\- Let \(\alpha_{\max }\) be the largest possible value of \(\alpha \pm m \hbar
.\) By definition then, we have that
\(\hat{L}_{z} \psi_{\alpha_{n u} \beta}=\alpha_{\max } \psi_{\alpha_{\max }
\beta}\)
\(\hat{L}^{2} \psi_{\alpha_{\max }^{p}}=\beta^{2} \psi_{\alpha_{\max } \beta}\)
and
$$
\hat{L}_{+} \psi_{\alpha_{\max } \beta}=0
$$
Operate on the last equation with \(\hat{L}_{-}\)to obtain
$$
\begin{aligned}
\hat{L}_{-} \hat{L}_{+} \psi_{\alpha_{\max }^{\beta}} &=0 \\
&=\left(\hat{L}^{2}-\hat{L}_{z}^{2}-\hbar \hat{L}_{z}\right)
\psi_{\alpha_{\max } \beta}
\end{aligned}
$$
and
$$
\beta^{2}=\alpha_{\max }^{2}+\hbar \alpha_{\max }
$$Use a parallel procedure on \(\psi_{\alpha_{\min }} \beta\) to obtain
$$
\beta^{2}=\alpha_{\min }^{2}-h \alpha_{\min }
$$
Now show that \(\alpha_{\max }=-\alpha_{\min }\), and then argue that the
possible values of the eigenvalues \(\alpha\) of \(\hat{L}_{z}\) extend from
\(+\alpha_{\max }\) to \(-\alpha_{\max }\) in steps of magnitude \(h\). This is
possible only if \(\alpha_{\max }\) is itself an integer (or perhaps a half-
integer) times \(\hbar\). Finally, show that this last result leads to
$$
\beta^{2}=l(l+1) \hbar^{2} \quad l=0,1,2, \ldots
$$
and
$$
\alpha=m \hbar \quad m=0, \pm 1, \pm 2, \ldots, \pm l
$$