Inadequacy of zero-width approximation for a light Higgs boson signal

TL;DR

The paper demonstrates that the commonly used zero-width approximation for the Higgs boson can lead to significant errors, especially when off-shell effects and interference are important, affecting precise signal normalization.

Contribution

It reveals the limitations of the zero-width approximation for a light Higgs and emphasizes the importance of including off-shell contributions for accurate predictions.

Findings

Off-shell effects can cause 5-10% corrections in Higgs signal predictions.

Interference effects are sizable in certain Higgs decay channels.

Experimental cuts can mitigate off-shell contributions in Higgs searches.

Abstract

The zero-width approximation (ZWA) restricts the intermediate unstable particle state to the mass shell and, when combined with the decorrelation approximation, fully factorizes the production and decay of unstable particles. The ZWA uncertainty is expected to be of O(Gamma/M), where M and Gamma are the mass and width of the unstable particle. We review the ZWA and demonstrate that errors can be much larger than expected if a significant modification of the Breit-Wigner lineshape occurs. A thorough examination of the recently discovered candidate Standard Model Higgs boson is in progress. For M_H ~ 125 GeV, one has Gamma_H/M_H < 10^(-4), which suggests an excellent accuracy of the ZWA. We show that this is not always the case. The inclusion of off-shell contributions is essential to obtain an accurate Higgs signal normalization at the 1% precision level. For gg -> H -> VV, V = W,Z, O(5-10%) corrections occur due to an enhanced Higgs signal in the region M_VV > 2 M_V, where also sizable Higgs-continuum interference occurs. We discuss how experimental selection cuts can be used to suppress this region in search channels where the Higgs mass cannot be reconstructed. We note that H -> VV decay modes in non-gluon-fusion channels are similarly affected.