So there are two unit vectors $\hat{m}$ and $\hat{n}$ with arbitrary directions in 3D space.

There is a spin operator along a particular direction in space, say that of $\hat{r}$, is:

$\sigma_r= \begin{bmatrix} r_z & r_x-ir_y \\ r_x+ir_y & -r_z \end{bmatrix}$

..which has the eigenvalues +1,-1.

Here is the question:

Suppose that a spin is prepared so that $\sigma_m = +1$. The apparatus is then rotated to the $\hat{n}$ direction and $\sigma_n$ is measured. What is the probability that the result is $+1$?

Now I've tried this using spherical coordinates, and Cartesian coordinates but it always ends up in a huge spaghetti of algebra. I have a feeling I'm doing something subtly wrong or missing some fundamental facts.

Here is what my procedure of attempt was:

I find the eigenvector (the possible state) of $\sigma_m$ corresponding to the eigenvalue +1. I get the normalized vector:

$|m_+> \ = \frac{1}{\sqrt{2(1-m_z)}} \begin{bmatrix} m_x-im_y \\ 1-m_z \end{bmatrix}$

Similarly, I find the same for the $\sigma_n$ operator:

$|n_+> \ = \frac{1}{\sqrt{2(1-n_z)}} \begin{bmatrix} n_x-in_y \\ 1-n_z \end{bmatrix}$

Then, for the probability I simply find (or try to..) $P = |<n_+|m_+>|^2$. The answer given is $\cos^2(\rho/2)$ where $cos(\rho)=\hat{n}.\hat{m}$.

Am I attempting this correctly?


If you want to reduce the "spaghetti of algebra" you can reorient the coordinates. If $\hat{z}'=\hat{m}$ then in spherical coordinates you have $$ \sigma_\hat{m}= \left[ \begin{array}{cc} \cos(\theta) & \sin(\theta) \\ \sin(\theta) & -\cos(\theta) \\ \end{array} \right] $$ where $\cos(\theta)=\hat{n} \cdot \hat{m}$. It's easier to find the eigenvectors of this matrix using: $$ \sin(\theta/2)= \sin(\theta-\theta/2)= \sin(\theta)\cos(\theta/2)-\cos(\theta)\sin(\theta/2) $$ and the analogous equation for cosine.

Therefore $$ |n_{+}> = \left[ \begin{array}{c} \cos(\theta/2) \\ \sin(\theta/2) \\ \end{array} \right] $$ and finally: $$ |<m_+|n_+>|^2=\cos^2(\theta/2) $$

  • $\begingroup$ Thank you for your answer. Effectively, that is re-orienting our 3D coordinates such that m and n lie in the same plane formed by a pair of new axis - am I right? Also, what does the z' mean? $\endgroup$ – Zeeshan Ahmad Nov 10 '15 at 20:12
  • $\begingroup$ You're right, that way you have less variables. z' mean the z axis of the new coordinate system that way there is no azimuthal angle in the equations. $\endgroup$ – Hu Al Nov 11 '15 at 1:27

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.