The t Hooft symbol is a collection of numbers which allows one to express the generators of the SU(2) Lie algebra in terms of the generators of Lorentz algebra. The symbol is a blend between the Kronecker delta and the Levi-Civita symbol. It was introduced by Gerard 't Hooft. It is used in the construction of the BPST instanton.
a | |
η | |
\mu\nu |
a | |
η | |
\mu\nu |
=\begin{cases}\epsilona\mu\nu&\mu,\nu=1,2,3\ -\deltaa\nu&\mu=4\ \deltaa\mu&\nu=4\ 0&\mu=\nu=4\end{cases}
\deltaa\nu
\deltaa\mu
\epsilona\mu\nu
In other words, they are defined by
(
a=1,2,3;~\mu,\nu=1,2,3,4;~\epsilon1=+1
ηa=\epsilona+\deltaa\delta\nu-\deltaa\delta\mu
\barηa=\epsilona-\deltaa\delta\nu+\deltaa\delta\mu
In matrix form, the 't Hooft symbols are
η1\mu\nu=\begin{bmatrix} 0&0&0&1\\ 0&0&1&0\\ 0&-1&0&0\\ -1&0&0&0\end{bmatrix}, η2\mu\nu=\begin{bmatrix} 0&0&-1&0\\ 0&0&0&1\\ 1&0&0&0\\ 0&-1&0&0\end{bmatrix}, η3\mu\nu=\begin{bmatrix} 0&1&0&0\\ -1&0&0&0\\ 0&0&0&1\\ 0&0&-1&0\end{bmatrix},
\bar{η}1\mu\nu=\begin{bmatrix} 0&0&0&-1\\ 0&0&1&0\\ 0&-1&0&0\\ 1&0&0&0\end{bmatrix}, \bar{η}2\mu\nu=\begin{bmatrix} 0&0&-1&0\\ 0&0&0&-1\\ 1&0&0&0\\ 0&1&0&0\end{bmatrix}, \bar{η}3\mu\nu=\begin{bmatrix} 0&1&0&0\\ -1&0&0&0\\ 0&0&0&-1\\ 0&0&1&0\end{bmatrix}.
They satisfy the self-duality and the anti-self-duality properties:
ηa\mu\nu=
1 | |
2 |
\epsilon\mu\nu\rho\sigmaηa\rho\sigma , \barηa\mu\nu=-
1 | |
2 |
\epsilon\mu\nu\rho\sigma\barηa\rho\sigma
Some other properties are
\epsilonabcηb\mu\nuηc\rho\sigma=\delta\mu\rhoηa\nu\sigma+\delta\nu\sigmaηa\mu\rho-\delta\mu\sigmaηa\nu\rho-\delta\nu\rhoηa\mu\sigma
ηa\mu\nuηa\rho\sigma=\delta\mu\rho\delta\nu\sigma-\delta\mu\sigma\delta\nu\rho+\epsilon\mu\nu\rho\sigma ,
ηa\mu\rhoηb\mu\sigma=\deltaab\delta\rho\sigma+\epsilonabcηc\rho\sigma ,
\epsilon\mu\nu\rho\thetaηa\sigma\theta=\delta\sigma\muηa\nu\rho+\delta\sigma\rhoηa\mu\nu-\delta\sigma\nuηa\mu\rho ,
ηa\mu\nuηa\mu\nu=12 , ηa\mu\nuηb\mu\nu=4\deltaab , ηa\mu\rhoηa\mu\sigma=3\delta\rho\sigma .
The same holds for
\barη
\barηa\mu\nu\barηa\rho\sigma=\delta\mu\rho\delta\nu\sigma-\delta\mu\sigma\delta\nu\rho-\epsilon\mu\nu\rho\sigma .
and
\epsilon\mu\nu\rho\theta\barηa\sigma\theta=-\delta\sigma\mu\barηa\nu\rho-\delta\sigma\rho\barηa\mu\nu+\delta\sigma\nu\barηa\mu\rho ,
Obviously
ηa\mu\nu\barηb\mu\nu=0
Many properties of these are tabulated in the appendix of 't Hooft's paper[1] and also in the article by Belitsky et al.[2]