Computations in asymptotically flat coordinates

51 Views Asked by At

For reference, I'm following the first line in proof of Lemma 6.1 of this paper. This is my first time working with asymptotic flatness, and I would greatly appreciate it someone would tell me if the general ideas in my calculations are correct. If I haven't made any major errors, I'd appreciate if someone can help me finish the proof.

We work in an asymptotically flat 3-dimensional manifold. My question concerns the second equality of $$\nabla|\nabla u|=\nabla(g^{11})^{1/2}=-\frac{1}{2}\nabla g_{11}+O(|x|^{-1-2q}),$$ where $u=x^1$ is a function defined previously, and $x^1,x^2,x^3$ are coordinates satisfying the decay conditions $|\partial^k(g_{ij}-\delta_{ij})|=O(|x|^{-q-k})$ for some $q>1/2$ and $k=0,1,2$. For simplicitly, let me denote $O(|x|^{-p})$ by $O_{-p}$ for all decay rates $p$. For completeness, the first equality follows from $$|\nabla u|^2=g(\nabla u,\nabla u)=g(\nabla x^1,\nabla x^1)=g(dx^1,dx^1)=g^{11}.$$ From here, we take the gradient of $\delta^i_j=g^{ik}g_{kj}$. $$0=\nabla(\delta^1_1)=\nabla(g^{11}g_{11}+g^{12}{g_{12}}+g^{13}g_{13})=(\nabla g^{11})g_{11}+g^{11}(\nabla g_{11})+\underbrace{\nabla(g^{12}g_{12}+g^{13}g_{13})}_{=O_{-2q-1}},$$ The decay on the last term follows from the following calculation. By AF, both $g_{ij},g^{ij}$ can be replaced by their Euclidean counterparts up to an error of $O_{-q}$. Therefore, $$\nabla (g^{12}g_{12})=\nabla((\underbrace{\delta^{12}}_{=0}+O_{-q})(\underbrace{\delta_{12}}_{=0}+O_{-q}))=\nabla(O_{-2q})=O_{-2q-1},$$ and similarly for $g^{13}g_{13}$ thereby finishing the claim. We continue, \begin{align*} 0&=(\nabla g^{11})(\underbrace{\delta_{11}}_{=1}+O_{-q})+(\underbrace{\delta^{11}}_{=1}+O_{-q})(\nabla g_{11})+O_{-2q-1}\\ &=\nabla g^{11}+(\nabla g^{11})O_{-q}+\nabla g_{11}+O_{-q}(\nabla g_{11})+O_{-2q-1}\\ &=\nabla g^{11}+O_{-q-1} O_{-q}+\nabla g_{11}+O_{-q}O_{-q-1}+O_{-2q-1}\\ &=\nabla g^{11}+\nabla g_{11}+O_{-2q-1}. \end{align*} It follows that $\nabla g^{11}=-\nabla g_{11}+O_{-2q-1}$. From here, we calculate \begin{align*} \nabla (g^{11})^{1/2}&=\frac{1}{2}(g^{11})^{-1/2}\nabla g^{11}\\ &=\frac{1}{2}(1+O_{-q})^{-1/2}(-\nabla g_{11}+O_{-2q-1})\\ \end{align*} From here, I'm unsure how to proceed with dealing with the factor $(1+O_{-q})^{-1/2}$.