Let $\mathbf{\sigma}=\sigma_1+\sigma_2+\sigma_3$, where the $\sigma_i$ are the Pauli matrices and define:

$$M:=i\mathbf{\sigma}\cdot \mathbf{v}$$

The claim is that if I change $M$ through a similarity change $M'=SMS^{-1}$, then I can find another vector $\mathbf{v}'$ such that $M'=i\mathbf{\sigma}\cdot \mathbf{v}'$. I checked it by explicitly finding the coefficients $\{f^i_1,f^i_2,f^i_3\}$ for each $\sigma_i$ such that

$$S\sigma_iS^{-1} = \sum_j f^i_j\sigma_j$$

For example, if $S=\begin{pmatrix} a & b \\ c & d \end{pmatrix}$, then

$$S\sigma_1S^{-1} = (bd-ac)\sigma_1 - i(bd-ac)\sigma_2+ (ab+bc)\sigma_3.$$

My question is: could I have convinced myself that I can find $\mathbf{v}'$ without actually going through the calculation?

  • $\begingroup$ What is the meaning of $\sigma\cdot v$? $\endgroup$
    – Daniel
    Nov 23, 2014 at 20:06
  • $\begingroup$ that $\sigma$ should actually be bold, but anyway $\mathbf{\sigma}\cdot v=v_1\sigma_1+v_2\sigma_2+v_3\sigma_3$ $\endgroup$
    – Rodrigo
    Nov 24, 2014 at 1:47

1 Answer 1


Let $M_2$ denote the set of complex matrices of order 2. Notice that $\sigma_1,\sigma_2,\sigma_3$ is a basis for the subspace $W=\{M\in M_2,\ tr(M)=0\}$

Notice that if $M$ is a linear combination of $\sigma_1,\sigma_2,\sigma_3$ then $M\in W$. Thus, $SMS^{-1}\in W$, because $tr(SMS^{-1})=tr(MS^{-1}S)=tr(M)=0.$ Thus, $SMS^{-1}$ is a linear combination of $\sigma_1,\sigma_2,\sigma_3$.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .