Normalised basis

Description

By substituting

\[\vec{E}_i = H_i \vec{\hat{E}}_i\]

into the covariant relation:

\[\dedx{i}{j}{k},\]

I obtain

\[\begin{split}\pder{}{X^j} \left( H_i \vec{\hat{E}}_i \right) & = \pder{H_i}{X^j} \vec{\hat{E}}_i + H_i \pder{\hat{E}_i}{X^j} \\ & = - \sum_k \delta_{ij} \frac{H_j}{2 H_k} \pder{H_i}{X^k} \vec{\hat{E}}_k - \sum_k \delta_{ij} \frac{H_i}{2 H_k} \pder{H_j}{X^k} \vec{\hat{E}}_k + \pder{H_i}{X^j} \vec{\hat{E}}_i + \pder{H_j}{X^i} \vec{\hat{E}}_j,\end{split}\]

or simply (\(j = i\) is used to unify the first two terms)

\[\dehatdx{i}{j}{k}.\]

This relation is written down explicitly below for an intuitive understanding.

\[\begin{split}\newcommand{\diag}[2]{ - \frac{1}{H_{#2}} \pder{H_{#1}}{X^{#2}} \vec{\hat{E}}_{#2} } \newcommand{\nodiag}[2]{ \frac{1}{H_{#1}} \pder{H_{#2}}{X^{#1}} \vec{\hat{E}}_{#2} } \begin{pmatrix} \pder{\vec{\hat{E}}_1}{X^1} & \pder{\vec{\hat{E}}_2}{X^1} & \pder{\vec{\hat{E}}_3}{X^1} \\ \pder{\vec{\hat{E}}_1}{X^2} & \pder{\vec{\hat{E}}_2}{X^2} & \pder{\vec{\hat{E}}_3}{X^2} \\ \pder{\vec{\hat{E}}_1}{X^3} & \pder{\vec{\hat{E}}_2}{X^3} & \pder{\vec{\hat{E}}_3}{X^3} \\ \end{pmatrix} = \begin{pmatrix} \diag{1}{2} \diag{1}{3} & \nodiag{2}{1} & \nodiag{3}{1} \\ \nodiag{1}{2} & \diag{2}{1} \diag{2}{3} & \nodiag{3}{2} \\ \nodiag{1}{3} & \nodiag{2}{3} & \diag{3}{1} \diag{3}{2} \end{pmatrix}.\end{split}\]

Example

In cylindrical coordinates, I have

\[\begin{split}\begin{pmatrix} \pder{\vec{\hat{E}}_1}{X^1} & \pder{\vec{\hat{E}}_2}{X^1} & \pder{\vec{\hat{E}}_3}{X^1} \\ \pder{\vec{\hat{E}}_1}{X^2} & \pder{\vec{\hat{E}}_2}{X^2} & \pder{\vec{\hat{E}}_3}{X^2} \\ \pder{\vec{\hat{E}}_1}{X^3} & \pder{\vec{\hat{E}}_2}{X^3} & \pder{\vec{\hat{E}}_3}{X^3} \\ \end{pmatrix} = \begin{pmatrix} 0 & 0 & 0 \\ \vec{\hat{E}}_2 & - \vec{\hat{E}}_1 & 0 \\ 0 & 0 & 0 \\ \end{pmatrix}.\end{split}\]