Ive been reading Griffiths's introduction to Quantum mechanics and just reached the chapter about spin. I really dont understand Griffiths's derivation of the eigenspinors of for instance $S_x$. This is supposed to be a very simple calculation, basically just calculating the eigenvectors to the following matrix:

$$S_x = \begin{pmatrix} 0 & \frac{h}{4\pi} \\ \frac{h}{4\pi} & 0 \\ \end{pmatrix}$$

When calculating the eigenvectors for this very simple matrix i get: $$v_1 = \begin{pmatrix} 1\\ 1 \end{pmatrix}$$ and $$v_2 = \begin{pmatrix} -1\\ 1 \end{pmatrix}$$

This, however, seems to be the incorrect result. because Griffiths gets the same $v_1$ but $$v_2 = \begin{pmatrix} 1\\ -1 \end{pmatrix}$$

I really don't understand this, because $v_2$ is not even an eigenvector to $S_x$. Letting Griffiths $v_2$ transform with $S_x$ should give a vector that is proportional to his $v_2$, but that ends up not happening. It's instead proportional to my $v_2$, and therefore $v_2$ can't be an eigenvector of $S_x$. I've obviously messed up somewhere, but I really don't see what it might be. I'd really appreciate some help. Thanks.

  • $\begingroup$ Note that if $\Lambda \vec v=\lambda \vec v$ for some operator $\Lambda$ and some eigenvector $\vec v$, then $\alpha \vec v$ is also an eigenvector $\forall \alpha\in \Bbb C-\{0\}$ (assuming your vector space is defined over $\Bbb C$ as is the case here) $\endgroup$
    – Charlie
    Commented Jul 30, 2022 at 19:51

1 Answer 1


His $v_2$ and your $v_2$ simply differ by a $-1$ factor, so they're colinear. If one is an eigenvector of $S_x$, the other is too, for the same eigenvalue.

  • $\begingroup$ Ohh i see, that makes sense. Thanks for your answer, but why is that the normal convention for defining the eigenstate? Feels so unnatural $\endgroup$
    – lumpludde
    Commented Jul 30, 2022 at 19:58
  • $\begingroup$ As far as I know, there is no convention. For a eigenvalue with multiplicity 1, the eigenvectors form a dimension-1 vector space, in other words they're all colinear. In quantum mechanics, the only important thing is that eigenvectors are usually normalized, which means that there are a infinite number of them, differing by a phase. $\endgroup$
    – Miyase
    Commented Jul 30, 2022 at 20:01

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.