. The plot presents the distribution of invariant mass of $\mu^+ \mu^-$ pair coming from KKMC (in red, or darker-grey) and KORALZ (in green, or lighter-grey); the black line is the ratio of the two normalised distributions. The effects due to different types of exponentiation are small, albeit noticeable, and constitute 0.066% overall difference.
• Figure 4: Acoplanarity distributions (in the $Z^0 \rightarrow \mu^- \mu^+ \gamma \gamma$ channel) produced by the KKMC generator running in exponentiated $O(\alpha^2)$ mode (in red, or darker-grey) and exponentiated $O(\alpha)$ mode (in green, or lighter grey); the ratio of the two distributions plotted in black. For more details see subsection \ref{['sub:Iterating-emission-kernels']}.
• Figure 5: Acoplanarity distributions (in the $Z^0 \rightarrow \mu^- \mu^+ \gamma \gamma$ channel) produced by the KKMC generator running in exponentiated $O(\alpha^2)$ mode (in red, or darker-grey) and the exponentiated PHOTOS algorithm (in green, or lighter grey); the ratio of the two distributions is plotted as dotted black curve. For more details see subsection \ref{['sub:Iterating-emission-kernels']}.