So, take for instance the x and y pauli matrices $\sigma_1$ and $\sigma_2$, which have a representation in the $\sigma_3$ basis as

$$ \sigma_1 = \begin{pmatrix} 0 & 1 \\ 1 & 0 \end{pmatrix},\ \ \sigma_2 = \begin{pmatrix} 0 & -i \\ i & 0 \end{pmatrix} $$

Now consider dynamic evolution given by a Hamiltonian like $H = \lambda\sigma_3$. In the Heisenberg picture, $\sigma_1$ and $\sigma_2$ become time-dependent operators, whose time derivatives are

$$ \frac{d \sigma_1}{d t} \propto \sigma_2,\ \ \ \frac{d \sigma_2}{d t} \propto \sigma_1 $$

Now, clearly the above matrix form for $\sigma_1$ and $\sigma_2$ (which I've always kinda taken as a definition) becomes inappropriate, since all its elements are constant numbers.

My question is: how do I integrate these equations to obtain the 4 time-dependent elements of each matrix?


1 Answer 1


Integrating your equations is a bit silly. You just evaluate $$ \sigma_1(t) \equiv e^{-i\theta \sigma_3} \sigma_1 e^{i\theta \sigma_3}\\ \sigma_2(t) \equiv e^{-i\theta \sigma_3} \sigma_2 e^{i\theta \sigma_3} , $$ with $\theta\equiv t\lambda/\hbar$.

But Pauli vectors exponentiate trivially, $$ e^{i\theta \sigma_3} =\cos \theta +i \sigma_3 \sin\theta , $$ whence you just evaluate, $$ \sigma_1(t)= \sigma_1 \cos (2\theta) + \sigma_2\sin(2\theta) ,\\ \sigma_2(t)= \sigma_2 \cos (2\theta) - \sigma_1\sin(2\theta), $$ a uniform rotation, as you'd expect (why?).

  • $\begingroup$ Thank you! For some reason I had forgotten about the time evolution operator. I take it the uniform rotation of $\sigma_1$ and $\sigma_2$ is the Heisenberg picture statement of the Larmor precession, right? $\endgroup$ Mar 23, 2021 at 16:55
  • $\begingroup$ Basically. Rotation round the z-axis. $\endgroup$ Mar 23, 2021 at 17:22

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.