In Coleman's paper Fate of the false vacuum: Semiclassical theory while working out the exponential coefficient for tunneling probability through a potential barrier, he studies the problem with Wick's rotation $\tau=it$, getting to the Euclidean Lagrangian

$$L_E = \frac{1}{2}\left(\frac{dq}{d\tau}\right)^2+V(q),\tag{2.14}$$

where clearly the potential is inverted. The potential as given in the paper is this

enter image description here

He then states that from conservation of energy formula


i quote

"By eq. (2.12)"-the conservation of energy-"the classical equilibrium point, $q_0$, can only be reached asymptotically, as $\tau$ goes to minus infinity" $$\lim_{\tau\rightarrow-\infty}q = q_0.\tag{2.15}$$

  • Q1. Why is this true? How you define infinity for a complex number?

Then, by translation invariance, he sets the time at which the particle reaches $\sigma$ as $\tau=0$ and that


He goes on by saying that this condition

"[...] also tells us that the motion of the particle for positive $\tau$ is just the time reversal of its motion for negative $\tau$; the particle simply bounces off $\sigma$ at $\tau=0$ and returns to $q_0$ at $\tau=+\infty$."

  • Q2. Even this isn't very clear for me. Why should the condition for zero velocity at $\sigma$ imply that?

Is there something really basic that I'm missing? I'm not very competent in Wick's rotations and such and I have to understand every little bit of this paper for my bachelor's thesis.

  • $\begingroup$ $\tau$ is real. That's why we call it a rotation -- otherwise it would just be a change of symbol, i.e., merely notation. $\endgroup$ – AccidentalFourierTransform Sep 8 at 12:52
  1. Here we will work in the Euclidean$^1$ formulation, where Euclidean time $\tau\in\mathbb{R}$ is real. The potential is assumed to satisfy $$ V(q_0)~=~0~=~V(\sigma) ,\tag{A} $$ cf. OP's figure. The bounce solution satisfies the following boundary conditions (BCs) $$ \dot{q}(\tau_i)~=~0\quad \wedge\quad q(\tau_i)~=~q_0\quad \wedge\quad q(0)~=~\sigma,\tag{B} $$ where $q\equiv |\vec{q}|$ and dot means differentiation wrt. $\tau$.

  2. The potential is minus $V$. Energy is conserved (and equal to zero, since that's what it was in the beginning $\tau\!=\!\tau_i$). Therefore $$ \frac{1}{2} \dot{q}^2-V(q)~=~0 \qquad\Leftrightarrow\qquad \dot{q}~=~\pm \sqrt{2V(q)} .\tag{C} $$ Therefore we find that (minus) the initial time is $$ -\tau_i~=~ \int_{q_0}^{\sigma}\frac{dq}{\sqrt{2V(q)}}.\tag{D} $$

  3. Now since we assume that the potential is approximately quadratic $$V(q) ~\propto ~(q-q_0)^2 \quad\text{near}\quad q~\approx~ q_0\tag{E},$$ cf. OP's figure, we see that the integrand (D) has a simple pole at the lower integration limit $q\!=\! q_0$, implying that the integral $\tau_i=-\infty$ cannot be finite. This answers OP's first question$^1$.

  4. From energy conservation we see that $$ \dot{q}(0)~=~0.\tag{F} $$ By time reversal symmetry and uniqueness of solutions to the first-order ODE (C), it follows that the final time $$\tau_f~=~ \int_{q_0}^{\sigma}\frac{dq}{\sqrt{2V(q)}}\tag{G} $$ is given by the very same formula (D). This answers OP's second question.

  5. For more details, see this related Phys.SE post and links therein.


$^1$ Concerning Wick rotation, see e.g. this Phys.SE post and links therein.

  • $\begingroup$ Thank you very much! Now everything is clear $\endgroup$ – Davide Morgante Sep 8 at 14:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.