Zγ production at NNLO
Tobias Neumann, University at Buffalo
in collaboration with John Campbell and Ciaran Williams
June 1, 2017
Motivation / Overview
- Second implementation of \(Z\gamma\) at NNLO
using jettiness-subtractions (Gaunt, Stahlhofen, Tackmann, Walsh '15; Boughezal, Focke, Liu, Petriello '15)
- Validation of previous calculation (Grazzini, Kallweit, Rathlev, Torre ’13+’15)
using \(q_T\)-subtractions (Catani, Grazzini '07)
- How can we improve/extend two year old computation?
- Performance: Power corrections? Slicing parameter?
- Updated SM phenomenology
- BSM: Anomalous couplings & Dark Matter
Setup and assembly
- MCFM NNLO 0-jettiness subtractions implementation
(Boughezal, Petriello, et al. '16)
- Amplitude ingredients:
- 2-loop \(Z\gamma\) amplitudes (Gehrmann, Tancredi '12)
- Electron radiation: 2-loop form-factor (Matsuura, van der Marck, van Neerven '89)
- Tree and one-loop amplitudes reimplemented
(Campbell, Hartanto, Williams '12), (Bern, Dixon, Kosower '98)
- Anomalous \(Z\gamma\gamma\) and \(ZZ\gamma\) couplings:
- Vertices (Hagiwara, Peccei, Zeppenfeld, Hikasa '87)
- Four parton + V off-shell current (Berends, Giele, Kuijf '89)
- V decay to three partons one-loop current (Garland, Gehrmann, Glover '02)
and analytical continuation (Gehrmann, Remiddi '02)
- Our data points/performance:
- NLO: 150-250 cpu hours on 2009/2010 Intel Xeons
- NNLO: 600-1200 cpu hours (24 cores, 24-48 hours)
How to extrapolate the slicing parameter to zero?
(Grazzini, Kallweit, Rathlev ’15)
Non-linear fit to
$$\sigma_\text{NLO}^\text{asympt.} = \Delta\sigma_\text{NLO} + c_2 \hat\tau_\text{cut} \log^2\hat\tau_\text{cut} + \ldots$$
$$\sigma_\text{NNLO}^\text{asympt.} = \Delta\sigma_\text{NNLO} + c_3 \hat\tau_\text{cut} \log^3\hat\tau_\text{cut} + \ldots$$
\(\hat\tau_\text{cut} = \tau_\text{cut}/Q\)
Stat. uncertainty \(\gg\) model uncertainty
No higher-power corrections in fit.
(Moult, Rothen, Stewart, Tackmann, Zhu '16)
Power corrections
Subleading terms for \(q\bar q \to \text{color-singlet}\)
(Moult, Rothen, Stewart, Tackmann, Zhu '16), (Boughezal, Liu, Petriello '17)
Standard cuts
\(p_T^l>25\,\text{GeV},\)
\(p_T^\gamma>40\,\text{GeV},\)
\(\Delta R(l,\gamma)\)>0.7
Loose cuts
\(p_T^l>20\,\text{GeV},\)
\(p_T^\gamma>10\,\text{GeV},\)
\(\Delta R(l,\gamma)>0.1\)
Jettiness definition: Boosted vs. hadronic center of mass system
Boosted: 0-jettiness \(\tau_0\) defined in color singlet c.o.m. frame
MCFM-8.0: \(\tau_0\) defined in hadronic c.o.m. frame
Comparison to Grazzini, Kallweit, Rathlev ’15
Charged lepton decay
Neutrino decay
Comparison for neutrino decay
Our NLO results agree to better than 0.2%
\(\sigma_\text{NNLO} = 80.7(1)~\text{fb}\), \(\sigma_\text{NNLO}^\text{GKR} = 80.8(4)~\text{fb}\)
(0.1% num. uncertainty)
Comparison for charged lepton decay
\(\sigma_\text{NNLO} = 178.6(4)~\text{fb}\), \(\sigma_\text{NNLO}^\text{GKR} = 180(1)~\text{fb}\)
(0.2% num. uncertainty)
How about \(13~\text{TeV}\) results?
(ATLAS neutrino-channel cuts for \(13~\text{TeV}\))
Total cross section for \(Z\to \nu\bar\nu\): NLO coefficient
Total cross section for \(Z\to \nu\bar\nu\): NNLO coefficient
\(\sigma_\text{NNLO} = 86.0(2)\text{fb}\) (0.2% num. uncertainty)
Total cross section for \(Z\to l\bar l\): NNLO coefficient
\(\sigma_\text{NNLO} = 306(2)\text{fb}\) (0.7% num. uncertainty)
\(13~\text{TeV}\,\, p_T^\gamma\) for \(Z\to \nu\bar\nu\)
\(13~\text{TeV}\,\, m_T^{\nu\bar\nu\gamma}\) for \(Z\to \nu\bar\nu\)
Anomalous \(ZZ\gamma\) and \(Z\gamma\gamma\) couplings
ATLAS, 1604.05232
using NNLO SM prediction (Grazzini, Kallweit, Rathlev '15) + NLO anomalous couplings (MCFM-8.0)
Anomalous Couplings: \(h_3^Z\), \(13~\text{TeV}\, Z\to\nu\bar\nu\)
"Now": NNLO SM + NNLO anomalous couplings in MCFM
\( p_T^\gamma>400~\text{GeV}\), no jets
No form factor applied. Very similar results for \(h_4^\gamma\).
How much do current limits change?
Another topic: \(Z(\to\nu\bar\nu)\gamma\) mono-photon signal:
Background for Dark Matter
Summary
- Second implementation of \(Z\gamma\) at NNLO
- Successful verification of previous results ✔
- Numerically stable and fast even without power corrections ✔
(~1000 cpu hours for sub-percent precision NNLO results)
- Ready for \(13~\text{TeV}\) phenomenology
- NNLO Standard Model ✔
- NNLO BSM: Anomalous \(ZZ\gamma\) \(Z\gamma\gamma\) couplings ✔
-
Public in MCFM (soon) and usable by non-experts (already)
(no excessive parameter tinkering)