4
$\begingroup$

Consider a spinor field $\psi(x)$. Its vacuum expectation value is given by $$v=\langle 0|\psi(x)|0\rangle.$$ Using the fact that the vaccum is invariant under Lorentz transformation, we get, $$v=\langle 0|\psi(0)|0\rangle.$$ Why is it that, if $v\neq 0$, the Lorentz invariance is broken?

$\endgroup$
9
$\begingroup$

The $v$ you write is itself a spinor, not a scalar. A non-zero spinor is obviously not invariant under Lorentz transformations, so a non-zero spinorial VEV breaks Lorentz invariance of the 1-point function.

$\endgroup$
  • $\begingroup$ @ACM But are 1-point functions related to measurables. Right? The LSZ reduction formula does not contain 1-point function (but $n$-point functions in general with $n>1$) and therefore, is not related to scattering amplitude. Am I wrong? So if 1-point functions are not measurables, should one care? $\endgroup$ – SRS Jan 18 '17 at 7:54
  • $\begingroup$ @SRS the derivation of the LSZ formula explicitly assumes that $v\equiv 0$. $\endgroup$ – AccidentalFourierTransform Jan 18 '17 at 14:09
  • $\begingroup$ @AccidentalFourierTransform Hmm. But if 1-point functions are Lorentz non-invariant, how does that make (the predictions of) the theory problematic? $\endgroup$ – SRS Jan 18 '17 at 14:13
  • 1
    $\begingroup$ @SRS because if $v$ is not invariant then $U(\Lambda)$ cannot exist (as in my answer below), which means that Lorentz transformations are not a symmetry of the theory (and, in particular, of the $S$ matrix). $\endgroup$ – AccidentalFourierTransform Jan 18 '17 at 14:16
4
$\begingroup$

To make ACM's argument more explicit, consider \begin{align} v&=\langle 0|\psi|0\rangle\\ &=\langle 0|\overbrace{UU^\dagger}^1\psi\overbrace{UU^\dagger}^1|0\rangle\\ &=\overbrace{\langle 0|U}^{\langle 0|}\overbrace{U^\dagger\psi U}^{D_\Lambda \psi}\overbrace{U^\dagger|0\rangle}^{|0\rangle}\\ &=D_\Lambda v \end{align} where $U=U(\Lambda)$ is the unitary operator that corresponds to Lorentz transformations in the Hilbert space, and $D_\Lambda$ its representation in the space of spinors.

Considering $\Lambda$ to be, say, a rotation around the $z$ axis with angle $\theta$, and expanding to first order in $\theta$, we get $$ S^zv=0 $$ which is impossible for representations of the Lorentz Group with half-integer spin, as $S^z$ has eigenvalues $$ -j,-j+1,\cdots,+j $$ none of which is zero.

Therefore, we must conclude that $U(\Lambda)$ doesn't exist, that is, the Lorentz symmetry is broken.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.