# Conservation Law from Metric

• I

## Main Question or Discussion Point

Consider the following metric for a 2D spacetime:
$g_{tt} = -x$
$g_{tx} = g_{xt} = 3$
$g_{xx} = 0$

i.e.
$$g_{\mu \nu} = \left( \begin{array}{cc} -x & 3\\ 3 & 0 \end{array} \right)$$

Now, since the metric is independent of time (t), there is supposedly a conservation law containing $\frac{dx(\tau)}{d\tau}$ and $\frac{dt(\tau)}{d\tau}$. What is this conservation law, and why does the time independence of the metric imply it?

Related Special and General Relativity News on Phys.org
haushofer
To find the isometries of such a spacetime, you have to solve for the Killing equations. The tt-component of this equation should answer your question.

A note: GR in two dimensions is a bit problematic; the Einstein-Hilbert action is a topological term, meaning that the Einstein equations are an identity.

• vanhees71
pervect
Staff Emeritus
Consider the following metric for a 2D spacetime:
$g_{tt} = -x$
$g_{tx} = g_{xt} = 3$
$g_{xx} = 0$

i.e.
$$g_{\mu \nu} = \left( \begin{array}{cc} -x & 3\\ 3 & 0 \end{array} \right)$$

Now, since the metric is independent of time (t), there is supposedly a conservation law containing $\frac{dx(\tau)}{d\tau}$ and $\frac{dt(\tau)}{d\tau}$. What is this conservation law, and why does the time independence of the metric imply it?
Consider a curve $u^0(\tau) = t(\tau), u^1(\tau) = x(\tau)$. The tangent to this curve is a vector, i.e.

$$\left( \frac{dt}{d\tau}, \frac{dx}{d\tau} \right)$$

is a vector.

Because none of the components of the metric is a function of time, the vector (in the above sense) $\xi^a = (1,0) = dt/d\tau$ is a Killing vector.

The dot product of a tangent vector of a geodesic curve and a Killing vector is a conserved quantity - it is the same everywhere along the geodesic curve. If we let $u^a = (u^0(\tau), u^1(\tau) ) = (t(\tau), x(\tau))$ be a geodesic curve then we can say that the dot product is given by $g_{ab} \xi^a u^b = \xi_b u^b$ is constant. Because $\xi^a = (1,0)$ and $\xi_a = (-x,3)$ we can say that $-x (dt/d\tau) + 3 (dx/d\tau)$ is constant along the curve, i.e. is independent of $\tau$.

As a check, we can consider $(d/d\tau) (-x (dt/d\tau) + 3 (dx/d\tau)) = 0$, we know $d/d\tau$ of a constant is zero. Expanding this out using the chain rule, we get:
$$-\frac{dx}{d\tau} \frac{dt}{d\tau} + -x \frac{dt^2}{d\tau^2} + 3 \frac{dx^2}{d\tau^2} = 0$$

We can derive this from the geodesic equation. We compute the christoffel symbols

$$\Gamma^t{}_{tt} = \frac{1}{6} \quad \Gamma^x{}_{tt} = \frac{x}{18} \quad \Gamma^x{}_{tx} = -\frac{1}{6}$$

and confirm that we can derive the above from the geodesic equations, which in tensor notation are:

$$\frac{d^2 u^i}{d\tau^2} + \Gamma^i{}_{jk} \frac{du^j}{d\tau}\frac{du^k}{d\tau} = 0$$

Tensor notaton is very compact - the non-repeated index i generates two equations, one where we substitute i=0, the other where we substitute i=1. For each of these equations, j and k are repeated indices, so we sum over the repeated indicies, i.e we sum , for all four combinations of j,k, (0,0), (0,1), (1,0), (1,1).