Stability of the Standard Model vacuum with respect to vacuum tunneling to the Komatsu vacuum in the cMSSM

Abstract

We investigate the stability of the Standard Model vacuum with respect to vacuum tunneling to the Komatsu vacuum, which exists when $m_{L}^{2}+m_{H_{u}}^{2}<0$, in the cMSSM. Employing the numerical tools SARAH, SPheno and CosmoTransitions, we scan and constrain the parameter space of the cMSSM up to 10 TeV. Regions excluded due to having a vacuum tunneling half-life less than the age of the observable universe are concentrated near the regions where the Standard Model vacuum is tachyonic and are more stringent at smaller $m_{0}$, larger and negative $A_{0}$, and larger $\tan\beta$. New excluded regions, which satisfy $m_{h}\simeq 125\text{GeV}$, are found.

I Introduction

The Minimal Supersymmetric Standard Model (MSSM) Martin (1998); Aitchison (2005); Csaki (1996) is one of the most compelling models for physics beyond the Standard Model, as its minimal combination of the Standard Model and supersymmetry not only provides a solution to the hierarchy problem but also predictions that have been confirmed by experiments. First, assuming gauge coupling unification and a supersymmetry mass scale in the range of $10^{2}$ to $10^{4}$ GeV, the MSSM correctly Amaldi et al. (1991); Martens et al. (2010) predicted Einhorn and Jones (1982); Dimopoulos and Georgi (1981) a relation between the Standard Model gauge couplings. Second, the MSSM has consistently predicted the Higgs mass $m_{h}\lesssim 130$ GeV since well before the LHC Hempfling and Hoang (1994); Degrassi et al. (2001); Brignole et al. (2002). Combined with the experimental bound of $m_{h}\gtrsim$ 114 GeV from the Large Electron–Positron Collider (LEP) Barate et al. (2003), the MSSM thus predicted the Higgs boson mass within a $10\%$ range of the value observed at the LHC Chatrchyan et al. (2012); Workman et al. (2022).

In the MSSM, the superpartners of the Standard Model particles were expected to be found in the range of $10^{2}$ to $10^{4}$ GeV and have been searched for in experiments such as the Large Hadron Collider (LHC). The observation of the Higgs boson mass at $125$ GeV and recent searches have constrained the stop and gluino masses to be above 1 TeV Draper et al. (2012); Heinemeyer et al. (2012); Brummer et al. (2012); Slavich et al. (2021); Aaboud et al. (2018); Sirunyan et al. (2019). More generally, the LHC has increased the lower bound of the supersymmetric particle masses Aaboud et al. (2018); Sirunyan et al. (2019); Han et al. (2017); Adam and Vivarelli (2022); Aad et al. (2014); Khachatryan et al. (2016), and the difference between the electroweak scale and the supersymmetric particle mass scale has become a fine-tuning problem known as the little hierarchy problem.

The Muon g-2 experiment conducted by Brookhaven National Laboratory Bennett et al. (2006) and Fermilab Abi et al. (2021); Aguillard et al. (2023) has measured the anomalous magnetic moment of the muon to 20 significant figures. Theoretical estimates of muon g-2 in the Standard Model derived using electron collision data Aoyama et al. (2020) show a $5\sigma$ significance difference with the experiment, but lattice QCD simulation of the Standard Model Borsanyi et al. (2021) agrees with the experimental results. Which calculation is correct will determine whether the g-2 experiment is evidence of physics beyond the Standard Model.

Despite a large region of parameter space being ruled out by experiments, the MSSM is still a compelling candidate for physics beyond the Standard Model, especially compared to its alternatives, such as extra dimensions Arkani-Hamed et al. (1998); Randall and Sundrum (1999), composite Higgs Miransky et al. (1989); Chivukula et al. (1999); Hill and Simmons (2003) and cosmological relaxation Graham et al. (2015), which lack successful predictions. It is expected to see clear evidence for the MSSM in a future collider that can detect sparticles with masses up to 10 TeV Abada et al. (2019a, b, c).

The MSSM has a high dimensional scalar field space, whose potential may have multiple vacua Drees et al. (1985); Kusenko et al. (1996); Komatsu (1988); Casas et al. (1996); Strumia (1996), which should satisfy the cosmological requirement that the tunneling half-life from the Standard Model vacuum to the other vacua should not be much less than the age of the observable universe.

Calculating tunneling in one-dimensional field space is straightforward, but finding the instanton path in multi-dimensional field space is difficult. However, various numerical packages have been developed over the last decade Wainwright (2012); Sato (2021); Guada et al. (2020) and used to constrain MSSM parameters Camargo-Molina et al. (2013, 2014); Bechtle et al. (2016); Chattopadhyay and Dey (2014); Hollik (2016); Duan et al. (2019); Hollik et al. (2019a, b). In Ref. Camargo-Molina et al. (2013), the authors scan and constrain the cMSSM parameters up to $4\mathinner{\mathrm{TeV}}$ for vacuum tunneling to vacua formed by $h_{u},h_{d},u_{3},\bar{u}_{3},e_{3},\bar{e}_{3}$. In Ref. Camargo-Molina et al. (2014), the parameter space of the Natural MSSM is constrained using vacuum and thermal tunneling with non-zero $h_{u},h_{d},u_{3},\bar{u}_{3}$. No points satisfying the Higgs mass constraint $m_{h}\simeq 125\mathinner{\mathrm{GeV}}$ were excluded by vacuum tunneling, but some were by thermal tunneling. In Ref. Bechtle et al. (2016), the authors assume neutralino dark matter with stau-coannihilation and the theoretical estimate of muon g-2 and find points which best fit with dark matter abundance, muon g-2, and the Higgs mass and decay rate. These points are tested by vacuum tunneling to the Komatsu vacuum and they are found to be safe. Several other papers Chattopadhyay and Dey (2014); Hollik (2016); Duan et al. (2019); Hollik et al. (2019a, b) use vacuum tunneling to further reduce the parameter space that survives after making various assumptions and applying constraints. In this paper, we scan and constrain the full cMSSM parameter space up to $10\mathinner{\mathrm{TeV}}$ by vacuum tunneling to the Komatsu vacuum.

The Komatsu vacuum Komatsu (1988); Abel and Savoy (1998) lies in the direction

$$\mu H_{u}L_{k}=\lambda_{d}^{ij}\bar{d}_{i}Q_{j}L_{k}+\lambda_{e}^{ij}\bar{e}_{i}L_{j}L_{k}$$ (1)

along which the $\mu$ term contribution to $LH_{u}$’s mass squared is cancelled. At large values of $LH_{u}$, the D-term constrains $|L|\simeq|H_{u}|$ and if

$$m_{L}^{2}+m_{H_{u}}^{2}<0$$ (2)

the potential descends into a deep vacuum, where it is stabilized by higher-order terms such as supersymmetric neutrino mass terms.

In Section II, we present the potential used in this paper and briefly review the calculation of vacuum tunneling. In Section III, we illustrate the tunneling calculation process, including a brief explanation of the numerical tools: SARAH Staub (2014), SPheno Porod and Staub (2012) and CosmoTransitions Wainwright (2012). In Section IV, we plot the region excluded by vacuum tunneling to the Komatsu vacuum in cMSSM parameter space. We summarize the results and suggest future work in Section V.

II Model

The MSSM superpotential is

$$W=\lambda_{u}^{ij}\bar{u}_{i}Q_{j}H_{u}-\lambda_{d}^{ij}\bar{d}_{i}Q_{j}H_{d}-\lambda_{e}^{ij}\bar{e}_{i}L_{j}H_{d}+\mu H_{u}H_{d}$$ (3)

There are two types of Komatsu vacuum: quark Komatsu vacuum $\mu H_{u}L_{k}=\lambda_{d}^{ij}\bar{d}_{i}Q_{j}L_{k}$ and lepton Komatsu vacuum $\mu H_{u}L_{k}=\lambda_{e}^{ij}\bar{e}_{i}L_{j}L_{k}$. We have sampled the tunneling rate for the quark Komatsu vacuum and found it to be consistently less than that for the lepton Komatsu vacuum. Thus, we restrict to the lepton Komatsu vacuum

$$\mu H_{u}L_{k}=\lambda_{e}^{ij}\bar{e}_{i}L_{j}L_{k}$$ (4)

Note that $L_{j}L_{k}=0$ if $j=k$, so we need to consider at least two lepton generations. We neglect the off-diagonal Yukawa couplings and reduce the field space by the following three criteria. First, we build the potential relevant to the Standard Model and lepton Komatsu vacua by using $H_{u},H_{d},L$ and $\bar{e}$, and setting $Q=\bar{u}=\bar{d}=0$. Second, we choose the combination of lepton generations which gives the largest tunneling rate. When the Yukawa coupling $\lambda_{e}^{ij}$ is large, $\bar{e}_{i}L_{j}$ can be small but still cancel the $\mu$ contribution to $LH_{u}$’s mass squared, resulting in a saddle point close to the origin and a large tunneling rate. Hence, we use $\lambda_{e}^{33}\bar{e}_{3}L_{3}$ and set $\bar{e}_{1}=\bar{e}_{2}=0$. On the other hand, a larger Yukawa coupling renormalises the corresponding slepton mass squared to smaller values at low energy, increasing the tunneling rate. Hence we set $L_{1}=0$ and use $L_{2}$. Third, we use $\mathinner{SU\mathopen{\left(2\right)}}$ gauge freedom to set $h_{u}^{+}=0$. Then $e_{2}=\nu_{3}=h_{d}^{-}=0$ follows because the fields have no destabilizing linear terms. Thus, the field configuration we use in this paper is

$$\begin{aligned} H_{u}=\begin{pmatrix}0\\ h_{u}\end{pmatrix}\begin{matrix},\ \end{matrix}H_{d}=\begin{pmatrix}h_{d}\\ 0\end{pmatrix}\begin{matrix},\ \end{matrix}L_{2}=\begin{pmatrix}\nu_{2}\\ 0\end{pmatrix}\begin{matrix},\ \end{matrix}L_{3}=\begin{pmatrix}0\\ e_{3}\end{pmatrix}\begin{matrix},\ \end{matrix}\bar{e}_{3}=\bar{e}_{3}\\ \end{aligned}$$ (5)

and other fields zero.

At tree-level our potential is

$$V_{\mathrm{tree}} =m_{H_{u}}^{2}{|h_{u}|}^{2}+m_{H_{d}}^{2}{|h_{d}|}^{2}+m_{L_{2}}^{2}{|\nu_{2}|}^{2}+m_{L_{3}}^{2}{|e_{3}|}^{2}+m_{\bar{e}_{3}}^{2}{|\bar{e}_{3}|}^{2}$$ (6)

$$+{[}(a_{e}^{33}\bar{e}_{3}e_{3}h_{d}-B{\mu}h_{u}h_{d})+\mathrm{c.c.}]+{|\lambda_{e}^{33}\bar{e}_{3}h_{d}|}^{2}+{|{\mu}h_{u}-\lambda_{e}^{33}\bar{e}_{3}e_{3}|}^{2}$$

$$+\dfrac{1}{8}g_{1}^{2}{({|h_{u}|}^{2}-{|h_{d}|}^{2}-{|\nu_{2}|}^{2}-{|e_{3}|}^{2}+2{|\bar{e}_{3}|}^{2})}^{2}+\dfrac{1}{8}g_{2}^{2}{({|h_{u}|}^{2}-{|h_{d}|}^{2}-{|\nu_{2}|}^{2}+{|e_{3}|}^{2})}^{2}$$

The tunneling rate to the Komatsu vacuum is calculated by the well-known instanton method Coleman (1977), which gives a tunneling rate

$$\Gamma=Ae^{-S_{E}}$$ (7)

where the prefactor A comes from the measure of the path integral and includes fluctuations around the instanton, and $S_{E}$ is the Euclidean action given by

$$S_{E}=2\pi^{2}\int_{0}^{\infty}\rho^{3}d\rho\left[\dfrac{1}{2}h_{ab}\dfrac{\partial\phi^{a}}{d\rho}\dfrac{\partial\phi^{b}}{d\rho}+{\mathinner{V\mathopen{\left(\phi\right)}}}\right]$$ (8)

where $\rho$ is the radial distance in Euclidean space. The field space vector

$$\phi=(h_{u},h_{d},\nu_{2},e_{3},\bar{e}_{3})$$ (9)

and $h_{ab}$ is the metric on field space. The equation of motion for the instanton tunneling from false vacuum $\phi_{\mathrm{f}}$ to true vacuum $\phi_{\mathrm{t}}$ is

$$\dfrac{d^{2}\phi^{a}}{d\rho^{2}}+\dfrac{3}{\rho}\dfrac{d\phi^{a}}{d\rho}=h^{ab}\dfrac{\partial V}{\partial\phi^{b}}$$ (10)

with the boundary conditions $\dfrac{d\phi}{d\rho}(0)=0$, $\phi(0)=\phi_{\mathrm{t}}$ and $\phi(\infty)=\phi_{\mathrm{f}}$.

Figure 1: Full process of tunneling calculation

Figure 2: Detailed process of searching for the first estimate of the escape point

III Method

III.1 Numerical tools

We use SARAH Staub (2014) to generate the one-loop corrected potential and source file for SPheno Porod and Staub (2012). We calculate the supersymmetric particle spectrum with SPheno, including masses and Yukawa couplings at a two-loop level. We use CosmoTransitions Wainwright (2012) to calculate the multi-field tunneling. It decomposes the instanton equation into parallel and perpendicular to the path and searches for the instanton path, which is the solution for both parallel and perpendicular equations.

III.2 Tunneling to Komatsu vacuum

The calculation process is illustrated in Figure 1. We work in the context of the cMSSM Kane et al. (1994), which simplifies MSSM parameters into five parameters: the universal scalar mass $m_{0}$, the ratio between the MSSM Higgs vacuum expectation values $\tan\beta$, the universal gaugino mass $m_{1/2}$, and the universal trilinear coupling $A_{0}$, as well as the sign of $\mu$. The scanning range for the cMSSM parameters is shown in TABLE 1.

Parameter	Range
$m_{0}/\mathinner{\mathrm{TeV}}\;$	$\{0.1,0.2,0.5,1,2,5,10\}$
$\tan\beta$	$\{10,20,30,40,50\}$
$\log_{10}(m_{1/2}/\mathinner{\mathrm{TeV}})$	$[-1,1]$
$\log_{10}(|A_{0}|/\mathinner{\mathrm{TeV}})$	$[-1,1]$
sign of $\mu$	$+$

Table 1: Scanning range in cMSSM parameter space

We start by setting the scanning range of the cMSSM parameters. SPheno checks whether the Standard Model vacuum is correctly broken and generates a SUSY spectrum. Next, the existence of the Komatsu vacuum is checked by the sign of $m_{L}^{2}+m_{H_{u}}^{2}$. If the Komatsu vacuum exists, we move to find the first estimate of the escape point.

We look for the escape point of the bounce solution for potentials unbounded from below in the following manner. First, we find the saddle point and set $\phi_{*}=5\times\max\left\{h_{u},h_{d},\nu_{2},e_{3},\bar{e}_{3}\right\}_{\rm saddle}$ to choose the initial scanning range for finding the escape point. We find the minimum of the potential within the range $\phi^{a}<\phi_{*}$ for each $a$ by using the minimize function in Scipy. If the Standard Model vacuum is found as the minimum, we continuously extend the scanning range to $\phi^{a}<n\phi_{*}$ for some positive integer $n$ until we find a proper minimum, which we take as a first estimate of the escape point. The process of finding the first estimate of the escape point is shown in Figure 2. Finally, we calculate the bounce action with CosmoTransitions and save the results, including the location of the Standard Model vacuum, the escape point and the value of the tunneling action.

IV Results

The constraints on the cMSSM parameter space derived by considering vacuum tunneling from the Standard Model vacuum to the lepton Komatsu vacuum are plotted in Figure 3. We set the prefactor A in Eq. (7) to $\mathinner{\mathrm{TeV}}^{4}$, and the action threshold for determining dangerous tunneling to 410, which corresponds to a tunneling half-life of the order of the age of the observable universe. We set the spread of the Higgs mass constraint as $1\mathinner{\mathrm{GeV}}$, which is the theoretical uncertainty of the Higgs mass calculation in $\verb|SPheno|$ Bahl et al. (2020); Staub and Porod (2017).

Figure 3: Constraints from vacuum tunneling to the lepton Komatsu vacuum in cMSSM parameter space. Left to right: $\tan\beta=10,20,30,40,50$; bottom to top: $m_{0}=0.1,0.2,0.5,1,2,5,10$ TeV.    tachyonic Standard Model vacuum;  ,    $S<410$;    ,   $S>410$;    ,   no Komatsu vacuum;   no electroweak symmetry breaking;  ,  ,   $m_{h}=125.25\pm 1\mathinner{\mathrm{GeV}}$;  ,  ,   $m_{h}\neq 125.25\pm 1\mathinner{\mathrm{GeV}}$.

Figure 4: Enlargement of the region satisfying the Higgs mass constraint but excluded by tunneling. Left to right: $\tan\beta=40,50$; Bottom to top: $m_{0}=0.1,0.2,0.5,1,2$ TeV. Color representation is the same as in Figure 3.

There are four notable trends in the tunneling constraints. First, regions excluded by tunneling (yellow) are located near the regions where the Standard Model is tachyonic (red). The sparticles are close to tachyonic near the red regions, so the height of the saddle is lower, reducing the action. Second, the constraints are stronger for smaller m0 because the mass squareds are smaller, lowering the height of the saddle. Third, the constraints are stronger for larger $\tan\beta$ since $\lambda_{e}$ is larger allowing $\bar{e}_{i}L_{j}$ to be smaller, but still cancel the $\mu$ contribution to mass squared of $LH_{u}$, resulting in a saddle point located nearer the origin and a smaller action. Fourth, the constraints are stronger for large and negative $A_{0}$. For negative $A_{0}$, the trilinear couplings are larger in magnitude at low energy, which renormalises the mass squareds to be smaller at low energy, lowering the height of the saddle.

Our main findings are as follows. For $\tan\beta\leq 30$, only a few points near the tachyonic region are excluded by tunneling and they don’t satisfy the Higgs mass constraint. However, for $\tan\beta=40,50$, we find a region, enlarged in Figure 4, satisfying the Higgs mass constraint but excluded by tunneling, which has not been reported previously.

We have also scanned using the one-loop corrected potential but the tunneling results are not significantly different compared to those of tree-level potential tunneling. We did not consider the full one-loop corrections to the tunneling Guada and Nemevšek (2020).

V Summary

In this paper, we have constrained the cMSSM parameter space up to $10\mathinner{\mathrm{TeV}}$ by requiring the vacuum tunneling half-life to the lepton Komatsu vacuum to be greater than the age of the observable universe. The results in Figure 3 show that the tunneling constraints are significant only at $\tan\beta=40,50$ and $m_{0}<2\mathinner{\mathrm{TeV}}$, and the excluded regions are situated near where the Standard Model is tachyonic. Figure 4 enlarges the region satisfying the Higgs mass constraint but excluded by tunneling. The constraints from vacuum tunneling are not strong, but they are model-independent and robust. We expect to see stronger constraints from thermal tunneling, but they will be less robust as they will depend on the cosmological history.