I have the followwing Lagrangian for the free electromagnetic field,

$$\mathcal{L} = -\frac{1}{4} F^{\mu \nu}F_{\mu \nu},$$

and the canonical stress tensor is,

$$T^{\alpha \beta}=\frac{\partial \mathcal{L}}{\partial \left(\partial_{\alpha} A^{\lambda}\right)}\partial^{\beta} A^{\lambda}-g^{\alpha \beta}\mathcal{L}.$$

My $T^{0i}$ is, $T^{0i} = (\vec{E}\times \vec{B})_{i}$

but it's the symmetric stress tensor, and I need to find it:

$$T^{0i} = (\vec{E}\times \vec{B})_{i} + \nabla \cdot (A_{i}\vec{E})$$

I don't understand where this divergence term came from. How can I do this asymmetric $T^{0i}$ tensor?

(Just for reference, see Jackson - Classical electrodynamics, Third ed.- page 606)


Answering my own question

We have:

$T^{ab} = -\frac{1}{4\pi} g^{ac}F_{cd}\partial^b A^d - g^{ab}L_{elec}$

$g^{0i} = 0$. In the first term we have, (using $F_{00} = 0$):

$T^{0i} = -\frac{1}{4\pi} g^{0c}F_{cd}\partial^i A^d - g^{0i}L_{elec}$

we need $c=0$, because $g^{00}=1$:

$T^{0i} = -\frac{1}{4\pi} g^{00}F_{0d}\partial^i A^d$

with $d=i,j,k$

$F_{0d}=(E_i + E_j + E_k)$

then, whe have:

$T^{0i} = -(E_i\partial^i A^i + E_j\partial^i A^j + E_k\partial^i A^k)$

$T^{0i}= - \frac{1}{4 \pi} \left( E_x \partial^x A^x + E_y \partial^x A^y + E_z \partial^x A^z \right )$

using $- B^j = \partial_k A^i - \partial_i A^k $:

$T^{0i}= \frac{1}{4 \pi} \left(E_i \partial_i A^i + E_j \partial_j A^i + E_j B_k - E_k B_j +E_k \partial_k A^i \right ) $

with $ \nabla \cdot \vec{E} = 0 $:

$ T^{0i} = \frac{1}{4\pi} \left( ( \vec{E} \times \vec{B} )_i + \nabla \cdot (A_i \vec{E}) \right)$


| cite | improve this answer | |

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.