LaTeX Accents and Umlauts

Quick answer

• In a LaTeX document: type the character. UTF-8 has been LaTeX’s default input encoding since 2018, so Gödel, café and Ångström work as they are. The commands (\"o, \'e, \AA) still work and give the same result.
• In a .bib file for classic BibTeX: write the accent command and put it in braces: G{\"o}del, caf{\'e}, {\AA}ngstr{\"o}m.
• In a .bib file for biblatex with biber: type the character in UTF-8, as in the document.

\documentclass{article}
\usepackage[T1]{fontenc}   % recommended with pdfLaTeX
\begin{document}
Gödel, café, Ångström, Erdős      % typed directly
G\"odel, caf\'e, \AA ngstr\"om, Erd\H{o}s   % the same with commands
\end{document}

author = {G{\"o}del, Kurt and Erd{\H{o}}s, Paul}
title  = {{\'E}tudes sur les fonctions de {\AA}ngstr{\"o}m}

Accent commands

Each command puts an accent on the letter that follows. Accents named by a symbol take the letter directly (\"o). Accents named by a letter need braces or a space (\c{c} or \c c). Every command works with any letter, upper or lower case.

Special letters

These letters are not built from a base letter and an accent. They have commands of their own.

Commands marked T1, and the ogonek \k, give the error Command unavailable in encoding OT1 unless the preamble contains \usepackage[T1]{fontenc}.

A command that ends in a letter swallows the space after it. Write Stra\ss e or Stra{\ss}e, and \o{} or {\o} before a space.

Letter by letter

The middle column is what you write in a LaTeX document if you can’t or don’t want to type the character itself. The right column is the form for a .bib file.

Umlauts: ä, ö, ü

German sharp s: ß

Example: Stra{\ss}e, Gau{\ss}. There is no \"s. The "s shorthand belongs to babel’s German support and should not be used in .bib files, see wrong forms.

é, è, ê and other vowels with accents

ç (c with cedilla)

Example: Fran{\c{c}}ois, Gon{\c{c}}alves.

ñ (n with tilde)

Scandinavian letters: ø, å, æ

Central European letters: ł, š, č, ž, ř, ő, ű

Turkish letters: ğ, ı, ş

Why BibTeX needs the braces

BibTeX looks at field values before LaTeX does: it sorts entries, changes the case of titles, abbreviates first names and builds labels such as [Göd31]. A group that starts with a backslash right after the opening brace, like {\"o}, is a special character to BibTeX and counts as one letter:

• Sorting. G{\"o}del is sorted as Godel. Typed as raw UTF-8 with classic BibTeX, Gödel ends up after Gzdel.
• Case changing. In a title, {\"U}ber is lowercased by sentence-case styles like any other letter, and a style that sets text in capitals turns {\"u} into {\"U}.
• Abbreviating names. {\"O}zge is abbreviated to {\"O}.
• Labels. The alpha style produces G{\"o}d31, with the accented letter counted once.

Accents on i and j

An accent replaces the dot of i and j, so the dotless forms \i and \j are used: {\"\i} for ï, {\'\i} for í, {\u{\i}} for ĭ. Current LaTeX removes the dot by itself for \'i, \`i, \^i and \"i, so na{\"i}ve prints correctly too. With other accents, write \i.

Forms that go wrong

Unicode in LaTeX documents

With pdfLaTeX, UTF-8 input works without any package since the 2018 release. Older installations need \usepackage[utf8]{inputenc}. Adding \usepackage[T1]{fontenc} is recommended: it provides real accented glyphs, so words with accents hyphenate correctly and can be copied from the PDF, and it provides ą, ð, þ, đ and ŋ.

With LuaLaTeX and XeLaTeX, input is Unicode throughout. Load neither inputenc nor fontenc, and choose a font that contains the characters you need.

If a character is missing from the font or unknown to LaTeX, pdfLaTeX stops with Unicode character … not set up for use with LaTeX. Use the command from the tables above instead, or switch to LuaLaTeX.

Unicode in .bib files

• biblatex with biber reads UTF-8 and sorts it correctly, including language-specific sorting. This is the simplest setup for bibliographies with many non-English names, see Using BibTeX.
• Classic bibtex works byte by byte. A UTF-8 ö is two bytes to it: it sorts after z, counts double in labels, and a label can be cut in the middle of a character, which stops LaTeX with an invalid UTF-8 error. Use the brace forms.
• bibtexu, included in TeX Live, is a BibTeX that understands UTF-8 and works with the usual .bst styles.
• bibtex8 handles 8-bit encodings such as Latin-1, not UTF-8.

Reference managers can do the conversion for you: JabRef and Zotero with Better BibTeX can export accented characters as LaTeX commands.

Accents in math mode

The text commands above do not work between $...$. Math mode has its own accents:

For several letters, use \widehat{abc}, \widetilde{abc} and \overline{abc}. Use \imath and \jmath for dotless i and j under an accent: \hat{\imath}. For words inside a formula, use \text{...} from amsmath with the text commands.

For characters such as &, %, $, # and _, see special characters.