5
$\begingroup$

It is known that classical electrodynamics is time reversal invariant if one assumes that the transformation laws under such operation are $$\mathbf E(t,\mathbf x)\mapsto\mathbf E(-t,\mathbf x)$$ $$\mathbf B(t,\mathbf x)\mapsto -\mathbf B(-t,\mathbf x)$$ $$\rho(t,\mathbf x)\mapsto \rho(-t,\mathbf x)$$ $$\mathbf J(t,\mathbf x)\mapsto -\mathbf J(-t,\mathbf x)$$

How are these transformations related to the time reversal $T$ of the full Lorentz group $O(1,3)$? Here I would like to assume matrix notation for simplicity, so that $$T = \begin{bmatrix}-1&0&0&0\\0&1&0&0\\0&0&1&0\\0&0&0&1\end{bmatrix}$$ which is the Jacobian of the transformation $(t,\mathbf x)\mapsto(-t,\mathbf x)$. If you think of the 4-current as a 1-form $J$ over space-time, and you assume this transformation, that is, time reversal, to be passive, i.e. just a change of coordinates, then $$(\rho,-\mathbf J)\mapsto(-\rho,-\mathbf J),$$ from which one actually deduces $\rho(t,\mathbf x)\mapsto-\rho(-t,\mathbf x)$ and $\mathbf J(t,\mathbf x)\mapsto\mathbf J(-t,\mathbf x)$. One has a similar situation when transforming the electromagnetic tensor $F$ with $T$, which then gives $\mathbf E\mapsto -\mathbf E$ and $\mathbf B\mapsto\mathbf B$, but on the other hand the constitutive tensor $\star F$ gives the expected transformation laws for the fields, namely the ones given above. Is this just a mere coincidence?

$\endgroup$
3
$\begingroup$

The problem here arises because the 4-current in the OP is assumed to be a 1-form, and after many years of accumulated rust on the subject I completely forgot that this is, strictly speaking, not the right geometrical object that can describe current density. Indeed, being a density, it must be a 3-form, and therefore the correct geometrical object is $$J = \rho\ \text dx\wedge\text dy\wedge\text dz + \text dt\wedge(\mathbf J\cdot\star\text d\mathbf x)$$ where $\star\text d\mathbf x$ is the Hodge dual in $\mathbb R^3$ of the formal vector $(\text dx,\text dy,\text dz)$. This object has now the correct transformation law under time reversal, since $\text dt\mapsto -\text dt$ and $\text d\mathbf x\mapsto\text d\mathbf x$, and therefore $$J\mapsto\rho\ \text dx\wedge\text dy\wedge\text dz - \text dt\wedge(\mathbf J\cdot\star\text d\mathbf x).$$

$\endgroup$

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.