Consider the principle part, i.e., the part which contains the highest derivatives of the metric (which is the $2^{nd}$ derivative) is $$\mathcal{P}\{R_{ab}\}=\frac{1}{2}g^{cd}\left(\partial_{a}\partial_{c}g_{db}+\partial_{b}\partial_{c}g_{da}-\partial_{a}\partial_{b}g_{cd}-\partial_{c}\partial_{d}g_{ab} \right)\tag{1}.$$

The metric tensor can be expressed in terms of the Heaviside function on either side of spacetime as follows $$g_{ab} = \Theta(l)g_{ab}^{+}+\Theta(-l)g_{ab}^{-}\tag{2}.$$ There exists a unique $C^{1}$ atlas on the spacetime manifold $\mathcal{M}$, which is expressed as the disjoint union of two $C^{3}$ spacetimes $\mathcal{M}^{+}$ and $\mathcal{M}^{-}$ (each with boundary $\Sigma^{+}$ and $\Sigma^{-}$),which induces the given $C^{3}$ structures on $\mathcal{M}^{+}$ and $\mathcal{M}^{-}$ sucgh that the metric $g$ admist a continuous extension on the whole $\mathcal{M}$ iff their first fundamental forms $g^{+}$ and $g^{-}$ agree, i.e., $g^{+}=g^{-}$, or in component form, $$\left[g_{ab}\right]=g_{ab}^{+}-g_{ab}^{-}=0\tag{3}.$$ This is basically the first junction condition. The problem with using a metric of the form mentioned in equation $(2)$ is that when differentiated it yields

$$g_{ab,c}=\Theta(l)g_{ab,c}^{+} +\Theta(-l)g_{ab,c}^{-}+\epsilon\delta(l)\left[g_{ab}\right]n_{c} \tag{4},$$

where $\epsilon=n^{c}n_{c}$ and I have used the fact that $\Theta'(l) = \delta(l)$. Now, the problem arises in the last term of the equation $(4)$ which is singular and generates terms which are proportional to $\Theta(l)\delta(l)$ which are not distributions and hence, causes problems in computing the to used in the Christoffel symbols. To fix this the first junction condition mentioned in equation $(3)$ is employed. A similar problem arises when we compute the Riemann tensor and the Ricci tensor. To fix this we then employ the second junction condition.

Now, using the metric of type $(2)$ in $(1)$ and employing junction conditions $(3)$, $$\mathcal{P}\{R_{ab}\}=\mathcal{P}\{R_{ab}^{+}\}+\mathcal{P}\{R_{ab}^{-}\}\tag{5},$$ and hence, $$\mathcal{P}\{g^{ab}R_{ab}\}=\mathcal{P}\{R\}=\mathcal{P}\{R^{+}\}+\mathcal{P}\{R^{-}\}\tag{6}.$$ Using equations $(5)$ and $(6)$, we can write the Einstein equation in terms of the principle parts, for a source placed on a shell, as follows $$\mathcal{P}\{G_{ab}\}=T_{ab}\tag{7}.$$

It is now possible to express the equations in the form of a linear operator (since only the principle parts were considered) acting on the metric tensor on the LHS equal to the stress-energy tensor of the source on the RHS. Firstly, is this correct? And can this problem now be intepreted as an electrostatics Poisson-like equation whose solution could be intepreted as a weak solution to Einstein's equations?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.