The Commutators of the Angular Momentum Operators


\begin{displaymath}\bgroup\color{black} [L_x,L_y]\neq 0 ,\egroup\end{displaymath}

however, the square of the angular momentum vector commutes with all the components.

\begin{displaymath}\bgroup\color{black}[L^2,L_z]=0\egroup\end{displaymath}

This will give us the operators we need to label states in 3D central potentials.

Lets just compute the commutator.

\begin{eqnarray*}[L_x,L_y]&=&[yp_z-zp_y,zp_x-xp_z]=y[p_z,z]p_x+[z,p_z]p_yx \\
&=&{\hbar\over i}[yp_x-xp_y]=i\hbar L_z \\
\end{eqnarray*}


Since there is no difference between \bgroup\color{black}$x$\egroup, \bgroup\color{black}$y$\egroup and \bgroup\color{black}$z$\egroup, we can generalize this to

\begin{displaymath}\bgroup\color{black} [L_i,L_j]=i\hbar\epsilon_{ijk}L_k \egroup\end{displaymath}

where \bgroup\color{black}$\epsilon_{ijk}$\egroup is the completely antisymmetric tensor and we assume a sum over repeated indices.

\begin{displaymath}\bgroup\color{black} \epsilon_{ijk}=-\epsilon_{jik}=-\epsilon_{ikj}=-\epsilon_{kji} \egroup\end{displaymath}

The tensor is equal to 1 for cyclic permutations of 123, equal to -1 for anti-cyclic permutations, and equal to zero if any index is repeated. It is commonly used for a cross product. For example, if

\begin{displaymath}\bgroup\color{black} \vec{L}=\vec{r}\times\vec{p} \egroup\end{displaymath}

then

\begin{displaymath}\bgroup\color{black} L_i=r_jp_k\epsilon_{ijk} \egroup\end{displaymath}

where we again assume a sum over repeated indices.

Now lets compute commutators of the \bgroup\color{black}$L^2$\egroup operator.

\begin{eqnarray*}
L^2&=&L_x^2+L_y^2+L_z^2 \\
{[L_z,L^2]}&=&[L_z,L_x^2]+[L_z,L...
..._y[L_z,L_y] \\
&=&i\hbar (L_yL_x+L_xL_y-L_xL_y-L_yL_x)=0 \\
\end{eqnarray*}


We can generalize this to

\begin{displaymath}\bgroup\color{black} [L_i,L^2]=0 .\egroup\end{displaymath}

\bgroup\color{black}$L^2$\egroup commutes with every component of \bgroup\color{black}$\vec{L}$\egroup.

Jim Branson 2013-04-22