At some point of an electrodynamics exercise, I had to prove:
$$ \nabla \frac{1}{|\textbf{r} - \textbf{r}'|} = - \frac{\textbf{r} - \textbf{r}'}{|\textbf{r} - \textbf{r}'|^3}$$
So, using Einstein Notation, I proceeded as follows:
$$\nabla [ (x_k - x_k')^2]^{-\frac{1}{2}} = \frac{\partial}{\partial x_i} [ (x_k - x_k')^2]^{-\frac{1}{2}} \textbf{e}_i = \ ...$$
and now the part I think I am getting wrong $$... \ (-1/2) [(x_k - x_k')^2]^{-\frac{3}{2}} \cdot 2\cdot (x_k-x_k') \cdot \frac{\partial x_k}{\partial x_i} \textbf{e}_i = $$
$$- [(x_k - x_k')^2]^{-\frac{3}{2}} (x_k - x_k') \delta_{ki} \textbf{e}_i = $$ $$ - [(x_i - x_i')^2]^{-\frac{3}{2}} (x_i - x_i') \textbf{e}_i = $$ $$- \frac{(x_i - x_i') \textbf{e}_i}{[(x_i - x_i')^2]^{\frac{3}{2}}}$$
which of course is not the quantity of the second-hand side. The quantity derived would be correct if it had a different index on the denominator, but the Kronecker delta reduced the double sum to a single sum. What did I do wrong?
$\textbf{Edit 1: Clarification}$
The operator nabla above is: $$\nabla = \frac{\partial}{\partial x_i} \textbf{e}_i$$ and the vectors $\textbf{r}$ and $\textbf{r}'$ are respectively: $$\textbf{r} = x_i \textbf{e}_i \text{ and } \textbf{r}' = x_i' \textbf{e}_i$$
Since $$\partial_i|\mathbf{r}-\mathbf{r}^\prime|^2=\partial_i[(x_k-x_k^\prime)(x_k-x_k^\prime)]=2(x_k-x_k^\prime)\partial_i(x_k-x_k^\prime)=2(x_k-x_k^\prime)\delta_{ik}=2(x_i-x_i^\prime),$$the chain rule gives$$\partial_i|\mathbf{r}-\mathbf{r}^\prime|^{-1}=-\frac12|\mathbf{r}-\mathbf{r}^\prime|^{-3}\partial_i|\mathbf{r}-\mathbf{r}^\prime|^2=-|\mathbf{r}-\mathbf{r}^\prime|^{-3}(x_i-x_i^\prime).$$This is the desired $i$th component.