# horizontal shift of sub- and superscript for \chi

I get a small horizontal displacement of sub- and superscript when using the letter \chi. It does not seem to appear for other letters. Does anybody know how to fix this?

MWE:

\documentclass{article}
\usepackage{amsmath}
\usepackage[charter,expert]{mathdesign}
\usepackage[scaled=.96]{XCharter}
\begin{document}
$\chi_{0}^{0}$ $\nu_{0}^{0}$ $\mu_{0}^{0}$ $u_{0}^{0}$ $\mathrm{u}_{0}^{0}$
\end{document}


output:

• Welcome to TeX.SX! – GuM Dec 6 '18 at 9:38

The chi is slanted and has in the tfm an italic value:

(CHARACTER O 37
(CHARWD R 0.562)
(CHARHT R 0.4835)
(CHARDP R 0.219)
(CHARIC R 0.051) %<-----
)


This is taken into account by TeX when placing the sub/superscripts.

You can suppress it, but I don't think that this improve the look:

\documentclass{article}
\usepackage{amsmath}
\usepackage[charter,expert]{mathdesign}
\usepackage[scaled=.96]{XCharter}
\begin{document}
$\chi_{0}^{0}$ $\chi{}_{0}^{0}$ $\chi{\vphantom{\chi}}_{0}^{0}$
\end{document}


• Note that \vphantom is implemented through \mathchoice, and that a “4-way-choice node”, being not an atom, cannot carry super-/subscripts. Indeed, you can check with \showlists that an empty atom is inserted after the phantom, and that the super-/subscripts are appended to this empty atom. This means that the result is the same as if you had said $\chi{}_{0}^{0}$. If you want the phantom to be effective, say $\chi{\vphantom{\chi}}_{0}^{0}$`. – GuM Dec 6 '18 at 9:30
• @GuM ah, yes, not quite sure why I thought an empty atom doesn't work either. – Ulrike Fischer Dec 6 '18 at 9:34