Gildener-Weinberg two-Higgs-doublet model at two loops

The Gildener-Weinberg two-Higgs doublet model (GW-2HDM) provides a naturally light and aligned Higgs boson, $H = H(125)$. It has been studied in the one-loop approximation of its effective potential, $V_1$. An important consequence is that the masses of the model's BSM Higgs bosons ($H',A,H^\pm$) are bounded by the sum rule $\left(M_{H'}^4 + M_A^4 + 2M_{H^\pm}^4\right)^{1/4} = 540\,{\rm GeV}$. Although they are well within reach of the LHC, searches for them have been stymied by large QCD backgrounds. Another consequence is that $H$ is highly aligned, i.e., $H$--$H'$ mixing is small and $H$ has only Standard Model couplings. A corollary of this alignment is that search modes such as $H',\,A \leftrightarrow W^+W^-,\,ZZ,\,HZ$ and $H^\pm \leftrightarrow W^\pm Z,\,W^\pm H$ are greatly suppressed. To assess the accuracy of the sum rule and Higgs alignment, we study this model in two loops. This calculation is complicated by having many new contributions. We present two formulations of it to calculate the $H$--$H'$ mass matrix, its eigenvectors $H_1,\,H_2$, and the mass $M_{H_2}$ while fixing $M_{H_1}= 125\,{\rm GeV}$. They give similar results, in accord with the one-loop results. Requiring $M_A = M_{H^\pm}$, we find $180\,{\rm GeV}M_{H_2}>125\,{\rm GeV}$, with $M_{H_2}$ decreasing as $M_{A,H^\pm}$ increase. The corrections to $H$-alignment are below ${\cal{O}}(1\%)$. So, the BSM searches above will remain fruitless. Finding the BSM Higgses requires improved sensitivity to their low masses. We discuss two possible searches for this.