Gravitationally dominated instantons and instability of dS, AdS and Minkowski spaces

Abstract

We study the decay of the false vacuum in the regime where the quantum field theory analysis is not valid, since gravitational effects become important. This happens when the height of the barrier separating the false and the true vacuum is large, and it has implications for the instability of de Sitter, Minkowski and anti-de Sitter vacua. We carry out the calculations for a scalar field with a potential coupled to gravity, and work within the thin-wall approximation, where the bubble wall is thin compared to the size of the bubble. We show that the false de Sitter vacuum is unstable, independently of the height of the potential and the relative depth of the true vacuum compared to the false vacuum. The false Minkowski and anti-de Sitter vacua can be stable despite the existence of a lower energy true vacuum. However, when the relative depth of the true and false vacua exceeds a critical value, which depends on the potential of the false vacuum and the height of the barrier, then the false Minkowski and anti-de Sitter vacua become unstable. We calculate the probability for the decay of the false de Sitter, Minkowski and anti-de Sitter vacua, as a function of the parameters characterizing the field potential.

1 Introduction

The decay of the false vacuum in quantum field theory (QFT), i.e. the transition from a metastable state to a more stable state through quantum tunneling, is a non-perturbative process that has significant implications for cosmology and particle physics. The false vacuum is a local minimum of the potential energy of a quantum field, but it is not the lowest energy state, which is the true vacuum. The process of the decay of the false vacuum involves the formation of bubbles in which small regions of the true vacuum spontaneously form within the false vacuum due to quantum fluctuations. Once a bubble of the true vacuum has formed, it can expand if the energy density inside the bubble is less than the energy density of the surrounding false vacuum and the bubble wall separates the two vacua.

The bubble grows when the energy gained from the false vacuum outweighs the energy cost of the bubble wall. As bubbles of true vacuum grow and merge, they percolate the entire space, leading to a transition from the false vacuum to the true vacuum. The decay of the false vacuum can be analyzed using the semiclassical framework Kobzarev:1974cp ; PhysRevD.15.2929 . In this framework, the decay rate per unit volume and unit time is $\Gamma\sim Ae^{-S}$, where $A$ is a dimensional coefficient that is inversely proportional to the fourth power of the bubble radius, and $S$ is the Euclidean action of the bounce solution.

The aim of this work is to analyse the decay of the false vacuum when gravity becomes important, and the QFT analysis is not valid. This happens when the height of the barrier separating the false and the true vacua is large, and it has implications for the instability of de Sitter (dS), Minkowski and anti-de Sitter (AdS) vacua. We will work within the thin-wall approximation Kobzarev:1974cp ; PhysRevD.15.2929 , where the bubble wall is thin compared to the size of the bubble, and show that: (i) the false dS vacuum is always unstable, regardless of the height of the potential and the relative depth of the true vacuum compared to the false vacuum, (ii) the false Minkowski vacuum and the AdS vacuum can be stable despite the existence of a true vacuum with lower energy. However, if the relative depth of the vacua exceeds a critical value, which depends on the value of the potential in the false vacuum and the height of the barrier, then the false vacuum in these two cases becomes unstable and decays into a true vacuum. We calculate the probability for the decay of the false dS, Minkowski and AdS vacua as a function of the parameters characterizing the field potential. Our analysis is performed in the framework of relativistic field theory coupled to gravity.

The paper is organized as follows. In section II, we will consider the structure and range of validity of gravitationally dominated instantons in the thin-wall approximation. In section III we will analyse the instabilities of dS, Minkowski and AdS vacua in the regime, where gravity plays an important role. Section IV will be devoted to a discussion.

2 Gravitationally dominated instantons

We will consider a scalar field with a potential shown in Fig.1, which is coupled to gravity.

Figure 1: A scalar field potential $V(\varphi)$ that has a local minimum false vacuum at $\varphi=0$, where $V\left(\varphi=0\right)=V_{+}$, which is separated by a barrier of height $V_{bar}$ from the true vacuum at $\varphi=\varphi_{0}>0$, with $V\left(\varphi_{0}\right)=V_{-}$.

2.1 The Newtonian approximation

To gain insight into the importance of gravity for the decay of the false vacuum, we first perform the field theory analysis in Minkowski space, taking into account the leading gravitational corrections in the Newtonian approximation. Let us assume that the scalar field potential $V(\varphi)$ has a local minimum corresponding to false vacuum at $\varphi=0$, where $V\left(\varphi=0\right)=V_{+}=\varepsilon$, which is separated from the true Minkowski vacuum at $\varphi=\varphi_{0}>0$, with $V\left(\varphi_{0}\right)=V_{-}=0$, by a barrier of height $V_{bar}$. The false vacuum is unstable and will decay by the formation of critical bubbles, whose size $\varrho_{b}$ can be estimated using the energy conservation. Working without gravity, and equating the total energy released during the transition, $4\pi\varepsilon\varrho_{b}^{3}/3$, with the surface energy of the bubble wall, $4\pi\sigma\varrho_{b}^{2}$, where $\sigma$ is the surface tension of the bubble, we obtain Kobzarev:1974cp ; PhysRevD.15.2929 :

$$4\pi\varepsilon\varrho_{b}^{3}/3=4\pi\sigma\varrho_{b}^{2}\implies\varrho_{b}=\frac{3\sigma}{\varepsilon}\ .$$ (1)

The gravitational corrections to (1) are parametrized by an effective dimensionless gravitational coupling $g_{eff}=\kappa M/\varrho$, where $\kappa\equiv 8\pi G$ with $G$ being the gravitational Newton constant, $\varrho$ and $M$ are the characteristic length and mass in the problem. In our case, $\varrho$ is the bubble size $\varrho_{b}$, and $M$ is the total energy released during the transition and the surface energy of the bubble wall. The first order gravitational corrections to the energy balance equation (1) come from the negative gravitational energy contributions and in the Newtonian approximation the balance equation reads:

$$\frac{4\pi}{3}\varepsilon\varrho_{b}^{3}-\frac{3}{5}\frac{\kappa(4\pi\varepsilon\varrho_{b}^{3}/3)^{2}}{8\pi\varrho_{b}}=4\pi\sigma\varrho_{b}^{2}-\frac{\kappa\left(4\pi\sigma\varrho_{b}^{2}\right)^{2}}{16\pi\varrho_{b}}\ ,$$ (2)

where we have neglected higher order terms in $g_{eff}$. Equation (2) shows the potential importance of gravity for bubble formation, but its range of validity is limited to $g_{eff}\ll 1$. Looking at the LHS of (3), this means that equation (2) is valid provided:

$$\frac{\kappa\frac{4\pi}{3}\varepsilon\varrho_{b}^{3}}{\varrho_{b}}\ll 1\implies\varrho_{b}\ll\frac{1}{\sqrt{\kappa\varepsilon}}\ .$$ (3)

This is the regime in which $\varrho_{b}\ll R_{+}$, where $R_{+}=\sqrt{3/(\kappa\varepsilon)}$ is the dS radius corresponding to the false vacuum. If we take into account the condition $g_{eff}\ll 1$ in the RHS of (3) we get

$$\varrho_{b}\ll\frac{1}{\kappa\sigma}\ ,$$ (4)

which can be related in the thin-wall approximation to the parameters of the potential since $\sigma\sim\varphi_{0}V_{bar}^{1/2}$ (see, e.g. Mukhanov:2005sc ). As we will see, the interesting phenomena caused by gravity occur when $g_{eff}\sim O(1)$, i.e. $\varrho_{b}\sim R_{+}$ or $\varrho_{b}\sim 1/(\kappa\varphi_{0}V_{bar}^{1/2})$, where the analysis requires general relativity, and this will be carried out in the next section.

2.2 Instantons basics

As argued in PhysRevD.15.2929 , the most important contribution to the sub-barrier transition between the false and the true vacua is due to the $O(4)$ Euclidean instantons, for which the value of the scalar field $\varphi$ depends on the Euclidean time $\tau$ and the spatial radial coordinate $r$ in the combination $\tau^{2}+r^{2}$. When the gravitational effects can be neglected, the background metric for these instantons is the flat Euclidean metric. The gravitational backreaction of the instantons on the metric that preserves the $O(4)$ symmetry, results in a conformally flat metric:

$$ds^{2}=f^{2}\left(\sqrt{\tau^{2}+r^{2}}\right)\left(d\tau^{2}+dr^{2}+r^{2}d\Omega^{2}\right),$$ (5)

where $d\Omega^{2}=d\theta^{2}+\sin^{2}\theta d\varphi^{2}$.

Define:

$$\tau=\bar{\xi}\sin\chi,\quad r=\bar{\xi}\cos\chi\ ,$$

where $\infty>\bar{\xi}\geq 0$ and $\pi/2\geq\chi\geq-\pi/2$, and then the metric reads:

$$ds^{2}=f^{2}\left(\bar{\xi}\right)\left[d\bar{\xi}^{2}+\bar{\xi}^{2}\left(d\chi^{2}+\cos^{2}\chi d\Omega^{2}\right)\right]\ .$$ (6)

Introducing:

$$d\xi=f\left(\bar{\xi}\right)d\bar{\xi},\qquad\varrho=f\left(\bar{\xi}\right)\bar{\xi}\ ,$$ (7)

we recast (6) as:

$$ds^{2}=d\xi^{2}+\varrho^{2}\left(\xi\right)\left(d\chi^{2}+\cos^{2}\chi d\Omega^{2}\right)\ .$$ (8)

The off-shell Euclidean action for a scalar field $\varphi\left(\xi\right)$ with potential $V\left(\varphi\right)$ propagating in the background metric (8) reads:

$$S_{E}=2\pi^{2}\int d\xi\left[\varrho^{3}\left(\frac{1}{2}\dot{\varphi}^{2}+V\right)+\frac{3}{\kappa}\left(\varrho^{2}\ddot{\varrho}+\varrho\left(\dot{\varrho}^{2}-1\right)-(\varrho^{2}\dot{\varrho})^{\centerdot}\right)\right]\ ,$$ (9)

where the dot denotes a derivative with respect to $\xi$, and the last term is a total derivative, corresponding to York-Gibbons-Hawking boundary term PhysRevLett.28.1082 ; Gibbons:1976ue .

The variation of the action with respect to $\varphi$ yields its field equation:

$$\ddot{\varphi}+3\frac{\dot{\varrho}}{\varrho}\dot{\varphi}-V_{,\varphi}=0\ ,$$ (10)

where $V_{,\varphi}=dV/d\varphi$, while the variation with respect to $\varrho$ gives:

$$2\varrho\ddot{\varrho}+\dot{\varrho}^{2}-1=-\kappa\varrho^{2}\left(\frac{1}{2}\dot{\varphi}^{2}+V\right)\ .$$ (11)

Multiplying (11) by $\dot{\varrho}$ and using (10) to express $\varrho^{2}\dot{\varrho}\dot{\varphi}^{2}$ in the RHS in terms of $\varrho$ and the derivative of $\left(\dot{\varphi}^{2}/2-V\right)$ we get:

$$\left(\varrho\left(\dot{\varrho}^{2}-1\right)\right)^{\centerdot}=\frac{\kappa}{3}\left(\varrho^{3}\left(\frac{1}{2}\dot{\varphi}^{2}-V\right)\right)^{\centerdot}\ .$$ (12)

This can be integrated to obtain:

$$\dot{\varrho}^{2}=1+\frac{\kappa}{3}\varrho^{2}\left(\frac{1}{2}\dot{\varphi}^{2}-V\right)\ ,$$ (13)

where we set the integration constant to zero in accordance with the $0-0$ component of the Einstein equations. Inserting (13) in (11), we get:

$$\ddot{\varrho}=-\frac{1}{3}\kappa\varrho\left(\dot{\varphi}^{2}+V\right)\ .$$ (14)

To find the instanton action that dominates the tunneling rate, we need to solve the equations (10) and (13) with the appropriate boundary conditions. Using these equations, we can rewrite the on-shell action in a form that is more convenient for calculating the instanton action. Substituting (13) in (9), the on-shell action simplifies to:

$$S_{E}=2\pi^{2}\int d\xi\left[\varrho^{3}\dot{\varphi}^{2}+\frac{3}{\kappa}\left(\varrho^{2}\ddot{\varrho}-(\varrho^{2}\dot{\varrho})^{\centerdot}\right)\right]\ ,$$ (15)

which using (14) can be written as:

$$S_{E}=-2\pi^{2}\int d\xi\left(\varrho^{3}V+\frac{3}{\kappa}(\varrho^{2}\dot{\varrho})^{\centerdot}\right)\ .$$ (16)

The decay rate of the false vacuum per unit volume per unit time can be estimated as Coleman:1980aw :

$$\Gamma\sim\varrho_{b}^{-4}e^{-(S_{E}^{f}-S_{E}^{i})},$$ (17)

where $\varrho_{b}$ is the size of the resulting critical bubble. $S_{E}^{i}$ and $S_{E}^{f}$ are the corresponding Euclidean actions for the initial false vacuum and for the final configuration, where the true vacuum has been created inside the critical bubble.

Note, that so far all the equations were exact, and we have not made any approximation. In order to obtain the quantitative results for general potentials $V$, we will have to make approximations. In the next section we will consider the thin-wall approximation in the presence of gravity. We will also denote the Euclidean action $S_{E}$ without the subscript $E$.

2.3 The thin wall approximation

Let us consider a false vacuum with potential $V_{+}$, located at $\varphi=0$, and separated from the true vacuum of depth $V_{-}<V_{+}$ at $\varphi_{0}$ by the barrier with positive height $V_{bar}\gg\left|V_{+}\right|$. Depending on $V_{-}$, the false dS vacuum decays either into a true dS vacuum with a lower potential or into Minkowski or AdS vacuum. The decay occurs via instantons, which lead to the formation of critical bubbles of size $\varrho_{b}$. We assume that the thickness of the bubble wall $\delta\varrho_{b}$ is much smaller than $\varrho_{b}$. This corresponds to the case, where the second term (“friction”) in the equation (10) can be neglected in the leading order approximation.

To determine the range of validity of this approximation, we consider the first integral of the equation (10)

$$\frac{1}{2}\dot{\varphi}^{2}-V=-\int\frac{3\dot{\varrho}}{\varrho}\dot{\varphi}^{2}d\xi\ .$$ (18)

In the thin-wall approximation, the integral on the RHS of (18), to which only the bubble wall contributes where $\dot{\varphi}^{2}$ does not vanish, must be small compared to each individual term on the LHS of (18) within the wall. Let us first assume that this condition is fulfilled, calculate the radius of the critical bubble, and determine for which potentials the thin-wall approximation can be applied. If we neglect the integral on the RHS in (18), then we obtain:

$$E\equiv\frac{1}{2}\dot{\varphi}^{2}-V\approx-V_{+}\ .$$ (19)

Assuming that $\left|V_{+}\right|\ll V_{bar}$, we can substitute $V\approx\dot{\varphi}^{2}/2$ within the wall into equation (14) and integrate over the wall to obtain:

$$\dot{\varrho}_{+}-\dot{\varrho}_{-}=-\frac{1}{2}\kappa\sigma\varrho_{b}\ ,$$ (20)

where $\dot{\varrho}_{-}$ and $\dot{\varrho}_{+}$ refer to the true and false vacuum immediately before and after the thin wall, and $\sigma$ is the surface tension defined as:

$$\sigma\equiv\int\dot{\varphi}^{2}d\xi\ .$$ (21)

For $\left|V_{+}\right|\ll V_{bar}$, $\sigma$ can be estimated as:

$$\sigma=\int\dot{\varphi}d\varphi\simeq\int\sqrt{2V}d\varphi\sim\sqrt{V_{bar}}\varphi_{0}\ .$$ (22)

Using (13) we have:

$$\dot{\varrho}_{+}=\pm\sqrt{1-\frac{\kappa V_{+}}{3}\varrho_{b}^{2},}\qquad\dot{\varrho}_{-}=+\sqrt{1-\frac{\kappa V_{-}}{3}\varrho_{b}^{2}}\ .$$ (23)

Note, that the space inside the resulting bubble expands and $\dot{\varrho}_{-}$ is therefore always positive, while outside the bubble $\dot{\varrho}_{+}$ can be either positive or negative, depending on the height of the barrier $V_{bar}$ and the other parameters characterising the field potential.

Solving (20) gives an expression for the size of the critical bubble

$$\varrho_{b}=\left[\frac{R_{+}^{2}}{1+\left(\frac{\varepsilon}{3\sigma}\left(1-\alpha\right)R_{+}\right)^{2}}\right]^{1/2},$$ (24)

where $R_{+}^{2}=3/\kappa V_{+}$, $\varepsilon\equiv V_{+}-V_{-}>0$ is difference between the potential heights in the false and true vacuum, respectively, and $\alpha$ is a dimensionless parameter:

$$\alpha\equiv\frac{3\kappa\sigma^{2}}{4\varepsilon}\ .$$ (25)

Inserting (24) into (23) we get:

$$\dot{\varrho}_{+}=\pm\frac{\varepsilon}{3\sigma}\left|1-\alpha\right|\varrho_{b},\quad\dot{\varrho}_{-}=\frac{\varepsilon}{3\sigma}\left(1+\alpha\right)\varrho_{b}\ .$$ (26)

Note, that the above calculations apply to both positive and negative $V_{+}$ and $V_{-}$, and can be used for the instability analysis of both dS and AdS spaces. For the false dS vacuum, $V_{+}>0$ and $R_{+}$ is the dS radius. When $\alpha>1$, we have to put a negative sign in front of the square root in $\dot{\varrho}_{b^{+}}$ in (23) to fulfill (20) and therefore:

$$\dot{\varrho}_{+}=\,\text{sgn}(1-\alpha)\sqrt{1-\frac{\kappa V_{+}}{3}\varrho_{b}^{2},}\qquad\dot{\varrho}_{-}=+\sqrt{1-\frac{\kappa V_{-}}{3}\varrho_{b}^{2}}\ .$$ (27)

Next, let us consider to which potentials the thin-wall approximation can be applied. As we can see from (14), $\dot{\varrho}$ always decreases as we go through the bubble wall, and therefore the integral on the RHS in (18) is bounded by

$$\int\frac{3\dot{\varrho}}{\varrho}\dot{\varphi}^{2}d\xi<\int\frac{3\dot{\varrho}_{-}}{\varrho_{b}}\dot{\varphi}^{2}d\xi=\varepsilon+\frac{3\kappa\sigma^{2}}{4}\ ,$$ (28)

and it must be much smaller than $V_{bar}$. Using $\sigma\sim\sqrt{V_{bar}}\varphi_{0}$, we find that the conditions for applicability of the thin-wall approximation are met when:

$$\frac{\varepsilon}{V_{bar}}\ll 1,\qquad\kappa\varphi_{0}^{2}\ll 1\ .$$ (29)

For the instantons with $\alpha>1$, the expanding dS solution inside the critical bubble is always matched with a contracting dS branch, describing the false vacuum outside the bubble. In this case, the value of $\dot{\varrho}$ should vanish inside the bubble wall, and its sign must change from positive to a negative inside the wall. As can be seen from (13), this would be impossible if the energy $E$ (19) remains exactly the same. However, due to the friction term in (10), the energy $E$ changes slightly. As can be seen from (13), this slight change is sufficient to violate the energy conservation law (19) by the amount $\Delta E\sim 3/(\kappa\varrho_{b}^{2})$ required to change the sign of the expansion rate $\dot{\varrho}$ within the wall, which is much smaller than $V_{bar}$ for the instantons with $\alpha>1$, if $\kappa\varphi_{0}^{2}\ll 1.$ Further, equation (19) which was used in the derivation (20) for the thin-wall bubble, where $\kappa\varphi_{0}^{2}\ll 1$ applies, is therefore also fulfilled in the leading order.

3 Instability of dS, AdS and Minkowski spaces

In this section we will analyse the instabilities of dS, Minkowski and AdS spaces, as well as the decay of the false vacuum in the regime where gravity is important and the QFT analysis is not valid.

3.1 The decay probability of the dS vacuum

To determine the decay rate of the dS false vacuum, we need to calculate the change in Euclidean on-shell action during the sub-barrier transition. In a false vacuum, $\dot{\varphi}=0$, and the equation (13) simplifies to:

$$\dot{\varrho}^{2}=1-\frac{\varrho^{2}}{R_{+}^{2}}\ ,$$ (30)

where $R_{+}=\sqrt{3/(\kappa V_{+})}$, and is solved by:

$$\varrho\left(\xi\right)=R_{+}\sin\left(\frac{\xi}{R_{+}}\right)\ .$$ (31)

When $\xi$ changes from $0$ to $\pi R_{+}$, the radius of the Euclidean dS instanton first increases, reaches its maximal value $\varrho=R_{+}$ at $\xi=\frac{1}{2}\pi R_{+}$, and then shrinks to zero.

Let us first consider the case in which $\alpha<1$. According to (27), in the final configuration where the critical bubble is created, we must match the expanding solution describing a true vacuum with a potential $V_{-}$ inside the bubble with an expanding dS solution corresponding to the false vacuum outside the bubble. The radius of the critical bubble $\varrho_{b}$ is smaller than $R_{+}$, and the wall is located at $\xi_{b}<\frac{1}{2}\pi R_{+}.$ The contribution of the false vacuum in the final configuration to the total action can be calculated by subtracting from the initial false vacuum action $S^{i}$ the action for the small bubble with radius $\varrho_{b}$, with a cosmological constant corresponding to the false vacuum:

$$S_{+}^{f}=S^{i}+2\pi^{2}\intop_{0}^{\varrho_{b}}\frac{V_{+}\varrho^{3}}{\sqrt{1-\kappa V_{+}\varrho^{2}/3}}d\varrho+\frac{6\pi^{2}}{\kappa}\varrho_{b}^{2}\dot{\varrho}_{+}=S^{i}+\frac{12\pi^{2}}{\kappa^{2}V_{+}}\left(1-\dot{\varrho}_{+}^{3}\right)\ .$$ (32)

In the derivation, we used (16) for the action, and replaced the integration over $\xi$ by the integration over $\varrho$ using (13) with the positive sign for the square root.

The action for the true vacuum expanding solution inside the bubble is calculated in the same way:

$$S_{-}^{f}=-2\pi^{2}\intop_{0}^{\varrho_{b}}\frac{V_{-}\varrho^{3}}{\sqrt{1-\kappa V_{-}\varrho^{2}/3}}d\varrho-\frac{6\pi^{2}}{\kappa}\varrho_{b}^{2}\dot{\varrho}_{-}=-\frac{12\pi^{2}}{\kappa^{2}V_{-}}\left(1-\dot{\varrho}_{-}^{3}\right)\ .$$ (33)

Finally, the contribution of the thin wall is obtained by integrating (15) over the wall, and taking into account the definition of the tension in (21):

$$S_{w}^{f}=2\pi^{2}\sigma\varrho_{b}^{3}\ .$$ (34)

Combining (32), (33) and (34), we obtain the total change in action between the initial and final configurations, which includes the critical bubble of the true vacuum separated from the false vacuum by the thin wall:

$$S^{f}-S^{i}=2\pi^{2}\sigma\varrho_{b}^{3}+\frac{12\pi^{2}}{\kappa^{2}}\left[\frac{1-\dot{\varrho}_{+}^{3}}{V_{+}}-\frac{1-\dot{\varrho}_{-}^{3}}{V_{-}}\right]$$

$$=\frac{12\pi^{2}}{\kappa^{2}}\left[\frac{1-\dot{\varrho}_{+}}{V_{+}}-\frac{1-\dot{\varrho}_{-}}{V_{-}}\right]=\frac{2\pi^{2}\sigma\varrho_{b}^{3}}{\left(1+\dot{\varrho}_{+}\right)\left(1+\dot{\varrho}_{-}\right)}\ ,$$ (35)

where we used (20) to simplify the final result.

Performing a similar calculation for the case $\alpha>1$, and taking into account that in this case we have to match the expanding true vacuum inside the bubble with the contracting false dS vacuum and accordingly take a negative sign for $\dot{\varrho}_{+}$, we get exactly the same expression. Thus, for each $\alpha>0$, the decay rate of the false vacuum is determined by (17), where the size of the critical bubble $\varrho_{b}$ is given in (24), and the change in the action (35) is calculated with $\dot{\varrho}_{+}$ and $\dot{\varrho}_{-}$ from (27).

As follows from (24), the size of the critical bubble never exceeds the radius of the false dS vacuum, and $\varrho_{b}=R_{+}$ is reached for $\alpha=1$. Given $V_{+}$ and $\varphi_{0}$, we aim to find out how the decay rate depends on the height of the potential barrier $V_{bar}$ separating false and true vacuum, and on the difference between the depths of the vacua $\varepsilon=V_{+}-V_{-}>0$. It is convenient to illustrate the results using the plane $\varepsilon-V_{bar}$ in Fig.2, where region I refers to the field theory instantons, and regions II and III describe gravitationally dominated instantons, with a radius of the order of the dS radius $R_{+}$ (II), and with radius much smaller than $R_{+}$ (III), respectively.

Figure 2: A plot of the plane $\varepsilon-V_{bar}$ on a logarithmic scale with $\varepsilon>0$ and $V_{bar}>V_{+}$. Region I refers to the field theory instantons. Regions II and III describe gravitationally dominated instantons, with a radius of the order of the dS radius $R_{+}$ (II), and with a radius much smaller than $R_{+}$ (III).

3.1.1 Field theory instantons

Region I in Fig.2: If we assume that $\alpha\ll 1$ and

$$\frac{\varepsilon}{3\sigma}(1-\alpha)R_{+}\gg 1\ ,$$ (36)

or, equivalently,

$$\varepsilon\gg\frac{3}{4}\kappa\sigma^{2}\left(1+\frac{4}{\kappa\sigma R_{+}}\right)\ ,$$ (37)

then the radius of the critical bubble (24) reads:

$$\varrho_{b}\simeq\frac{3\sigma}{\varepsilon}\ .$$ (38)

Using $\sigma\simeq\varphi_{0}V_{bar}^{1/2}$ and $\varepsilon\ll V_{bar}$ (29), it follows that for a given height of the barrier $V_{bar}$, the condition (37) can be rewritten as:

$$V_{bar}\gg\varepsilon\gg\frac{3}{4}\kappa\varphi_{0}^{2}V_{bar}\left(1+\sqrt{\frac{V_{+}^{*}}{V_{bar}}}\right)\ ,$$ (39)

where

$$V_{+}^{*}\equiv\frac{16V_{+}}{3\kappa\varphi_{0}^{2}}\ .$$ (40)

When this condition is met, we can neglect gravity and field theory instantons provide a valid approximation to the decay of the false vacuum. The false dS vacuum decays via these instantons to another dS vacuum with a smaller potential or to Minkowski space if $\varepsilon\leq V_{+}$. For $V_{bar}\gg\varepsilon\geq V_{+}$, the dS vacuum decays to an AdS space, whose radius $R_{AdS}=\sqrt{3/(\kappa\left|V_{-}\right|)}$ is always larger than the radius of the bubble.

Finally, we can set $\dot{\varrho}_{b^{+}}=\dot{\varrho}_{b^{-}}\simeq 1$ in (35), and obtain the known result for the field theory instanton action

$$S^{f}-S^{i}=\frac{\pi^{2}\sigma\varrho_{b}^{3}}{2}=\frac{27\pi^{2}}{2}\frac{\sigma^{4}}{\varepsilon^{3}}\ .$$ (41)

3.1.2 Gravitationally dominated instantons

Region II in Fig.2: If

$$\frac{\varepsilon}{3\sigma}\left|1-\alpha\right|R_{+}\leq 1\ ,$$ (42)

then, as follows from (24) the radius of the critical bubble is of the order of the dS radius:

$$\varrho_{b}\simeq R_{+}\ .$$ (43)

In this case gravity plays a significant role. The value of $\alpha$ in (42) can be either less than or greater than one. Therefore, the inequality (42) can be rewritten as:

$$\frac{3}{4}\kappa\varphi_{0}^{2}V_{bar}\left(1+\sqrt{\frac{V_{+}^{*}}{V_{bar}}}\right)>\varepsilon>\frac{3}{4}\kappa\varphi_{0}^{2}V_{bar}\left(1-\sqrt{\frac{V_{+}^{*}}{V_{bar}}}\right)\ .$$ (44)

If $V_{bar}\ll V_{+}^{*}$, both $\left(1+\dot{\varrho}_{+}\right)$ and $\left(1+\dot{\varrho}_{-}\right)$ in (35) are of order one, and the action that determines the transition rate from the false dS vacuum to the true dS vacuum with $V_{+}>V_{-}>V_{+}\left(1-\sqrt{V_{bar}/V_{+}^{*}}\right)$ is given by:

$$S^{f}-S^{i}\simeq O\left(1\right)\pi^{2}\sigma R_{+}^{3}\simeq O\left(1\right)\pi^{2}\left(\frac{V_{bar}}{V_{+}^{*}}\right)^{1/2}\frac{1}{\kappa^{2}V_{+}}\ .$$ (45)

For $V_{bar}\gg V_{+}^{*}$, the dS vacuum can decay via instantons of radius $R_{+}$ only into an AdS vacuum if $\varepsilon\simeq 3\kappa\varphi_{0}^{2}V_{bar}/4$. In this case

$$1+\dot{\varrho}_{-}\simeq 1+\sqrt{\frac{\varepsilon}{V_{+}}}\simeq\sqrt{\frac{3\kappa\varphi_{0}^{2}V_{bar}}{4V_{+}}}=\frac{1}{2}\kappa\sigma R_{+}\$$ (46)

and as follows from (35) the action is:

$$S^{f}-S^{i}\simeq\frac{4\pi^{2}}{\kappa}R_{+}^{2}=\frac{12\pi^{2}}{\kappa^{2}V_{+}}\ .$$ (47)

Region III in Fig.2: Finally, let us consider the case when $\alpha>1$, and

$$\frac{\varepsilon}{3\sigma}(\alpha-1)R_{+}\gg 1\ ,$$ (48)

i.e.,

$$\frac{3}{4}\kappa\varphi_{0}^{2}V_{bar}\left(1-\sqrt{\frac{V_{+}^{*}}{V_{bar}}}\right)\gg\varepsilon>0\ .$$ (49)

This only applies if the barrier is sufficiently high, $V_{bar}>V_{+}^{*}$. It then follows from (24) that

$$\varrho_{b}\simeq\frac{4}{\kappa\sigma}\sim\frac{O(1)}{\kappa\varphi_{0}V_{bar}^{1/2}}\ .$$ (50)

For such bubbles the negative contribution of gravity almost completely compensates the energy released during the transition in the wall. Taking into account that $\varrho_{b}\ll R_{+}$ for $V_{bar}\gg V_{+}^{*}$, and that the sign of $\dot{\varrho}_{+}$ in (27) is negative, the action (35) reads:

$$S^{f}-S^{i}=2\pi^{2}\sigma\varrho_{b}R_{+}^{2}\simeq\frac{8\pi^{2}}{\kappa}R_{+}^{2}=\frac{24\pi^{2}}{\kappa^{2}V_{+}}\ ,$$ (51)

which is equal to the dS action of the false vacuum, but with an opposite (positive) sign.

The decay rate of the false vacuum per unit time per unit volume is given by:

$$\Gamma\sim(\kappa\sigma)^{4}e^{-\frac{24\pi^{2}}{\kappa^{2}V_{+}}}\ ,$$ (52)

and it does not depend on how deep the true vacuum is compared to the false one. If $\varepsilon>V_{+}$, the dS vacuum decays to AdS vacuum via bubbles, whose size is smaller than the radius of the AdS. For $\varepsilon<V_{+}$ it decays to another dS true vacuum with a smaller cosmological constant. The dependence on the height of the barrier is in the pre-exponential factor. The higher the barrier is, the smaller the critical bubbles are, and therefore the decay rate increases as $V_{bar}^{2}$ with an increasing height of the barrier. Thus, regardless of how high is the barrier between the false and true vacuum, the false dS vacuum is always unstable and decays via either the known field theory instantons or the gravitationally dominated instantons, depending on the parameters of the potential.

3.2 The instability of the Minkowski vacuum

The Minkowski vacuum can decay if the true vacuum is an AdS vacuum separated from Minkowski vacuum by the barrier of height $V_{bar}$. The formulas describing this instability can be obtained by considering the corresponding limiting case in the formulas derived above. In particular, taking $R_{+}\rightarrow\infty$ in (24) leads to the following expression for the radius of the critical bubble:

$$\varrho_{b}=\frac{3\sigma}{\varepsilon\left|1-\alpha\right|}\ .$$ (53)

If $\alpha<1$, we obtain from (35) the following action

$$S^{f}-S^{i}=\frac{27\pi^{2}}{2}\frac{\sigma^{4}}{\varepsilon^{3}(1-\alpha)^{2}}\ .$$ (54)

For $\alpha\ll 1$, these results agree with the formulas describing the field theory instantons (see (38) and (41)).

The ratio of $\varrho_{b}$ to AdS radius $R_{AdS}\equiv\sqrt{3/(\kappa\left|V_{-}\right|)}$ is

$$\frac{\varrho_{b}}{R_{AdS}}=\frac{2\sqrt{\alpha}}{1-\alpha}\ ,$$ (55)

and therefore if

$$\alpha=\frac{3\kappa\sigma^{2}}{4\varepsilon}\simeq\frac{3\kappa\varphi_{0}^{2}V_{bar}}{4\varepsilon}\$$ (56)

is of order one, the gravitational effects become important. In particular, when $\alpha\rightarrow 1$, the action tends to infinity. Note, that the solution of (13) in Minkowski space $\varrho=+\xi$, which corresponds to $\dot{\varrho}_{+}=1$ for the false vacuum, covers the entire Euclidean space in the original coordinates (6), where $\bar{\xi}$ changes in the range from zero to infinity. To consider the case $\alpha>1$, in which the radius of the bubble is given by (50) for $\alpha\gg 1$, we should replace the expanding coordinates in the false Minkowski vacuum by the contracting coordinates. This in turn corresponds to an infinite change of the action, and the resulting action is infinite. This can also be seen from (35), where the denominator of $1+\dot{\varrho}_{+}$ vanishes. Consequently, the probability of the transitions vanishes for $\alpha>1$.

Therefore, for a given height of the barrier $V_{bar}$, the Minkowski vacuum is only unstable if the depth of the true AdS vacuum is large enough, i.e.

$$\left|V_{-}\right|=\varepsilon>\frac{3}{4}\kappa\varphi_{0}^{2}V_{bar}\ .$$ (57)

Otherwise the Minkowski vacuum is stable, despite the existence of the true AdS vacuum. The thin-wall approximation is of course only valid for $\left|V_{-}\right|\ll V_{bar}$. The results obtained are in complete agreement with the results for the dS false vacuum, if we take the limit $V_{+}\rightarrow 0$ in the formulas obtained in the previous section.

3.3 The transitions between the AdS vacua

Let’s look at the decay of the false AdS vacuum into another true AdS vacuum. In this case $V_{+}$ and $V_{-}$ are negative, and we assume that the magnitudes $\left|V_{+}\right|$ and $\left|V_{-}\right|$ are both much smaller than the positive height of the barrier $V_{bar}$. Taking into account that $V_{+}<0$, we find from (24) the radius of the critical bubble:

$$\varrho_{b}=\left[\frac{R_{+}^{2}}{\left(\frac{\varepsilon}{3\sigma}\left(1-\alpha\right)R_{+}\right)^{2}-1}\right]^{1/2}\ ,$$ (58)

where now $R_{+}=\sqrt{3/(\kappa\left|V_{+}\right|)}$ is the radius of the false AdS vacuum. It is always larger that the radius of the true AdS vacuum $R_{-}=\sqrt{3/(\kappa\left|V_{-}\right|)}$, because $\left|V_{-}\right|>\left|V_{+}\right|$.

It should be noted that the radius of the bubble is real when

$$\frac{\varepsilon}{3\sigma}\left|1-\alpha\right|R_{+}>1\ .$$ (59)

Comparing this inequality with (42), and taking into account the fact that in this case we have to replace $V_{+}>0$ for dS with $\left|V_{+}\right|$ for AdS, we find that the critical bubbles exist only if either

$$\varepsilon>\frac{3}{4}\kappa\varphi_{0}^{2}V_{bar}\left(1+\sqrt{\frac{\left|V_{+}^{*}\right|}{V_{bar}}}\right)\quad,$$ (60)

or

$$\varepsilon<\frac{3}{4}\kappa\varphi_{0}^{2}V_{bar}\left(1-\sqrt{\frac{\left|V_{+}^{*}\right|}{V_{bar}}}\right),$$ (61)

where

$$\left|V_{+}^{*}\right|\equiv\frac{16\left|V_{+}\right|}{3\kappa\varphi_{0}^{2}}\ .$$ (62)

The $\varepsilon-V_{bar}$ plane in the case of AdS is identical to Fig.2, if we replace $V_{+}$ by its magnitude of $\left|V_{+}\right|$, when we consider the instability of the AdS false vacuum. As we have shown above, in region II the equation for the critical bubble in the case of AdS has no solutions, unlike the dS decay. Therefore, we only need to calculate the probability of AdS decay in the regions I and III, which correspond to the cases $\alpha<1$ and $\alpha>1$.

Equation (13) in the AdS case takes the form:

$$\dot{\varrho}^{2}=1+\frac{\varrho^{2}}{R_{+}^{2}}\ ,$$ (63)

and its expanding solution,

$$\varrho\left(\xi\right)=R_{+}\sinh\left(\frac{\xi}{R_{+}}\right)\ ,$$ (64)

does not cover the entire AdS false vacuum. Returning to the original coordinates (6), we find that the $\bar{\xi}$ only changes in the interval $0\leq\bar{\xi}\leq 1$ for $0\leq\xi<\infty$. Therefore, when considering transitions between AdS vacua, it is convenient to work with the original coordinates (6). As can be easily verified:

$$\varrho\left(\bar{\xi}\right)=f\left(\bar{\xi}\right)\bar{\xi}=\frac{2R_{+}\bar{\xi}}{\left|1-\bar{\xi}^{2}\right|}\ ,$$ (65)

where $0\leq\bar{\xi}\leq 1$ describes the expanding branch of AdS, and $1\leq\bar{\xi}\leq\infty$ its contracting branch, thus completely covering the false AdS vacuum. Note, that $\bar{\xi}=1-0$ corresponds to $\xi=+\infty$, while at $\bar{\xi}=1+0$ we have $\xi=-\infty$. The metric (6) covering the entire AdS space thus becomes

$$ds^{2}=\frac{4R_{+}^{2}}{\left(1-\bar{\xi}^{2}\right)^{2}}\left[d\bar{\xi}^{2}+\bar{\xi}^{2}\left(d\chi^{2}+\cos^{2}\chi d\Omega^{2}\right)\right]\ ,$$ (66)

where $0\leq\bar{\xi}<\infty$. Note, that $\bar{\xi}$ in this metric can be shifted by an arbitrary constant $C$, $\bar{\xi}\rightarrow\bar{\xi}+C$, which shifts the range to $-C\leq\bar{\xi}<\infty$.

Case (a). Let us first consider $\alpha<1$, i.e., the region I in Fig.2, where $V_{+}$ is replaced by $\left|V_{+}\right|$. In this case we have to match the expanding true AdS vacuum solution with radius $R_{-}$:

$$\varrho\left(\bar{\xi}\right)=\frac{2R_{-}\left(\bar{\xi}+C\right)}{\left|1-\left(\bar{\xi}+C\right)^{2}\right|}\ ,$$ (67)

with the expanding false vacuum solution with $R_{+}>R_{-}$, given by (65). The matching point $\bar{\xi}_{b}$, and the constant $C$ are determined by the fact that $\varrho$ must be continuous at $\bar{\xi}_{b}$, and its derivatives must satisfy (20), from which it follows that $\dot{\varrho}_{+}<\dot{\varrho}_{-}$.

Figure 3: $\varrho$ versus $\bar{\xi}$ when $\alpha<1$: matching the expanding true AdS vacuum solution with the expanding false vacuum solution at the matching point $\bar{\xi}_{b}$.

As can be seen from Fig.3, in this case $\bar{\xi}_{b}<1-C$ and $C>0$. It follows from (16), using (7), that the contribution of the false AdS vacuum in the final configuration, which is calculated by subtracting the action for the bubble with radius $\varrho_{b}$ with potential $V_{+}$ from the initial false vacuum action, is:

$$S_{+}^{f}=S^{i}+2\pi^{2}V_{+}\intop_{0}^{\bar{\xi}_{b}}\frac{\varrho^{4}}{\bar{\xi}}d\bar{\xi}+\frac{6\pi^{2}}{\kappa}\varrho_{b}^{2}\dot{\varrho}_{+}=S^{i}+\frac{12\pi^{2}}{\kappa^{2}V_{+}}\left(1-\dot{\varrho}_{+}^{3}\right)\ ,$$ (68)

where

$$\dot{\varrho}_{+}=+\sqrt{1+\frac{\varrho_{b}^{2}}{R_{+}^{2}}}\ .$$ (69)

Similarly, we obtain the contribution of the true AdS vacuum inside the bubble to the final action:

$$S_{-}^{f}=-2\pi^{2}V_{-}\intop_{-C}^{\bar{\xi}_{b}}\frac{\varrho^{4}}{\left(\bar{\xi}+C\right)}d\bar{\xi}-\frac{6\pi^{2}}{\kappa}\varrho_{b}^{2}\dot{\varrho}_{-}=-\frac{12\pi^{2}}{\kappa^{2}V_{-}}\left(1-\dot{\varrho}_{-}^{3}\right)\ ,$$ (70)

where

$$\dot{\varrho}_{-}=+\sqrt{1+\frac{\varrho_{b}^{2}}{R_{-}^{2}}}\ .$$ (71)

Finally, if these results are combined with the wall contribution to the final action $S_{w}^{f}=2\pi^{2}\sigma\varrho_{b}^{3}$, the same expression (35) is obtained:

$$S^{f}-S^{i}=\frac{2\pi^{2}\sigma\varrho_{b}^{3}}{\left(1+\dot{\varrho}_{+}\right)\left(1+\dot{\varrho}_{-}\right)}\ ,$$ (72)

where $\sigma$ is the surface tension of the bubble and $\varrho_{b}$, $\dot{\varrho}_{+}$ and $\dot{\varrho}_{-}$ are given in (58), (69) and (71), respectively.

According to (58), the radius of the critical bubble in region I in Fig.2 can change from $3\sigma/\varepsilon$ to infinity, as we approach the edge of this region. We obtain the results of standard field theory (38) and (41), when $\varrho_{b}\ll R_{+},R_{-}$. For the bubble with the radius $R_{+}\gg\varrho_{b}\gg R_{-}$, gravity plays a significant role and the action (72) becomes:

$$S^{f}-S^{i}\simeq\pi^{2}\sigma\varrho_{b}^{2}R_{-}\ .$$ (73)

If $\varrho_{b}\gg R_{+},R_{-}$ then,

$$S^{f}-S^{i}\simeq 2\pi^{2}\sigma\varrho_{b}R_{+}R_{-}\ ,$$ (74)

and it becomes infinite when

$$\varepsilon\rightarrow\frac{3}{4}\kappa\varphi_{0}^{2}V_{bar}\left(1+\sqrt{\frac{\left|V_{+}^{*}\right|}{V_{bar}}}\right)\ .$$ (75)

Case (b). Let us now turn to the case $\alpha>1$, i.e., the region III in Fig.2, where $V_{+}$ is replaced by $\left|V_{+}\right|$, corresponding to

$$\frac{3}{4}\kappa\varphi_{0}^{2}V_{bar}\left(1-\sqrt{\frac{\left|V_{+}^{*}\right|}{V_{bar}}}\right)>\varepsilon>0.$$

Figure 4: $\rho$ versus $\bar{\xi}$ when $\alpha>1$: matching the expanding true AdS vacuum solution with the contracting false vacuum solution at the matching point $\bar{\xi}_{b}$.

In this case, the solution for the critical bubble exists and we have to calculate the action with this solution. As can be seen from Fig.4, we need to match the expanding branch of the true AdS vacuum with the contracting branch of the false AdS vacuum at the point $1+\left|C\right|>\bar{\xi}_{b}>1$, where $C$ in (67) must be negative. To calculate the contribution of the false AdS vacuum to the action after the bubble formation, we proceed as before, and subtract from the initial action the action for the false vacuum bubble with the radius $\varrho_{b}$. In this case, there is a singularity at $\bar{\xi}=1$, and to regularize this singularity, we divide the integration domain into two subdomains $(1-\delta)^{1/2}\geq\bar{\xi}\geq 0$, and $\bar{\xi}_{b}\geq\bar{\xi}\geq(1-\delta)^{1/2}$, where $\delta\ll 1$. The result is:

$$S_{+}^{f}=S^{i}+2\pi^{2}V_{+}\left[\intop_{0}^{(1-\delta)^{1/2}}\frac{\varrho^{4}}{\bar{\xi}}d\bar{\xi}+\intop_{(1+\delta)^{1/2}}^{\bar{\xi}}\frac{\varrho^{4}}{\bar{\xi}}d\bar{\xi}\right]+\frac{6\pi^{2}}{\kappa}\left[\left.\varrho^{2}\dot{\varrho}\right|_{0}^{(1-\delta)^{1/2}}+\left.\varrho^{2}\dot{\varrho}\right|_{(1+\delta)^{1/2}}^{\bar{\xi}_{b}}\right]$$

$$=S^{i}-\frac{24\pi^{2}}{\kappa^{2}V_{+}}\left(\frac{8+6\delta^{2}}{\delta^{3}}\right)+\frac{12\pi^{2}}{\kappa^{2}V_{+}}\left(1+\dot{\varrho}_{+}^{3}\right)\ ,$$ (76)

where $\dot{\varrho}_{+}$ is defined in (69). The contributions of the true vacuum and the wall are both finite. If we consider that $V_{+}$ is negative, we see that $S^{f}-S^{i}$ tends to $+\infty$ when $\delta\rightarrow 0$, and thus the probability of transitions for $\alpha>1$ (region III in Fig.2) vanishes, although the corresponding solution for the bubble exists. From this we conclude, that for a given $V_{+}$ and $V_{bar}$, the AdS false vacuum is stable if

$$\varepsilon=V_{+}-V_{-}<\frac{3}{4}\kappa\varphi_{0}^{2}V_{bar}\left(1+\sqrt{\frac{\left|V_{+}^{*}\right|}{V_{bar}}}\right)\ ,$$ (77)

despite the presence of a deeper true vacuum $V_{-}<V_{+}.$

4 Discussion

Working within the thin-wall approximation, we have shown that the false dS vacuum is always unstable, irrespective of the height of the barrier $V_{bar}$, which separate the false dS vacuum with positive potential $V_{+}$ from the true vacuum with potential $V_{-}$. The true vacuum can be either a dS vacuum with a smaller potential, a Minkowski vacuum or an AdS vacuum. This conclusion is not very intuitive from field theory perspective. Namely, if the radius of the bubble obtained by neglecting gravity, $\varrho_{b}=3\sigma/\varepsilon$, starts to exceed the dS radius of the false vacuum $R_{+}$, one might naively expect that tunneling out of the false vacuum cannot take place, and it becomes stable despite the presence of the true vacuum with a smaller potential. This can happen, for example, if, either the surface tension $\sigma\sim\varphi_{0}V_{bar}^{1/2}$ is too large or the distance between the depths of the vacua $\varepsilon=V_{+}-V_{-}$ is too small.

However, as we have seen, gravity becomes very important in such circumstances and the negative contribution of gravity to energy cannot be neglected, when either the vacua are almost degenerate, i.e., $\varepsilon\ll\varphi_{0}\sqrt{\kappa V_{+}V_{bar}}$ for $V_{bar}<V_{+}/\left(\kappa\varphi_{0}^{2}\right)$, or if the height of the barrier exceeds $V_{+}/\left(\kappa\varphi_{0}^{2}\right)$ independently of $\varepsilon$. In the first case, the vacuum decay occurs via critical bubbles with a radius of the order of the false dS radius $R_{+},$ and the decay rate is determined by the action (45), while in the second case for a very high potential barrier the size of the bubble can be much smaller than the dS radius and the decay rate is given by the action (51), which depends only on the potential in the false vacuum. Since $V_{bar}$ cannot exceed the Planck scale, the minimal size of these bubbles is about $l_{Pl}/\sqrt{\kappa\varphi_{0}^{2}}\sim 1/\varphi_{0}$. It should be noted that the decay rate via gravitationally dominated instantons is exponentially small for the small initial cosmological constant, and vanishes when $V_{+}$ approaches zero.

In contrast, the Minkowski or AdS false vacua are only unstable if the relative depth of the true vacuum exceeds the critical values given in (57) or (77). Otherwise, the false vacuum is stable despite the existence of the true vacuum with $V_{-}<V_{+}$.