The Kerr metric have two killing vectors:
$$t^{\mu} \equiv (k_{t})^{\mu} = (1,0,0,0)\hspace{5mm} \mathrm{and}\hspace{5mm} \phi^{\mu} \equiv (k_{\phi})^{\mu} = (0,0,0,1). \tag{1}$$
In general, it is possible to construct another killing vector given by,
$$\chi^{\mu} = a t^{\mu} + b \phi^{\mu}. \tag{2}$$
It is well-known that in Kerr spacetime the general killing vector is when: $a=1$ and $b=\Omega =- \frac{g_{t\phi}}{g_{\phi \phi}}|_{H}$.
So, using the fact that a horizon is a null surface, then we could try to find $a$ and $b$ using:
$$\chi_{\mu}\chi^{\mu}=0. \tag{3}$$
but in the end of the day my calculation resulted in:
$$\chi_{\mu}\chi^{\mu}= g_{tt}a^{2}+g_{\phi\phi}b^{2}+2g_{t\phi}ab=0. \tag{4}$$
So, my question is:
Given the Kerr metric $[1]$, and the general linear combination of the Killing vectors $(2)$, how can I prove that $a=1$ and $b=\Omega =- \frac{g_{t\phi}}{g_{\phi \phi}}|_{H}$?(*)
$[1]$ Kerr geometry: https://en.wikipedia.org/wiki/Kerr_metric
(*) It is possible to just define the well-known vector field $\chi^{\mu} = t^{\mu} + \Omega\phi^{\mu}$. You take it for granted and derive many important things. But at some point in black hole history someone faced the equation $(2)$ and by some kind of argument (a mathematical proof, say; or not) was stablished that the most interesting killing combination is when $a=1$ and $b=\Omega =- \frac{g_{t\phi}}{g_{\phi \phi}}|_{H}$, and that is the reasoning that I'm not grasping (in both physical and mathematical ways).