Suppose that the state $|p,\sigma\rangle$ (for a massless particle) has 3 momentum ${\bf p}=p_3$ (that is the momentum is in the $z$ direction) and that $J_3|p,\sigma\rangle=\sigma|p,\sigma\rangle$ where ${\mathbf{P }}=(P_1,P_2,P_3)$ is the momentum operator and ${\mathbf{J }}=(J_1,J_2,J_3)$ the angular momentum operator. Than

$$\frac{{\bf J} \cdot {\mathbf{P }}}{{|\mathbf{P}|}}|p,\sigma\rangle=\frac{{\bf J_3} {\mathbf{P_3 }}}{{|\mathbf{P}|}}|p,\sigma\rangle =\sigma|p,\sigma\rangle$$

Now suppose that we make a rotation around the $x$ axis than for a massless particle the state should transform like this $$U(\Lambda)|p,\sigma\rangle=e^{i\theta\sigma}| \Lambda p,\sigma\rangle.$$

Since $\mathbf{J_2}|\Lambda p,\sigma\rangle=0$, how can we proof that $$\frac{{\bf J} \cdot {\mathbf{P }}}{{|\mathbf{P }|}}|\Lambda p,\sigma\rangle=\frac{p'_3 \mathbf{J_3 }}{|\mathbf{P }|}|\Lambda p,\sigma\rangle=\frac{p'_3 \sigma}{|\mathbf{P }|}|\Lambda p,\sigma\rangle.$$

is equal to

$$\sigma|\Lambda p,\sigma\rangle$$

or since rotation leaves the norm of a vector invariant, that is the same as to proof that
$$\frac{p'_3}{|\mathbf{P }|}|\Lambda p,\sigma\rangle=\frac{p'_3}{p_3}|\Lambda p,\sigma\rangle$$

is equal to

$$|\Lambda p,\sigma\rangle$$

  • 1
    $\begingroup$ Why is $\boldsymbol J\cdot\boldsymbol P=J_3P_3$? $\endgroup$ Aug 15, 2019 at 15:17
  • $\begingroup$ Because i am assuming that the $|p,\sigma\rangle$ has momentum $p_3$ (in the $z$ direction) and angular momentum $\sigma=\sigma_3$ (in the $z$ direction) $\endgroup$ Aug 15, 2019 at 16:56
  • $\begingroup$ if $|p,\sigma\rangle$ has momentum in the $z$ direction, then $\frac{1}{|P|}|p,\sigma\rangle=\frac{1}{p_3}|p,\sigma\rangle$, and the factors of $p_3$ disappear. $\endgroup$ Aug 15, 2019 at 19:04
  • $\begingroup$ You are right but the norm of a vector is rotation invariant and so $\frac{1}{|P|}|\Lambda p,\sigma\rangle=\frac{1}{p_3}|\Lambda p,\sigma\rangle $ and then we would have $\frac{p'_3}{|\mathbf{P }|}|\Lambda p,\sigma\rangle=\frac{p'_3}{p_3}|\Lambda p,\sigma\rangle$. But the factor $\frac{p'_3}{p_3}$ does not disapear $\endgroup$ Aug 16, 2019 at 6:54

1 Answer 1


Your mistake is overlooking the fact that the parameter $\sigma$ measures the helicity referring to the actual direction of ${\bf p}$.

Therefore, following again the first (correct) part of your reasoning where $p$ was directed along $x_3$, simply replacing the unit vector ${\bf e}_3$ for the rotated axis $\Lambda {\bf e}_3$ you find that $|\Lambda p, \sigma\rangle$ satisfies $$\frac{{\bf J}\cdot {\bf P}}{|{\bf P}|}|\Lambda p, \sigma\rangle = (\Lambda {\bf e}_3)\cdot {\bf J}|\Lambda p, \sigma\rangle = \sigma |\Lambda p, \sigma\rangle\:.$$ As $\Lambda {\bf e}_3$ generally admits components along both $x_2$ and $x_3$, you cannot argue that $$J_2 |\Lambda p, \sigma\rangle = 0,$$ $$J_3 |\Lambda p, \sigma\rangle =\sigma |\Lambda p, \sigma \rangle,$$ and your claim does not follows.

  • $\begingroup$ Suppose that $\Lambda {\bf e}_3=\cos\theta{\bf e}_3+\sin\theta{\bf e}_2$ i am not seeing why $(\cos\theta J_3+\sin\theta J_2)|\Lambda p,\sigma\rangle=(\sigma \cos\theta +\sin\theta J_2)|\Lambda p,\sigma\rangle$ is equal to $\sigma |\Lambda p,\sigma\rangle$ $\endgroup$ Aug 16, 2019 at 9:29
  • 1
    $\begingroup$ By definition. It is the definition of the states $|p,\sigma\rangle$. Your second identity above is wrong $J_3|\Lambda p,\sigma\rangle \neq \sigma|\Lambda p,\sigma\rangle$. That is the same mistake you took in the post. $\endgroup$ Aug 16, 2019 at 9:41
  • $\begingroup$ If states do transform like this $U(\Lambda)|p,\sigma\rangle=e^{i\theta\sigma}| \Lambda p,\sigma\rangle.$ that means that the $z$ component of angular momentum does not change so we would have $ J_3|\Lambda p,\sigma\rangle =\sigma|\Lambda p,\sigma\rangle$ or am i understanding it wrong? $\endgroup$ Aug 16, 2019 at 9:48
  • 1
    $\begingroup$ $|p,\sigma\rangle$ has the property that $\sigma$ is the component along ${\bf p}$ of ${\bf J}$. If ${\bf p}$ is directed along ${\bf e}_3$, then $\sigma$ is the value of the component $J_3$. Therefore, keeping ${\bf p}$ as said, if $\Lambda$ is a rotation $\Lambda p$ has not spatial components directed along ${\bf e}_3$ and consequently $\sigma$ cannot coincide to the component $J_3$ any more. $\endgroup$ Aug 16, 2019 at 9:49
  • 1
    $\begingroup$ Please, stop insisting on your erroneous claims and try to focus on what I am trying to communicate you. Otherwise this discussion is useless. $\endgroup$ Aug 16, 2019 at 9:51

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.