0

Working through David Tong's sheet here https://www.damtp.cam.ac.uk/user/tong/qft/oh1.pdf and can't follow how to get the Levi-Cevita symbol out the front? Its equation 15. I was looking at trying to use an identity with the Kronecker Deltas in the energy momentum tensor but I am really stuck. Below is the conserved charge, where does the $\epsilon_{ijk}$ come from?

$$Q_i = \epsilon_{ijk} \int d^3x (x^jT^{0k} - x^kT^{0j})$$

I can follow up to the conserved current and get the correct value for charge but I am missing the Levi-Cevita

Qmechanic
  • 201,751
milkcookie
  • 23
  • 5
  • 1
    it's just the cross product. The expression is just a fancy way of writing $({\vec x} \times {\vec p})_{i}$ – Zo the Relativist Oct 29 '20 at 00:53
  • 1
    To be specific, $(\mathbf{a} \times \mathbf{b})^i = \epsilon_{ijk} a^j b^k$ – Nihar Karve Oct 29 '20 at 03:34
  • @JerrySchirmer thanks, I feel silly asking this but where exactly does the cross product come into this? I've tried expanding by substituting in what I know $T^{0k}$ is but its getting kinda messy – milkcookie Oct 29 '20 at 14:44
  • @NiharKarve thanks, where are you getting the cross product in the equation? I'm a bit stuck still sorry, i have the conserved current $j^\mu$ but struggling to use the Levi-Civita still in the charge – milkcookie Oct 29 '20 at 14:45
  • @milkybean: $T^{0k}$ is the same thing as the three-momentum ${\vec p}^{k}$. Then, remember that determinant rule for calculating the cross product? Well, the signs in the terms in the determiant alternate in exactly the same way that $\epsilon_{ijk}$ do, right? So, the $x$ component of ${\vec a} \times {\vec b}$ is $a_{y}b_{z} - a_{z}b_{y}$, for example. Put it all together, and you can work out that the $i$ component of ${\vec a} \times {\vec b}$ is $\epsilon_{ijk}a^{j}b^{k}$ – Zo the Relativist Oct 29 '20 at 15:24
  • Luboš Motl's excellent answer here explains the difference between $L_{ij}$ and $L_k$ – Nihar Karve Oct 29 '20 at 17:51

1 Answers1

0

The conserved quantity that arises due to spatial rotational invariance is $$ L^{jk} = Q^{jk} = \int \mathrm{d}^3 x\ (x^j T^{0k} - x^k T^{0j}) $$

Unlike the classical angular momentum pseudovector $\mathbf{x \times p}$, this is expressed as a matrix:

$$ L^{ij} = \begin{pmatrix} 0 & L_{xy} & L_{xz} \\ -L_{xy} & 0 & L_{yz} \\ -L_{xz} & -L_{yz} & 0 \end{pmatrix} $$

whose components may look familiar - they are similar to the magnetic field components in the electromagnetic field tensor $F^{jk}$. Again, similar to the electromagnetic case, to convert $L^{jk}$ back into the classical vector form, you have to use the Levi-Civita symbol: $$ L_i = \epsilon_{ijk}L^{jk} $$

However, as Luboš Motl remarks here, it is usually more efficient to leave it tensor form.

Nihar Karve
  • 8,445