1
$\begingroup$

While reading Shankar's book on Quantum Mechanics, I encountered an interesting problem:

Compute $\Delta T\cdot\Delta X$, where $T = P^2/2m$.

I found several solutions online which arrive at the result $\Delta T\cdot\Delta X \ge 0$.

My question is: does there exist a state $|{\psi}\rangle$ which saturates this inequality, i.e. for which $\Delta T\cdot\Delta X = 0$? We know $\Delta X\ne 0$ (from the uncertainty relation between $X$ and $P$), so then we must surely have that $\Delta T = 0$. But I'm struggling to imagine a physical state with a well-defined kinetic energy! If it is indeed possible, please provide an example of such a state. Thanks!

$\endgroup$
0

1 Answer 1

2
$\begingroup$

The minimum value of $\Delta x \Delta T$ is dictated by by the magnitude of the commutator $|[x,T]|=\hbar p/m$. Only if this vanishes can the inequality be saturated. This requires a momentum eigenstate with $p=0$, which means a constant wave function $\psi=c$.

Whether such a wave function can exist is a trickier question. Obviously, the domain of the wave function must be finite, or else $\psi$ will not be normalizable. If the particle is confined to a finite region by a potential, then $p=0$ will not be an energy eigenstate. This leaves us with the case of a finite region with periodic boundary conditions.

For a real particle, if $x$ is the Cartesian coordinate, this is not a realistic arrangement. (Periodic boundary conditions are often used in statistical mechanics, but there they are idealizations that are useful when the thermodynamic limit is taken.) However, if $x$ represents something like the position of a bead on a hoop, then the zero-momentum state is the real physical ground state. More generally, such a state will exist when the coordinate variable represents an angular position. For instance, in three dimensions, an S state wave function in a central potential has zero angular momentum and zero angular momentum uncertainty, although there is still position-energy uncertainty, because of the radial dependence of the kinetic energy.

$\endgroup$
6
  • $\begingroup$ I see. So for something like $\psi = c$, the momentum wave function would be a delta function at $x=0$, and so I guess it would follow that both momentum and kinetic energy are identically zero, hence $\Delta T = 0$. Right? The only shady thing though is that $\Delta X = \infty$ in this case... $\endgroup$
    – J-J
    Commented May 21, 2019 at 18:00
  • $\begingroup$ It's seeming more likely that no such "physical" state exists (constant wave functions aren't normalisable, after all). I'm fine if that's the case. $\endgroup$
    – J-J
    Commented May 21, 2019 at 18:02
  • $\begingroup$ +1: This is interesting tho. For the theory of a particle on a ring, a physical state can exist which saturates this bound. Since $\Delta x$ can only be finite as the eigenspectrum is compact, I can simply choose a momentum eigenstate with $p=0$ as you suggest and it would saturate this uncertainty bound, right? $\endgroup$
    – user87745
    Commented May 21, 2019 at 18:27
  • 1
    $\begingroup$ It doesn't require $p=0$, it's sufficient to have $p=\mathrm{const}$. I.e. $\psi=\exp(ikx)$. $\endgroup$
    – Ruslan
    Commented May 21, 2019 at 20:20
  • 1
    $\begingroup$ @J-J Yes, the eigenstates of the momentum operator are not physically realizable as they are not normalizable. But that is not the trick here as far as I understand. The point is that $\Delta T \Delta X$ is, in fact, $0\cdot \infty$ but, since $T\sim p^2$, if we regularize the product properly and evaluate this $0\cdot \infty$, it will turn out to be $0$. Intuitively, you can see this without explicitly performing the regularization via recalling that the $0\cdot \infty$ of $\Delta p \Delta x$ turns out to be finite after such a regularization. $\endgroup$
    – user87745
    Commented May 28, 2019 at 12:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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