Physics Stack Exchange is a question and answer site for active researchers, academics and students of physics. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Can anyone point me to a reference (preferably either something online or something a small liberal arts school would be likely to have in its library) that goes through a derivation of the minimum uncertainty wavefunction in more detail than in the Griffiths?

Edit I've moved the second part of my original question to a separate post: 3D Minimum uncertainty wavepackets

share|cite|improve this question
great question, but it would be much better if you split the two things you've asked into two separate posts. You could edit this one to remove one of the parts, and then ask a separate question for that part alone. – David Z Mar 18 '11 at 5:33
For the first question, do you want to derive equation on that minimizing wavefunction (i.e. annihilation of anticommutator and certain colinearity condition) or do you want the solution to that equation? – Marek Mar 18 '11 at 9:26
okay, I answered the first part. But I agree with David that you should ask the second question separately. Nevertheless, what do you mean w.r.t. $\mathbf r$ and $\mathbf p$? The Gaussian does satisfy HUP w.r.t. each individual component, which is a same thing in my opinion. Also, you should remember that the HUP gets modified in more dimensions essentially because of Pythagorean theorem: the square of total error is a sum of squares of individual errors. Therefore your equation for $\Sigma$ should have 3 instead of 1 on the r.h.s. – Marek Mar 18 '11 at 11:52
up vote 3 down vote accepted

First some preliminaries. Suppose you have hermitian operators $A$ and $B$ and some state $\left| \psi \right>$. Denote by $\left<X\right>$ the expectation of $X$ in the state $\psi$, i.e. $\left<\psi\right| X \left| \psi\right>$. Denote by $\bar A := A - \left< A \right>$ and $\bar B := B - \left< B \right>$ the part of of $A$, resp. $B$ with vanishing expectation.

So, let's compute $\left< \bar A \right>^2 \left<\bar B \right>^2$. According to Cauchy-Schwarz inequality that this is always greater or equal than $\left|\left< \bar A \bar B \right>\right|^2$ (just plug in $\psi$ and interpret these expressions as scalar products). Now, we can express the product as sum of hermitian and antihermitian component $$\left|\left< \bar A \bar B \right>\right|^2 = {1\over4}\left< [A,B]/i \right>^2 + {1\over 4}\left< \{\bar A, \bar B\} \right>^2$$

(here we used the fact that $[\bar A, \bar B] = [A,B]$). If the commutator is just a number times identity operator then we can discard the expectations and after removing the anticommutator term (because it doesn't have any important interpretation and it doesn't spoil the inequality) we are left with HUP. But we're not interested in this application right now. Instead, we want to minimize the error term and that means we want equalities everywhere (it's not clear that it's possible to attain them, but let's assume this for a while). First, Cauchy-Schwarz inequality becomes equality if the vectors in the scalar product are colinear $$\bar B \left| \psi \right> = c \bar A \left| \psi \right>$$ Second, we want the expectation of anticommutator to vanish $$\left<\psi\right| \{\bar A, \bar B\} \left| \psi \right> = 0$$ So this gives us two equations for $\psi$. Let's see what we can get from them for $A = x$ and $B = p$. For simplicity let's assume that $\left<x\right> = \left<p\right> = 0$ (the general solution doesn't change anything much).

From first condition we obtain $$(p - cx) \left | \psi \right> = 0$$ which is a differential equation $$ (i \partial_x + cx) \psi(x) = 0$$ with a solution $\psi(x) = K \exp(-\alpha x^2)$ with ${\rm Re} \alpha > 0$ (so that this is indeed a vector from our Hilbert space) and $K$ being just a normalization constant. Finally from the anticommutator relation we get $\alpha = {1 \over 4(\Delta x)^2}$ and we're done.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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