Context: (not asking for solution)

I'm attempting to show $\langle n,l',m'|\hat z|n,l,m \rangle = 0$ for $m\neq m'$ using the explicit form of $Y_{l,m}(\theta,\phi)$.


I wasn't sure how to handle the position operator $\hat z$, so my idea was to express it in spherical coordinates.

It is not clear to me if I'm allowed to do something like $\hat z = \hat r \cos(\theta)$

Do position operators work like this?

  • 1
    $\begingroup$ This would suggest that the radial unit vector $\hat{r}$ is always in the $z$-direction, which can't be true. See here for how to go between Cartesian and spherical unit vectors. $\endgroup$
    – march
    Commented Feb 20, 2020 at 1:48
  • 2
    $\begingroup$ You can also try to express the spherical harmonics in Cartesian. $\endgroup$ Commented Feb 20, 2020 at 3:46
  • $\begingroup$ @march I may be confused but I was under the impression that the position operator $/hat x$ was different than the unit vector $/hat x$ $\endgroup$ Commented Feb 20, 2020 at 4:24
  • 1
    $\begingroup$ I see. I misunderstood the hat to be indicating a unit vector rather than an operator. (That always gets me). Yes, then you are correct. $\endgroup$
    – march
    Commented Feb 20, 2020 at 4:57

1 Answer 1


You can write $z$ as

$$ z = r\left(\frac{4\pi}{3}\right)^{1/2} Y_{1,0}(\theta,\phi) $$

This amounts to evaluating the matrix element in the following way

$$ \langle n\;l^\prime m^\prime | z | nlm\rangle = \sqrt{\frac{4\pi}{3}}\int R_{n^\prime l^\prime}(r)R_{nl}(r)r^3\,dr\int Y_{l^\prime m^\prime}(\theta, \phi)Y_{10}(\theta, \phi)Y_{lm}(\theta, \phi)\,d\Omega$$

I carried out the full calculation just to give an idea of how to evaluate this integral in the case of linear polarization on the $z$ axis.

Since we are only interested into the projection of the angular momentum, we just have to consider the part of the spherical harmonics which depends upon $m$. This part, in the integral, is given by the product of the three exponential factors of the three spherical harmonics which depend on $\phi$, so $$ Y_{l^\prime m^\prime}Y_{1q}Y_{lm} \propto e^{-im^\prime \phi}e^{iq\phi}e^{im\phi}$$ in our case $q=0$. The integral in $\phi$ is non-vanishing only when $m^\prime = m+q$ which in our case amounts as saying that the integral is non-vanishing only is $m^\prime = m$

  • $\begingroup$ There's too much detail here. We don't provide complete answers to homework-like questions. They don't do the OP much good. $\endgroup$
    – garyp
    Commented Feb 20, 2020 at 11:31
  • 1
    $\begingroup$ @garyp Yes, i understand, but this calculations can easily be found everywhere, even in many books. I thought that giving an example to the question asked by the OP would help him understand the basics so that he could do the other calculations in a similar manner. I'll edit my answer anyway $\endgroup$ Commented Feb 20, 2020 at 11:35

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.