When quantizing the free scalar field, we define positive frequency modes according to
$$\frac{\partial}{\partial t} \phi_{\omega}=-i \omega \phi_{\omega}. $$
In the mode expansion we then separate positive and negative frequency solutions,
$$\phi = \int dw \ a_w \phi_{\omega} + a^{\dagger}_w \phi^*_{\omega},$$
and consequently define the vacuum state $|0\rangle$ as being annihilated by the operators corresponding to the coefficients of the positive frequency modes,
$$ a_w |0\rangle.$$
This means that if we quantize the field using a different time $t'$, the definitions do not necessarily agree. I wonder how I can show that for those cases, where $t$ and $t'$ are related by a Poincare transformation, the definitions of the vacuum state match?