2
$\begingroup$

This question already has an answer here:

Prove: Product of any polynomial of $x$ with $\psi(x)$ or any of its derivatives goes to zero in the limit $x\to\pm\infty$.

This comes from a footnote written by the professor in his quantum mechanics problem set and I am curious to know why. I have unsuccessfully attempted L'Hôpital, Taylor expansion, and many other weird methods. Any help is appreciated.

$\endgroup$

marked as duplicate by AccidentalFourierTransform, Cosmas Zachos, Valter Moretti, ACuriousMind quantum-mechanics Jun 29 at 16:27

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • 1
    $\begingroup$ What is $\Psi (x)$ equal to? $\endgroup$ – DanielC Jun 28 at 23:17
  • 4
    $\begingroup$ Then what you have to prove is wrong. There are normalizable wavefunctions whose product with a polinomial in x go to infinity, when x goes to infinity. $\endgroup$ – DanielC Jun 28 at 23:42
  • 2
    $\begingroup$ physics.stackexchange.com/q/30228 $\endgroup$ – DanielC Jun 28 at 23:51
  • 4
    $\begingroup$ @Zachary Any function is a solution to the Schrödinger equation, for some suitable potential. For a given $\psi$, just solve for $V$ in the Sch. eq. $\endgroup$ – AccidentalFourierTransform Jun 28 at 23:59
  • 2
    $\begingroup$ You've either dropped some of the crucial context, or your professor did by accident when writing the question. $\endgroup$ – knzhou Jun 29 at 1:44
2
$\begingroup$

In general, not all states in the Hilbert space satisfy your professor's requirement. One guess for what your professor had in mind is that they wanted the wave function to be "nice" enough that the $\hat{X}$ and $\hat{P}$ operators can act on the state arbitrarily many times while still outputting a physically valid state in the Hilbert space. In general, the $\hat{X}$ and $\hat{P}$ operators are not closed on the Hilbert space: there exist wave functions that are square-integrable, but if you act $\hat{X}$ on them enough times (i.e. multiply by them by a high enough power of $x$), then the result is no longer square-integrable. So technically, the $\hat{X}$ and $\hat{P}$ operators are not actually linear operators on the entire Hilbert space $\mathcal{L}_2(\mathbb{R}^n)$. But they are closed on a (non-complete) subspace of the Hilbert space called the Schwartz space. It's pretty straightforward to show that the Schwartz space respects your professor's claim. Your professor may have been considering this special subset of "nice" wave functions, but a general state doesn't meet the stated requirement.

$\endgroup$
  • 2
    $\begingroup$ Hilbert space means complete complex vector space with respect to the norm induced by the scalar product.for a quantum system it is unique. No other definitions exist. X and P are never defined on the whole Hilbert space, so " as they should be" does not mean anything. This answer is misleading in my view. $\endgroup$ – Valter Moretti Jun 29 at 6:28
  • 2
    $\begingroup$ Observables, described by selfadjoint operators, are used in physics because they posses an associated decomposition in terms of eigevectors (also generalized, more properly in terms of a projection valued measure). The existence of such a decomposition is one of the most important facts for the physical interpretation of the general formalism of Quantum Theory. Completeness of the space is necessary to assure the existence of this decomposition for every selfadjoint operator. $\endgroup$ – Valter Moretti Jun 29 at 6:44
  • 2
    $\begingroup$ So, this requirement cannot be relaxed dealing for instance with spaces of differentiable functions instead of the whole Hilbert space where most elements are not even continuous. $\endgroup$ – Valter Moretti Jun 29 at 6:44
  • $\begingroup$ @ValterMoretti You're right, I fixed my answer. $\endgroup$ – tparker Jun 29 at 11:04
3
$\begingroup$

A counterexample where one can provide an explicit Hamiltonian is the case of a particle under a constant force (linear potential).

$$ \frac{\mathrm{d}^2}{\mathrm{d}x^2} \psi(x) = F\cdot \,(x-x_0)\,\psi(x)\,. $$ The solution is famously given by the Airy function. $$ \psi(x) = c_A \mathrm{Ai}(\sqrt[3]{F}\cdot(x-x_0)) + c_B \mathrm{Bi}(\sqrt[3]{F}\cdot(x-x_0))\,. $$ If $x < x_0$ and $|x-x_0|\gg1$, then, letting $z= - \sqrt[3]{F}\cdot(x-x_0)$, $$ \mathrm{Ai}(-z) \sim \frac{\sin \left(\frac23z^{\frac{3}{2}}+\frac{\pi}{4} \right)}{\sqrt\pi\,z^{\frac{1}{4}}} \left(1 + O(z^{-3/2})\right)\,. $$ This clearly does not go to zero when multiplied by a polynomial.

$\endgroup$
  • 1
    $\begingroup$ I agree. But this is a physical system that exists and so any general claim about QM must admit the possibility that a Hilbert state containing these functions is a valid choice. $\endgroup$ – MannyC Jun 29 at 3:44
  • 1
    $\begingroup$ A free particle under a constant force. $\endgroup$ – MannyC Jun 29 at 3:49
  • $\begingroup$ Parenthetically, I'd say that "free particle under a constant force" is a contraction in terms; if the particle experiences a linear potential then it isn't free. $\endgroup$ – tparker Jun 29 at 3:50
  • 2
    $\begingroup$ I borrowed the terminology "free" from QFT, where interactions are those from cubic degree onwards. Anyway physics is no stranger to idealized model to describe nature. It's not necessary for the force to be constant out to spatial infinity. It's sufficient that it's constant on a scale much larger than the particle at the point that no difference can be detected experimentally. $\endgroup$ – MannyC Jun 29 at 4:07
  • $\begingroup$ If there are linear and quadratic terms, then you can always complete the square and eliminate the linear term via a constant shift in the field or $\hat{X}$ operator, so the linear term is indeed qualitatively unimportant. But if there's only a linear term then this trick doesn't work, and the space of solutions clearly changes qualitatively from the free case. $\endgroup$ – tparker Jun 29 at 12:00

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