-
The scalar potential of 2HDM [11] has 14 parameters, including the charge and CP violations. The general term for this scalar potential is
$ \begin{aligned}[b] V=&m^2_{11} \Phi^{\dagger}_1 \Phi_1 + m^2_{22} \Phi^{\dagger}_2 \Phi_2 - m^2_{12} \biggl\{ \Phi^{\dagger}_1 \Phi_2 + h.c. \} + \frac{\lambda_1}{2} (\Phi^{\dagger}_1 \Phi_1)^2 \\&+ \frac{\lambda_2}{2} (\Phi^{\dagger}_2 \Phi_2)^2 + \lambda_3 (\Phi^{\dagger}_1 \Phi_1)(\Phi^{\dagger}_2 \Phi_2) + \lambda_4 (\Phi^{\dagger}_1 \Phi_2) (\Phi^{\dagger}_2 \Phi_1) \\&+ \biggl\{ \frac{\lambda_5}{2} (\Phi^{\dagger}_1 \Phi_2)^2 + \left[ \lambda_6 (\Phi^{\dagger}_1 \Phi_1) + \lambda_7 (\Phi^{\dagger}_2 \Phi_2) \right] (\Phi^{\dagger}_1 \Phi_2) + h.c. \biggl\}, \end{aligned} $
(1) where
$ (\lambda_{i}, i = 1,2,3,...., 7) $ are dimensionless coupling parameters, and$ m^2_{11}, m^2_{22} $ , and$ m^2_{12} $ are squares of masses. To treat the 2HDM potential as charge and parity conserving potential, all the parameters should be real. The vacuum expectation value VEV is obtained by each scalar-doublet when the electroweak symmetry breaks. The two doublets are$ <\Phi_1>=\left(\begin{array}{*{20}{c}}{0}\\{\dfrac{V_1}{\sqrt{2}}}\end{array} \right),\quad <\Phi_2>=\left(\begin{array}{*{20}{c}}{0}\\{\dfrac{V_2}{\sqrt{2}}} \end{array}\right). $
(2) These two doublets lead to eight fields, among which three correspond to massive
$ W^{\pm} $ and$ Z^0 $ vector bosons, and the remaining five fields lead to five physical Higgs bosons:$ \Phi_i=\left(\begin{array}{*{20}{c}}{\Phi^+_i}\\{\dfrac{(V_i+\rho_i+\iota\eta_i)}{\sqrt{2}}}\end{array}\right), $
(3) where i = 1, 2 with
$V_1=V\cos\beta,~ V_2=V\sin\beta~ {\rm and}~ V_1,V_2 \geq 0$ . The condition$V_{\rm SM} = \sqrt{V^2_1+V^2_2}$ is satisfied. The experimentally obtained value of$ V_{SM} $ is 246.22 GeV. The obtained fields are expressed as$ \begin{aligned}[b] &\binom{\rho_a}{\rho_b}= \begin{pmatrix} \cos\alpha & -\sin\alpha\\ \sin\alpha & \cos\alpha \end{pmatrix} \binom{H}{h} , \\& \binom{\eta_a}{\eta_b} = \begin{pmatrix} \cos\beta & -\sin\beta\\ \sin\beta & \cos\beta \end{pmatrix} \binom{G^0}{A} \end{aligned} $
(4) and
$ \left(\begin{array}{*{20}{c}} {\Phi^{\dagger}_a}\\{\Phi^{\dagger}_b}\end{array}\right)= \begin{pmatrix} \cos\beta & -\sin\beta\\ \sin\beta & \cos\beta \end{pmatrix} \left(\begin{array}{*{20}{c}}{G^{\dagger}}\\{H^{\dagger}}\end{array}\right). $
(5) The mass-matrix of charged Higgs states is diagonalized by a rotational angle defined as
$\tan\beta = \dfrac{V_{2}}{V_{1}}$ . Similarly, the mass-matrix of scalar Higgs states is diagonalized by a rotational angle α that satisfies the relation$ \tan(2\alpha)=\frac{2(-m^2_{12}+\lambda_{345}V_1V_2)}{m^2_{12}(V_2/V_1 - V_1/V_2)+\lambda_1V^2_1-\lambda_2V^2_2} $
(6) where
$ \lambda_{345}=\lambda_3+\lambda_4+\lambda_5 $ . -
The general potential of 2HDM is complex compared to the one in the standard model SM. In particular, 2HDM imposes theoretical constraints on the potential to guarantee the stability of the potential.
The Higgs potential should be positive throughout the field space. This ensures that a stable vacuum configuration for asymptotically large field values is maintained. The quartic terms are the leading terms at large field space values. In this context, the following substitutions are helpful:
$\mid \Phi_1 \mid=r \cos \phi , \quad \mid\Phi_2\mid=r\sin\phi \quad {\rm and} \quad \frac{\Phi_1\Phi^{\dagger}_2}{\mid\Phi_1\mid\mid\Phi_2\mid} =\rho\exp(\iota\theta), $
(7) where
$\rho=\mid0-1\mid , \quad \theta=\mid0-2\pi\mid \quad {\rm and} \quad \phi=\mid0-\dfrac{\pi}{2}\mid$ . After making these substitutions and omitting the common factor$ r^4 $ , the quartic terms of the potential can be expressed as$ \begin{aligned}[b] V_{(4)}=&\frac{1}{2}\lambda_1 \cos^4\phi + \frac{1}{2}\lambda_2 \sin^4\phi + \lambda_3\cos^2\phi \sin^2\phi\\&+\lambda_4\rho^2\cos^2\phi \sin^2\phi + \lambda_5\rho^2 \cos^2\phi \sin^2\phi \cos2\theta \\&+ [\lambda_6 \cos^2\phi + \lambda_7 \sin^2\phi]2\rho \cos\phi \sin\phi \cos\theta. \end{aligned} $
(8) From the above equation, we can see that the potential will be positive if
$ \begin{aligned}[b]& \lambda_1>0 , \quad \lambda_2>0 , \quad \lambda_3>-\sqrt{\lambda_1\lambda_2}\; , \\& \lambda_3 + \lambda_4 - |\lambda_5| > -\sqrt{\lambda_1\lambda_2} \end{aligned} $
(9) If
$ \lambda_{6} = 0 = \lambda_{7} $ , there are necessary and sufficient conditions to ensure the positivity of the quartic potential.If either
$ \lambda_{6} \not= 0 $ or$ \lambda_{7} \not= 0 $ for the case they are real, one finds that$ 2 |\lambda_6+\lambda_7| ~ < ~ \frac{\lambda_1 + \lambda_2}{2}+\lambda_3+\lambda_4+\lambda_5 . $
(10) Scattering matrices are unitary in order to conserve probability. In the theory of weak couplings, the contribution of higher order terms decreases gradually. By contrast, in the theory of strong couplings, individual contributions increase arbitrarily. The eigenvalues (
$ L_i $ ) of S-matrices must satisfy the condition$ L_i \leq 16\pi $ in order to achieve the tree-level unitarity that means the saturation of S-matrices up to tree-level unitarity.Perturbation constraints require that the quartic Higgs couplings must satisfy the condition
$ \mid C_{H_{i},H_{j},H_{k},H_{l}} \mid \leq 4\pi $ . One can imagine that some interaction channels are non-perturbative while others are perturbative. Note that setting$ \mid \lambda_i \mid = 4\pi\xi $ is an alternative approach to explain this constraint, where$ \xi = 0.8 $ 1 . This sets$ \mid \lambda_i \mid \leq 10 $ for$ \lambda_i $ as the upper bound.Alongside theoretical constraints, there are also experimental constraints coming from B-Physics and various experiments on different colliders from recent Higgs searches. Here we discuss some important constraints. Moreover, using SuperIso V.3.2 [12], we list the SM prediction values provided in this category. The Standard Model BR for
$(B_\mu \rightarrow \tau\nu)_{\rm SM}$ reported in [12] is$ {\rm BR} (B_\mu\rightarrow \tau\nu)_{\rm SM} = (1.01 \pm 0.29) \times 10^{-4} . $
(11) The standard model estimation may, in fact, be contrasted with the most recent heavy flavour averaging (HFAG) result [13]:
$ {\rm BR} (B_\mu\rightarrow \tau\nu)_{\rm exp} = (1.64 \pm 0.34) \times 10^{-4}, $
(12) and the ratio becomes
$ R^{\rm exp}_{\rm SM} = \frac{{\rm BR}(B_\mu\rightarrow \tau\nu)_{\rm exp}}{{\rm BR}(B_\mu\rightarrow \tau\nu)_{\rm SM}} = 1.62 \pm 0.54 . $
(13) This causes the exclusion of two sectors of
$(\tan\beta)/m_{H^\pm}$ ratio in 2HDM [14]. This implies that for$\tan\beta \geq 1$ , the mass of charged Higgs must be greater than 800 GeV for Type-II 2HDM[15]. Given that the ($ B_\mu \rightarrow {\tau}{\nu}_{\tau} $ ) decay depends on$ \mid V_{ub} \mid $ , the ($ B_\mu \rightarrow Dl \nu $ ) (semi-leptonic) decay also depends on$ \mid V_{ub} \mid $ , which is more precisely known than$ \mid V_{ub} \mid $ . Moreover, the branching ratio of ($ B_\mu \rightarrow {\tau}{\nu}_{\tau} $ ) is fifty times greater than the branching ratio of ($ B_\mu \rightarrow {\tau}{\nu} $ ) in the standard model, but it is still difficult to detect because two neutrinos exist in its final state. The 2HDM deals only with the numerator of the ratio$ {\xi}_{Dl\nu_\tau} = \frac{{\rm BR}(B\rightarrow D{\tau}{\nu}_{\tau})}{{\rm BR}(B\rightarrow Dl{\nu}_{\tau})} $
(14) and allows reducing theoretical uncertainties to some extent. The experimental outcomes by BaBar collaborations and SM predictions [14] are
$ {\xi}^{\rm SM}_{Dl{\nu}_{\tau}} = (29.7 \pm 3) \times 10^{-2} ,$
(15) $ {\xi}^{\rm exp}_{Dl{\nu}_{\tau}} = (44.0 \pm 5.8 \pm 4.2) \times 10^{-2} . $
(16) For (
$ B \rightarrow X_s\gamma $ ), this special transition is mediated by$ H^{\pm} $ and includes Flavor-changing neutral current (FCNC) and$ W^{\pm} $ contributions. Regarding the respective BR, the contribution of charged Higgs is always positive when it comes to probing Type-II 2HDM; this can be used efficiently. For this transition, the NNLO-SM predicted$(3.34\pm0.22)\times 10^{-4}$ for${\rm BR} (B \rightarrow X_s\gamma)_{\rm SM}$ [16]. Therefore, for${\rm BR} (B \rightarrow X_s\gamma)_{\rm exp}$ , the recently experimentally calculated value is$ (3.32\pm0.15)\times 10^{-4} $ . For Type-II Yukawa interactions, this constraint excludes the light-charged Higgs. Higher order analysis [17] estimated a lower limit of 380 GeV at 95% C.L. for$ M_{H^{\pm}} $ . However, it is important to mention that the bound in [17] does not include novel experimental and theoretical predictions and hence the numerical results may be outdated.For (
$ D_S \rightarrow \tau\nu $ ), the SM prediction is$ (3.32\pm0.15)\times 10^{-4} $ [12] for$f_{Ds} = 0.248 \pm 2.5~{\rm GeV}$ [18] and the updated experimental calculation for${\rm BR}(D_s \rightarrow \tau\nu)_{\rm exp}$ is$ (5.51\pm 0.24)\times10^{-2} $ [19] for ($ B_{d/s} \rightarrow \mu^+\mu^- $ ). At large$\tan \beta$ values, the lower limit on charged Higgs mass$ m_{H^{\pm}} $ is given in [20]. For decays${\rm BR}(B_s \rightarrow {\mu}^+{\mu}^+)_{\rm SM}$ and${\rm BR} (B_d \rightarrow {\mu}^+{\mu}^+)_{\rm SM}$ , SM predictions are$ (3.54\pm0.27)\times10^{-9} $ and$(1.1\pm0.1)\times 10^{-9}$ , respectively [12]. Experimental results for the limits of these decays at 95% C.L. are${\rm BR} (B_s \rightarrow {\mu}^+{\mu}^+)_{\rm exp} < 4.5 \times 10^{-9}$ and${\rm BR} (B_d \rightarrow {\mu}^+{\mu}^+)_{\rm exp} < 1.0 \times 10^{-9}$ reported by the LHCb collaboration [21]. If the results from ATLAS and CMS [22] concerning the aforementioned limits are also added, then stricter limits are${\rm BR} (B_s \rightarrow {\mu}^+{\mu}^+)_{\rm exp} < 4.2 \times 10^{-9}$ and${\rm BR}(B_d \rightarrow {\mu}^+{\mu}^+)_{\rm exp} < 8.1 \times 10^{-10}$ . -
Multivariate Analysis (MVA) [25] methods have been widely used in many studies conducted in ATLAS and CMS to discriminate signals over backgrounds in searches that include complex multi-particle final states, e.g., the current analysis of
$pp \rightarrow H^{+} t \rightarrow t\bar{t}b \rightarrow 3bjets + 4 jets$ . Based on machine learning, the multivariate classification technique is fundamental for many data analysis methods. It is included in the ROOT framework; the TMVA toolkit encompasses a large diversity of classification algorithms. Training, testing, progress assessment, and use of all accessible classifiers are executed simultaneously and are simple to use. Supervised Machine Learning is used in all TMVA methods, which exploit training events to determine the required outputs. When implementing a machine learning algorithm for an analysis, a training process is required in which the algorithm observes already known events (i.e., simulated samples where right or wrong is predefined) and learns the difference between background and signal events.We selected three different techniques for charged Higgs production in association with top quark: Boosted Decision Trees (BDT), maximum Likelihood (LH) method, and Multilayer Perceptron (MLP). The production samples are based on Monte Carlo (MC) simulations of the signal and most backgrounds at
$ \sqrt{s} = 100 $ TeV. The main background consists of production of single top quarks in association with W boson, single top production in s-channel process, and top anti-top quark pair production. The charged Higgs mass was set at 800 GeV, while the signal contained a physical single charged Higgs boson associated with top quark. Both signal and background processes were produced with Pythia 8 embedded in MadGraph [26] and the detector simulation was performed by Delphes [27]. Both the signal and background samples were analyzed with the Toolkit for MultiVariate Analysis (TMVA) available in ROOT, which includes a large variety of multivariate classification algorithms. Events were selected according to the presence of a required number of objects in the final state. In signal events including at least 7 jets, there must be at least three b-jets and 4 light jets. In addition, a missing transverse energy larger than 30 GeV, a transverse momentum greater than 20 GeV, and an absolute pseudo rapidity less than 3.0 were applied as selection cuts for background suppression and signal enhancement.The responses of the three classifiers are shown in Fig. 11, and their values for Area Under the Curve (AUC) are extracted in the form of Table 1. These results are satisfactory except for the Likelihood method.
Figure 11. (color online) The y-axis represents the background rejection, i.e., how much background is lost after a given cut on the BDT, MLP, and Likelihood outputs, while the x-axis represents how much signal is kept. This gives an idea on how a set of chosen variables is discriminated.
MVA Classifier AUC(with cuts) MLP 0.777 BDT 0.781 Likelihood 0.582 Table 1. MVA Classifier Area Under the Curve (AUC) with cuts; the values are satisfactory, except for the Likelihood method.
We also calculated the statistical significance (SS) from signal (S) and background (B) events, where significance is defined as
$ SS= S /\sqrt{S+B} $ . The results are presented in Table 2. This enables the observability of the charged Higgs at FCC designed integrated luminosity.MVA Classifier Optimal-Cut $\dfrac{S}{\sqrt{S+B}}$ Sig-Eff Bkg-Eff MLP 0.2046 105.575 0.924 0.6079 Likelihood −3.9495 100.004 1 0.9998 BDT −0.0944 105.89 0.9306 0.6141 Table 2. Optimal-Cuts, Signal-Background ratio, signal, and Background efficiency for a number of signals and backgrounds under the application of cuts.
For a given number of signal and background events, Fig. 12 depicts the significance
$ S/\sqrt{S + B} $ as a function of the BDT cut value in (a) and Likelihood cut in (b). This is the parameter that most properly indicates the training performance. The significance output is used to evaluate the corresponding method performance: a higher significance indicates a better classifier. However, statistics must be considered, which means that the signal efficiency that corresponds to the ideal cut value must be considered. A lot of signal are filtered out if this number is really low. If the real data exhibit low statistics, the quality of the cut cannot be guaranteed because the output distribution differs from that of the training sample. Figure 13 shows the MLP cut efficiencies as a function of the applied cut value. MLP and BDT present similar values of significance. These values are higher than that of the Likelihood method.
Observability of parameter space for charged Higgs boson in its bosonic decays in two Higgs doublet model Type-1
- Received Date: 2023-10-28
- Available Online: 2024-02-15
Abstract: This study explores the possibility of discovering